"Much Madness is divinest Sense" by Emily Dickinson:
Tinglong Dai
Tinglong Dai is the Bernard T. Ferrari Professor of Business and Professor of Operations Management and Business Analytics at the Carey Business School, Johns Hopkins University, with expertise in the areas of healthcare analytics, global supply chains, the interfaces between marketing and operations, and human–AI interaction. Dai's research primarily examines the health care ecosystem using analytics approaches, with a focus on behavioral, incentive, and policy issues related to healthcare operations management. More on Wikipedia
-
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
Carnegie Mellon University
@CarnegieMellon
Today, Carnegie Mellon University and Ken Griffin, founder and CEO of Citadel and founder of Griffin Catalyst, announced a historic $3 billion gift to elevate Carnegie Mellon in Pittsburgh and launch Carnegie Mellon University Miami. #cmumiami
Learn more:This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Saturday Night Live
@nbcsnl
Dario Amodei is at the desk to assure that the future of humanity is safe from AI
This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Avi Goldfarb
@avicgoldfarb
The 2026 NBER Economics of AI conference is available on YouTube. The theme was the economics of AI in China. Here are some of my takeaways.
Day 1: www.youtube.com/watch?v=pwODfPHl7WY
Day 2: www.youtube.com/watch?v=8cqRUc0W7Sc -
Tinglong Dai reposted
Andrea Presbitero
@a_presbitero
Great discussion on #geoeconomics @JHUCarey, touching upon how trade and financial flows adjust to a multipolar world, weaponized interdependence, and coercion. @TinglongDai @arebucci1 @LaynaMosley @sdonnan
-
Tinglong Dai reposted
The Chicago Maroon
@ChicagoMaroon
All social science Core classes at UChicago will transition to an “analog” approach for the upcoming academic year, restricting in-classroom technology use and banning AI-assisted writing and grading for students and professors.
chicagomaroon.com/53262/news/sosc-core-to-institute-ai-ba… -
Tinglong Dai reposted
The Lancet
@TheLancet
"...the most powerful argument for health AI is not cost savings or diagnostic speed; it is workforce survival."
Viewpoint—Global advances in health artificial intelligence: a workforce imperative www.thelancet.com/journals/lancet/article/PIIS0140-6736(2… -
Tinglong Dai reposted
Daniel C. Baumgart
@colitiscrohn
Is @BillGates right? He predicted #ai will replace many doctors within a decade — that humans “won’t be needed for most things.” @TinglongDai, @KathyMcDonald, (@JohnsHopkins University) and I argue the opposite in @TheLancet 🔗 doi.org/10.1016/S0140-6736(26)00693-8 #digital #health #medicine
-
Tinglong Dai reposted
PBS News
@NewsHour
The State Department said it takes "full responsibility" for "hastily" altering and then presenting a map of Africa at a global AIDS conference that mislabeled every single country it highlighted.
An analysis by Reuters found a watermark in the map image that suggests it was made with OpenAI. -
Tinglong Dai reposted
David Decosimo
@DavidDecosimo
University of Chicago Law School had decided to ban *not just AI* but laptops from the first year law classroom. This is the way!
-
Tinglong Dai reposted
François Chollet
@fchollet
Eventually, much of AI will converge towards intuition-guided symbolic world modeling, i.e. deep learning-guided program synthesis. It is inevitable. Symbolic modeling lets a system construct a compact, reusable, highly generalizable mental model of a problem space using minimal data.
-
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
New research from Carey faculty @TinglongDai and @TheHaiyang finds that waiving copays increases patients’ preference for medical AI, but they still seek reconfirmation from a doctor.
carey.jhu.edu/news/patients-warm-medical-ai-no-copay?__cf… -
Tinglong Dai reposted
Andrew Hammel
@AndrewHammel1
The cardiac unit of the Düsseldorf University Hospital, where people recover from heart surgery, is currently measuring 38°C/ 100.4°F. As you can see, the building is only about 15 years old. And built without central air-conditioning.
-
Tinglong Dai reposted
Seth Benzell - the LA-atic Stranger
@SBenzell
It’s in his paper on the St Petersburg Paradox (JEL 1977)
-
Tinglong Dai reposted
Michael D Abramoff
@MichaelAbramoff
Trust, reimbursement & AI
Or: the three stages of trust
What can physicians and other healthcare providers learn from AI creators — and what can AI creators learn from the provider experience?
There is a lot of debate about AI in healthcare and how it relates to physicians and other providers. Too much of it is physicians telling AI creators how it should be, and too little is about listening to how it can be.
It is interesting to look at this from a financial perspective. In this thread I assume the US third-party payer system as given, and I primarily use Medicare data, since commercial payers often follow Medicare and there is substantial documentation on the why and how. -
Tinglong Dai reposted
Benjamin Mazer
@BenMazer
Health care is typically among the last fields to adopt a new technology; I still use a pager, and I send faxes on a regular basis. But now chatbots are everywhere in medicine.
My latest story.
www.theatlantic.com/health/2026/06/ai-healthcare-uber-mom… -
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
Faculty, researchers, and clinicians from across Johns Hopkins gathered for a convening of the Workgroup on AI and Healthcare: carey.jhu.edu/news/ais-breakthrough-handshake
-
Tinglong Dai reposted
Kathy McDonald
@KathyMcDonald
AI must strengthen, not replace clinicians - with a special wish for nurses around the world to be deeply involved in shaping the future with AI for us all. #patientsafety #patientcentered #systems #workforce #globalhealth
@TheLancet with @TinglongDai and @colitiscrohn -
Tinglong Dai reposted
Timothy Beck Werth
@beck_werth
Thanks to the AI industry and RAMageddon, laptops, smartphones, gaming consoles, and other products are getting more expensive. Johns Hopkins professor @TinglongDai told me the AI industry should be worried about a backlash.
"I do think there is a backlash risk...I'm not sure AI leaders fully appreciate that risk. Jensen Huang calling the scarcity 'fantastic for us' may be true from NVIDIA's standpoint, but it is exactly the kind of quote that can look terrible in a policy debate." (1/3) -
Tinglong Dai
@TinglongDai
BE THE WHOLE PROBLEM
Address to Johns Hopkins MBA Class of 2026
May 18, 2026
Deans, faculty, staff, families, friends, and most importantly, the Johns Hopkins MBA Class of 2026, congratulations!
This is the third year in a row that Carey MBAs have invited me to do this. What an honor.
In 2024, I talked about death. In 2025, I talked about destruction. Despite these dark themes, you gave me a third chance.
So this year, I promised myself something sweet... Cheerful... Emotionally safe... Perfect for graduates and families who are celebrating...
I will talk about: DOOM.
Doom as in: We're All Doomed.
For those of you who took my artificial intelligence course, you have heard me say "we're doomed" in every class. Parents, yes, this is exactly how your children have been educated at Johns Hopkins.
❀ ❀ ❀
Let me begin with the story of an alien.
When you think of doom, what do you think of? Climate catastrophe. A.I. But the classic form of doom is alien invasion.
So let me ask: how many people in this room are worried about alien invasion? Raise your hands.
Just a couple of hands? Good. Very brave.
Now, how many people in this room have seen an alien in your lifetime? Raise your hands.
No one?
Well, I have. I was an alien for 14 years. Not metaphorically. Legally.
The American immigration and taxation system is so poetic and so imaginative. DHS and IRS looked at a person like me and declared: you must be an alien.
I was a non-resident alien for nine years and then a resident alien for five years. Apparently, over time, my species had adapted to local conditions.
I was proud to be one. Still am. An alien is someone who arrives to become a different person, with everything behind them and everything ahead.
If being an alien makes me mysterious, wait for it — I was illegal before I was an alien.
I was born in China in the 1980s, a millennial under the one-child policy. I had four siblings. I was the middle child. You see, my parents are true believers and practitioners of the Scaling Law. As far as the law of the land was concerned, however, I should not have existed.
I was an illegal child. Just as we debate illegal immigrants today, by some measure, I am an illegal person.
Before I had done anything wrong, before I had spoken my first words, I was already redundant. A rounding error in human form. I was the problem.
My parents went into huge debt so my family could exist. I feel two things at once. The money. And the love.
When I was born, the government saw an illegal being and an account receivable. My parents saw me.
When I came to America, the law saw an alien. Johns Hopkins saw a professor. And you saw someone worth listening to.
That is why the question of aliens is not really about outer space. It is about who gets to be fully human.
I have been thinking about what happened in Minneapolis. Watching the stories of Renée Nicole Good and Alex Pretti, I could not stop crying.
Renée Nicole Good. Alex Pretti. Good and Pretti. Those names say it all.
❀ ❀ ❀
Now let me tell you a bit about the beginning of my academic career.
In Spring 2013, I earned my PhD from Carnegie Mellon, and I received a faculty offer from Johns Hopkins. My wife was pregnant and we were expecting a baby in the summer.
Amid all the celebrations, I realized one thing that seemed unmistakable: my career was doomed before it even started.
All my papers were being rejected by top journals. The same journal rejected my job-market paper, the one that got me the job at Hopkins, twice.
And the rejections were not even subtle. They said: not only is this paper not good, it has no chance of getting better.
Within months, I experienced success and failure, so I feel qualified to compare them.
Success is sweet. Success makes every detour look like strategy. Success also makes people less funny; that is why Jerry Seinfeld feels less funny now than before.
Failure is rude. Failure is painful. And failure is personal.
Failure tells you what matters, who matters, and who celebrates you when there is nothing yet to celebrate.
Failure is discontinuity. Discontinuity is the precondition for poetry. Harold Bloom said the god of poets is not Apollo. Apollo is too steady, too continuous. The god of poets is Error, who lives at the back of a cave and only comes out in the dark of the moon.
I am not romantic about failure. Failure made me ridiculous.
I failed so much that during my first year at Carey, I bought a book by Scott Adams, with a title of "How to Fail at Almost Everything—and Still Win Big." I saw the first half of the book title and was like, "wow, this book is just for me."
You may not like Scott Adams' politics and I certainly don't. But the book has some good stuff. One trick the book taught me was affirmations. Pick a goal, phrase it as "I, xxx, will be a great astronaut." Then repeat it. Write it fifteen times a day. Say it out loud on the treadmill. The point wasn't magic. It was focus. His own version was "I, Scott Adams, will be a famous cartoonist." That worked out for him.
So I tried it.
There were days when I drove alone on I-95 and yelled, almost screamed, "I, Tinglong Dai, will become a great business school professor." Just like that, 15 times. I screamed without worrying about people driving next to me. This is Maryland. They've seen worse. The driver next to me was probably screaming too.
Sometimes it was raining. Very cinematic. Sometimes tears flowed nonstop, because I knew how ridiculous and how pathetic it sounded.
This was desperation in its rawest form.
This is what failure does. It strips away the audience. It leaves you alone in the car, in the rain, asking one question: do I still believe in a life that does not yet believe in me?
Years later, those rejected papers were accepted by the same journals after many, many revisions, tears, and smiles. And I have been appointed to edit the same journals.
People now look at my CV and assume I had an easy life publishing. But I failed more than anyone I knew, and I was, by any measure, the biggest failure I knew.
Failure is clarifying. It teaches us gratitude, if we let it.
Failure teaches us humility, if we survive it.
Failure teaches us love, if we are attentive to who is behind us.
My wonderful colleague Kathleen Sutcliffe, a pioneering scholar of management and organizations, writes that high-reliability organizations are first and foremost "preoccupied with failure." I love that phrase. Preoccupied with failure.
Failure forces us to change. It meets us in the corner we painted ourselves into. It finds us at the edge of the map. It speaks when the world we knew has ended, and asks: now what will you become?
Class of 2026, some of you came to Carey to change careers. Some of you came to change countries. All of you came to change the world.
But changing the world starts with changing yourself. Your Johns Hopkins MBA doesn't change yourself, but it makes you more interpretable to yourself.
❀ ❀ ❀
Now, let's go back to doom — and somehow, to a love story.
When I moved from Hong Kong in 2007, I walked around downtown Pittsburgh and was shocked. Everything felt messy and exposed. In Hong Kong, the streets felt orderly. In America, even the sidewalks seemed to be having an identity crisis.
The healthcare system was broken. The country was sleepwalking into a financial crisis. National debt was unsustainable. Infrastructure was crumbling. Do I even need to mention crime?
For a while, I genuinely suspected America was a scam. A conspiracy theory the rest of the world had quietly agreed to play along with.
And yet, over time, I began to understand. America's genius is not perfection. America's genius is a permanent sense of impending doom.
This country has been one step from complete collapse since 1776. We have been declaring "this is the end" roughly three times a week for two hundred and fifty years. We are very, very good at it.
Look out the window. See the harbor. See the fort across the water. Let's think about the War of 1812.
To be fair, one of America's first major moves in that war was a terrible decision: invading Canada.
But wars have a way of coming home. By 1814, the British had burned Washington and sailed for Baltimore. Their warships filled this harbor. They gave the fort an ultimatum: lower your flag by dusk, or be erased.
The flag stayed up. Through a night of fire, the defenders refused to lower it. At dawn, the flag was still there. Detained on a British ship in this harbor, Francis Scott Key watched the flag survive. The words he wrote that morning became our anthem. We still sing them, even though it is known to be hard to sing.
The harbor is still here. The fort is still here. America was one breath from gone. And then it wasn't.
Every administration, every crisis, every news cycle: the institutions buckle, the markets tremble, the center cannot hold. And then somehow, maddeningly, imperfectly, something moves.
A country recovers. An institution recovers. A career recovers. A person recovers.
America once called me alien. Then it let me become American. That contradiction is America.
And maybe that contradiction is also Carey: a young business school in America's oldest research university, still building, still self-advocating, still proving, still becoming.
A school that takes risks on people. I know this because it took one on me.
And maybe that contradiction is Johns Hopkins itself.
The largest donor in this university's history is Michael Bloomberg. Class of 1964, electrical engineering. C student in high school. Only listed extracurricular: math club. Hopkins said yes anyway. Almost by accident. Even ChatGPT wouldn't recommend admitting him.
In a recent Johns Hopkins Magazine essay, alongside one of my own, he said it himself:
"I was fortunate that Hopkins opened its doors to me. I was never a stellar student... I've always wondered how, after graduating, I got accepted to Harvard Business School, given my middling grades and lack of connections. Maybe the admissions office figured that since Hopkins' students believed in me, they would give me a shot, too."
One could argue Hopkins made a mistake. By his own admission, he was just "middling" at Hopkins. And yet he has given more than four billion dollars back to Hopkins, and counting. Without him, would Hopkins still be Hopkins as we know it?
You are a proud product of Johns Hopkins. And Johns Hopkins was a product of love, accident, stubbornness, and people making bets that looked irrational at the time.
In the 1810s, around the same time the U.S. almost ceased to exist, Johns Hopkins, the man, fell in love with his first cousin Elizabeth. Or so the story goes. Their Quaker family forbade the marriage. Neither of them ever married anyone else. He had no child to bear his name, and yet, Hopkins has become the most beloved and most trusted name in higher education. And his Hopkins family has grown bigger than anyone could have imagined.
A love that could not become a marriage became, in some mysterious way, a lovable institution.
A heartbreak became a gift. Without that love, without that unspent devotion, without that strange conversion of longing into legacy, do we even have Johns Hopkins?
Not every breakup should become a university. That would be too much pressure on modern love. But the deepest human energies do not disappear. If we are lucky, disciplined, and generous, they become something larger than ourselves.
And Hopkins Medicine, the jewel of this university, also came into being through people others were ready to underestimate.
By the 1890s, Hopkins was in trouble. The endowment had collapsed with the B&O Railroad. There was no money to open the medical school the will required. M. Carey Thomas, her lover Mary Elizabeth Garrett, and their circle of women raised the money. And raised the bar. Their conditions: women admitted on equal terms with men, and admission standards higher than any American medical school. The trustees did not know what hit them. They helped create not just a medical school, but a medical school worthy of Johns Hopkins.
M. Carey Thomas. If that name sounds familiar, it should. Same Carey family. Her father was one of the original trustees of Johns Hopkins. Six generations later, a Carey descendant gave us our business school.
Without them, the School of Medicine may never have opened in the form that made it famous.
If that is not the Hopkins story, I do not know what is.
❀ ❀ ❀
Here's the pattern:
Hopkins said yes to a mediocre student, and he became its greatest benefactor.
Hopkins said yes to women told they did not belong in medicine, and they built one of the world's greatest medical schools.
Hopkins said yes to a forbidden love, and that love became a university.
And Hopkins said yes to you, and to me — an illegal human being and an alien.
Hopkins saw the problem you and I were going to be, and said yes anyway.
So graduates, the world will tell you: don't be part of the problem.
I disagree.
Be the kind of problem that forces the world to become more alive.
DON'T BE PART OF THE PROBLEM. BE THE WHOLE PROBLEM.
Or as Taylor Swift put it:
"It's me, hi, I'm the problem, it's me...
I'll stare directly at the sun but never in the mirror."
The best institutions are not built by people who only know how to say yes to the obvious. They are built by people willing to see what others miss, protect what others dismiss, and widen doors they did not have to widen.
❀ ❀ ❀
Now, before I end, I want you to look around this room.
Look at your families, your partners, your children, your parents, your siblings, your friends, your classmates, your faculty and staff. The existence of each of us is a miracle. That we are all in the same room is impossible. Almost obscene.
Please give them a round of applause.
They're part of your miracle. They're part of your problem. They are the whole problem.
May you never confuse being impressive with being loved. May you never confuse being successful with being human.
May you be preoccupied with failure as much as you are preoccupied with success.
May you lower your expectations just enough to be grateful, and raise your ambition enough to keep moving.
❀ ❀ ❀
One last thing.
You are living in a love story.
The love that sees a child who was not supposed to exist. The love that welcomes an alien. The love that believes in a non-stellar student. The love that builds a university. The love that makes a country thrive on a permanent sense of impending doom.
Congratulations, Class of 2026.
As you start this journey, I want you to think about this story. In 1783, Benjamin Franklin watched a hot-air balloon take flight over Paris. It was a marvel. A miracle. But one skeptic in the crowd was not impressed. He asked Franklin: "What is the use of it?"
Franklin did not give an answer. He gave a problem worth solving. He asked back: "What is the use of a newborn baby?"
Class of 2026, you will face no shortage of skeptics. They will ask: what is the use of an MBA degree if A.I. can do this and do that?
Do not give them an answer. Give them a problem. Be a problem. Be a problem worth solving.
BE THE WHOLE PROBLEM.
Because every newborn baby, every new idea, every new institution, every new life begins as a problem the world does not yet know how to love.
Class of 2026: go be the problem the world will someday thank itself for having.
Go change the world! -
Tinglong Dai reposted
The Baltimore Sun
@baltimoresun
A $1 billion private container terminal breaking ground at Sparrows Point is being cast by officials and supply chain experts as a long-term bet on whether the Port of Baltimore can expand its role in an increasingly competitive East Coast shipping network.
The project at the Port of Baltimore is expected to expand container capacity by roughly 70%, adding two new berths and modern crane systems designed to handle larger cargo vessels. It comes as the port works to regain momentum following the collapse of the Francis Scott Key Bridge collapse, which temporarily shut down the main shipping channel and disrupted supply chains across the region.
Tinglong Dai, a supply chain expert at Johns Hopkins University’s Carey Business School, said the expansion represents a structural shift rather than incremental growth. “This is an important moment, a milestone,” he said. “It’s not just an expansion to the port. It’s really developing the status of the Port of Baltimore.”
Read more: www.baltimoresun.com/2026/05/01/port-of-baltimore-shippin…
🎥: Lloyd Fox and Natalie Jones, @baltimoresun -
Tinglong Dai reposted
steve hsu
@hsu_steve
Terrence Tao on AI and math: 5 stages of grief; denial fading
Same thing in theoretical physics! -
Tinglong Dai reposted
Nathan Lane
@straightedge
I keep thinking about this piece, which is enraging.
Nature Careers
@NatureCareers
Jane Ying Wu, who took her own life, was cleared of wrongdoing by the US National Institutes of Health. Her family alleges that her employer destroyed her career. go.nature.com/3OZDs51
-
Tinglong Dai reposted
INFORMS
@INFORMS
Congratulations to @Microsoft, the 2026 Franz Edelman Award winner! 🏆
-
Tinglong Dai
@TinglongDai
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
alz
@alz_zyd_
My advice to econ/finance PhD students in the age of AI:
Embrace the chaos. In Littlefinger's words, chaos is a ladder. Existing hierarchies will be disrupted, unlikely entrants will rise rapidly to the tops of fields thought to be impossible to disrupt -
Tinglong Dai
@TinglongDai
Shefali Patil, @ChrisGMyers, and I are so thrilled to share our latest publication on AI and medical decision making in @npjDigitalMed: www.nature.com/articles/s41746-026-02561-1
Much regulatory and policy attention on medical AI has been focused on vendors and users. We argue more should be on institutional implementation. This step shapes and scales clinical value judgment in a way never seen before.
In configuring AI systems, institutions calculate and embed normative trade-offs directly into their outputs, delivering a "black box" resolution to competing values. They operationalize hidden value tradeoffs at scale.
A sensitivity threshold redistributes attention, risk, and responsibility. It is an ethical choice.
We describe what this does to the people who have to live with these systems every day.
It creates the "attributability gap": a point where a clinician is compelled to execute a decision that is no longer attributable to their own ethical or clinical weighing of trade-offs. You sign off. You bear accountability. But the judgment was never yours.
Override the algorithm, and you face disciplinary action for "incorrectly" challenging algorithmic judgments.
Follow it, and you bear full accountability when it fails.
We call this "accountability ping-pong": oscillating between trusting and distrusting AI to avoid the most recent institutional sanction.
The immediate consequence is moral distress.
The longer-term effect is worse: the gradual atrophy of professional judgment.
When clinicians are repeatedly deterred from the "heavy lifting" of weighing competing clinical values, the foundations of holistic, patient-centered care erode.
We propose a collaborative governance framework, including federal mandates for standardized transparency via Model Cards, paired with multidisciplinary institutional reviews to deliberate and document value-laden configurations.
We call for preserving the clinical value judgment essential to patient-centered care, because the loss of it will be incalculably expensive. -
Tinglong Dai reposted
Dwarkesh Patel
@dwarkesh_sp
Terence Tao spent a year at the Institute for Advanced Study - no teaching, no random events of committees, just unlimited time to think. But after a few months, he ran out of ideas.
Terence thinks that mathematicians and scientists need a certain level of randomness and inefficiency to come up with new ideas.This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Margaretta Colangelo
@realmargaretta
Ophthalmology AI: Insights From The Frontier www.linkedin.com/pulse/ophthalmology-ai-insights-from-fro… @TinglongDai @JohnsHopkins @HopkinsMedicine #ai #ophthalmology @US_FDA @FDADeviceInfo
-
Tinglong Dai
@TinglongDai
New York
-
Tinglong Dai reposted
Chinmay Jani
@Jani_Chinmay
How will human expertise and machine intelligence reshape the future of health care delivery?
On Friday, March 6, I’ll be part of a panel at the @JohnsHopkins @JHUCarey Health Care Business Conference discussing “Human and Machine Synergy: The Next Frontier of Health Care Delivery.”
Joined by @VamsiVelcheti, Dr Risa Wolf, and Pinar Gorgulu, moderated by @TinglongDai
#aiinhealthcare #healthinnovation @SylvesterCancer @HemOncMiami -
Tinglong Dai reposted
Joshua Gans
@joshgans
As I collect these stories, Al Roth’s Nobel prize winning paper (“The Economics of Matching: Stability and Incentives,” Mathematics of Operations Research 1982) was initially rejected by the JPE.
-
Tinglong Dai reposted
-
Tinglong Dai
@TinglongDai
Colorado
-
Tinglong Dai reposted
NIHCM
@NIHCMfoundation
Announcing NIHCM's New Cohort of Research Grantees! nihcm.org/news/nihcm-awards-500-000-to-support-research-o…
10 grantees will examine pressing issues in #healthcare affordability, access, and quality—including AI scribes, perinatal market concentration, specialty drug pricing, and more.
@MSKCancerCenter @UCBerkeley @Wharton @ChildrensPhila @PublicHealthUMN @BrownHSPP @IndianaUniv @JohnsHopkins @WeillCornell @CM_Whaley -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Kevin Roose
@kevinroose
i follow AI adoption pretty closely, and i have never seen such a yawning inside/outside gap.
people in SF are putting multi-agent claudeswarms in charge of their lives, consulting chatbots before every decision, wireheading to a degree only sci-fi writers dared to imagine.
people elsewhere are still trying to get approval to use Copilot in Teams, if they're using AI at all.
it's possible the early adopter bubble i'm in has always been this intense, but there seems to be a cultural takeoff happening in addition to the technical one. not ideal! -
Tinglong Dai reposted
npj Digital Medicine
@npjDigitalMed
Commercial multimodal AI is entering healthcare fast. Yet, @TinglongDai shows that evidence still matters.
When evaluated for diabetic eye screening, off-the-shelf generative AI models underperformed retina specialists and did not meet FDA thresholds.
The authors argue this gap highlights the need for task-specific validation and new regulatory pathways before clinical use.
www.nature.com/articles/s41746-025-02216-7 -
Tinglong Dai
@TinglongDai
Tinglong Dai
@TinglongDai
Link to the @npjDigitalMed study:
www.nature.com/articles/s41746-025-02216-7 -
Tinglong Dai reposted
Dr. Marty Makary
FDA is now open to Bayesian statistical approaches. A leap forward!
Bayesian statistics can help:
✅ Clinical trial design
✅ Finding the optimal dose
✅ Extrapolation to children
✅ Leveraging phase 2 results in phase 3This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
SSRN
@SSRN
Check out the latest from the SSRN #blog with includes the top downloaded papers on #ai in Finance for Q4 2025.
Read more: blog.ssrn.com/2026/01/12/top-papers-on-ai-in-finance-q4-2…
#financetwitter #academicchatter #research -
Tinglong Dai reposted
INFORMS
@INFORMS
🎤 INFORMS member at #ces!
INFORMS Fellow Sridhar Tayur is speaking on “Quantum Leap: Computing’s Next Frontier in Health,” exploring how quantum computing could transform healthcare innovation.
🔗 www.ces.tech/schedule/quantum-leap-computings-next-fronti…
#informs #quantumcomputing #analytics #orms -
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
Tech with Mak
@techNmak
"Programmers will automate themselves out of existence."
That's what they said. And programmers laughed so hard they couldn't respond.
Two years into the AI revolution, here's what actually happened:
What the doomers predicted:
1./ AI writes all the code
2./ Developers become obsolete
3./ Only managers and marketers survive
4./ Programming becomes a dead career
What actually happened:
1./ AI writes code that needs debugging by developers
2./ Developers spend more time reviewing AI output than writing from scratch
3./ The skill gap between good and bad developers got WIDER, not narrower
4./ Demand for senior developers who understand what AI can't do went UP
The brutal irony nobody saw coming: AI didn't replace developers. It replaced the developers who thought AI would do their job for them.
Here's the real shift:
- Junior devs who learn to prompt AI? Productive.
- Senior devs who understand system design? Irreplaceable.
- Developers who just copy-paste without understanding? Already obsolete (AI just made it obvious).
- Managers who thought they could skip hiring developers? Drowning in unmaintainable AI-generated code.
We're not laughing because we're safe. We're laughing because the people who predicted our extinction don't understand what we actually do.
AI is a tool. And like every tool before it, it makes good developers better and exposes bad developers faster. -
Tinglong Dai
@TinglongDai
Excited to share my new paper with Prof. Terry Taylor @BerkeleyHaas, “Designing Enterprise AI Systems: Hallucination, Creativity, and Moral Hazard.”
Moral hazard is about hidden effort that’s essential to success at work. Since 2023, Terry and I have been wondering how GenAI changes moral hazard. When a system can draft, summarize, and generate on command, what is “effort” anyway?
In many settings, effort becomes vetting, correcting, refining, and knowing when to trust the output; the human role shifts from production to judgment.
By adopting GenAI, organizations are shaping the productive environment through AI system design. They can push the system toward more creative, exploratory output (more variability, more errors) or toward more predictable output (fewer errors, fewer surprises).
The same design choices that affect hallucinations also change incentives.
When output is more variable, employee effort matters more for success.
Outcomes become a sharper signal of human effort.
This means a firm may rationally tolerate a more hallucination-prone environment. In our model, that can reduce rent-sharing without necessarily reducing the agent’s surplus.
Tinglong Dai
@TinglongDai
Read our paper and share your thoughts: papers.ssrn.com/sol3/papers.cfm?abstract_id=5996714
-
Tinglong Dai reposted
Andrej Karpathy
@karpathy
I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become available over the last ~year and a failure to claim the boost feels decidedly like skill issue. There's a new programmable layer of abstraction to master (in addition to the usual layers below) involving agents, subagents, their prompts, contexts, memory, modes, permissions, tools, plugins, skills, hooks, MCP, LSP, slash commands, workflows, IDE integrations, and a need to build an all-encompassing mental model for strengths and pitfalls of fundamentally stochastic, fallible, unintelligible and changing entities suddenly intermingled with what used to be good old fashioned engineering. Clearly some powerful alien tool was handed around except it comes with no manual and everyone has to figure out how to hold it and operate it, while the resulting magnitude 9 earthquake is rocking the profession. Roll up your sleeves to not fall behind.
-
Tinglong Dai
@TinglongDai
Today, Professor Georgia Perakis of @MITSloan announced a new department at @INFORMS's flagship journal, Manufacturing and Service Operations Management Journal (M&SOM): "AI in Operations." @maxccohen (@mcgillu) and I are honored to serve as the founding Department Editors starting January 2026.
As Georgia put it, “AI has been rapidly evolving and is changing the field of Operations. Its impact is touching numerous Operations applications, ranging from healthcare to supply chains and more. One can argue the applications of AI arise, and can arise, in practically every area.”
And the department seeks to “help our community showcase and be an outlet for all the important work our community does in this area.”
In this new department, Maxime and I hope to spark and catalyze high-impact work on AI in operations, work that advances theory, empirics, behavioral foundations, and computational approaches, especially research that engages with generative and agentic AI.
Georgia, thank you for your leadership, and for giving our community a new home for this work.
This is our moonshot. This is our wheelhouse.
We’re in a golden age of Operations.
Tinglong Dai
@TinglongDai
Link to the announcement and our editorial statement: connect.informs.org/discussion/happy-holidays-and-new-dep…
-
Tinglong Dai reposted
Chris
@ChrisGPT
The majority of the public believes ChatGPT is looking up the exact answer in a database, like it’s some kind of Marvin Minsky era 1970s rule based expert system 😭.
We are so early bros..
Andy Masley
@AndyMasley
I've been wondering why the AI and copyright debate has been so bad, but this result makes it clear: 66% of people believe AI has all the art it trains on permanently stored inside it to reference and use.
-
Tinglong Dai
@TinglongDai
Ambient AI scribes may be the biggest healthcare development of 2025. No mandate. No reimbursement change.
They spread because doctors are drowning in documentation.
But while well-being gains are celebrated, a coding arms race is taking shape.
With @jkvedar and @healthecon_dan, we published “Policy brief: ambient AI scribes and the coding arms race” in @npjDigitalMed on early signals scribes can raise coding intensity: doi.org/10.1038/s41746-025-02272-z
Tinglong Dai
@TinglongDai
Link to our @npjDigitalMed article: “Policy brief: ambient AI scribes and the coding arms race,” with Dr. @jkvedar and Dr. @healthecon_dan: www.nature.com/articles/s41746-025-02272-z
-
Tinglong Dai reposted
François Chollet
@fchollet
LLMs represent the "library" phase of AI. The next phase will be the "scientist" phase. A library contains answers, but a scientist knows how to find answers that don't exist yet.
-
Tinglong Dai reposted
Analysis Group
@analysisgroup
Managing Principal Lisa Pinheiro recently joined a panel discussion on the strategic considerations involved in implementing AI in supply chain management, hosted by @JHUCarey.
-
Tinglong Dai reposted
SSRN
@SSRN
This vision statement examines #ai’s impact on five key areas of #supplychains, highlighting advances in forecasting, decision-making, & human–machine collaboration, as well as the supply chains that support AI.
Read more: papers.ssrn.com/sol3/papers.cfm?abstract_id=5792542
#aicommunity -
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
Pharmaceutical supply chain issues in the U.S. have become a health and national security threat. In the latest #speedbrief, @JohnsHopkinsSPH's Mariana Socal discusses Hopkins' answer to this growing problem.
#supplychain #drugshortagesThis media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
Professor @TinglongDai explores whether AI tools benefit doctors as gatekeepers or a second opinion for care. The answer depends on patient risk.
carey.jhu.edu/articles/ai-health-care-when-should-lead-wh… -
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
The U.S. has faced a shortage of the cancer drug leucovorin for 13 years. And it's not alone.
Drug shortages like this reflect deeper supply chain vulnerabilities that leave patients and providers with too little information and too few options.
Johns Hopkins experts launched the Pharmaceutical Supply Chain Dashboard—the first of its kind to track drug production, shortage history, and risk factors across over 60,000 drugs—to address the issue.
With this information, policymakers can see vulnerabilities in the supply chain before they cause harm.
Read more about the issue and Hopkins experts' answers: washingtondc.jhu.edu/news/hopkins-drug-supply-chain-dashb…
#drugshortages #supplychain #policy -
Tinglong Dai reposted
Ash Paul
@pash22
Drs may view peers who use AI as less competent, says new @TinglongDai et al research
www.futurity.org/doctors-artificial-intelligence-peers-33… -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Kevin Roose
@kevinroose
This is correct, and also true of every recent AI paper (the METR slow-down study, the Apple reasoning one) that casts doubt on AI's effectiveness. People are desperate to prove that LLMs don't work, aren't useful, etc. and don't really care how good the studies are.
rohit
@krishnanrohit
I think folks who know better, esp on twitter, are still underrating the extreme impact the MIT paper had about 95% of AI projects failing within enterprises. I keep hearing it over and over and over again.
-
Tinglong Dai reposted
Dr. Marty Makary
Beautiful teamwork. We’re rewarding companies that bring pharmaceutical manufacturing back to the U.S., address public health needs in new ways, and lower prices for Americans.
U.S. FDA
@FDA
FDA has awarded 6 additional National Priority Vouchers. The latest recipients tackle major health challenges: HER2 lung cancer, drug-resistant TB in children, rectal cancer, sickle cell disease, and obesity. When companies commit to accessibility and national priorities, we commit to speed. www.fda.gov/news-events/press-announcements/fda-awards-se…
-
Tinglong Dai reposted
François Chollet
@fchollet
ML research is an engineering discipline, not a philosophy seminar. You build, you test, you learn. Untested ideas are just speculation.
-
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
Global health emergencies are just one potential disruption that can put the American drug supply at risk.
#geopolitical tensions stemming from armed conflicts and trade wars, for example, also threaten the pharmaceutical supply chain, as do natural disasters that impact domestic supply.
🔗 Read the full story: washingtondc.jhu.edu/news/hopkins-drug-supply-chain-dashb… #drugprices #supplychain #publichealth -
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
The dashboard can help leaders make informed policy decisions to mitigate drug shortages in the short term and develop resilience strategies for the long term by:
1️⃣Identifying drugs of national interest
2️⃣Anticipating inspection-triggered shortages
3️⃣Spotting over-dependencies
4️⃣Making strategic procurement decisions
5️⃣Determining onshoring opportunities
#supplychain #publichealth #drugprices -
Tinglong Dai
@TinglongDai
St. Louis
-
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
Last month, the drug leucovorin made headlines as the FDA moved to approve the medicine for potential autism treatment.
Missing from the coverage? The fact that the U.S. has faced a shortage of leucovorin—typically used for cancer treatment—for 13 years, indicative of broader supply chain issues driving drug shortages in the U.S.
Thread ⤵️ #supplychain #drugprices #publichealth -
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
Supply chain disruptions risk more shortages, posing threats to public health and national security.
💵 About 1 in 4 Americans cannot afford the drugs they need
💊 Only 14% of active pharmaceutical ingredients are manufactured domestically
#supplychain #drugprices #publichealth -
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
The team has worked together for more than a year to identify the right #supplychain data sources, then link those sources together to help identify and address the key vulnerabilities that threaten the security of the U.S. drug supply chain.
-
Tinglong Dai reposted
David-Dan Nguyen
@DavidDan_Ngn
Our work adds to the regulatory & transparency research @JAMAplusAI
@JAMAHealthForum: lack of clinical validation resulting in early recalls
jamanetwork.com/journals/jama-health-forum/fullarticle/28…
@JAMANetworkOpen: absence of development details/generalizability
jamanetwork.com/journals/jamanetworkopen/fullarticle/2833324
(and more) -
Tinglong Dai reposted
JAMA+ AI
@JAMAplusAI
From @JAMAHealthForum: FDA-cleared #ai medical devices often experience recalls shortly after clearance, particularly those without clinical validation and produced by publicly traded companies.
jamanetwork.com/journals/jama-health-forum/fullarticle/10… -
Tinglong Dai
@TinglongDai
Atlanta
-
Tinglong Dai
@TinglongDai
With @US_FDA Commissioner @DrMakaryFDA, @dr_msocal, @drJoshS, and Gerard Anderson, what a great start to an insightful day at the @JHUBloombergCtr.
@JohnsHopkins × @theNASEM = advancing drug supply chain resilience through insight and action. -
Tinglong Dai reposted
Johns Hopkins Data Science and AI Institute
@HopkinsDSAI
Special thanks to our speakers @NTrayanova, @mdredze, @hsairmd, @alexander_baras, @TinglongDai, @alexisjbattle, Rick Carrick, Robert Stevens, Ashley Kieman, Agbessi Amouzou
-
Tinglong Dai reposted
Johns Hopkins Data Science and AI Institute
@HopkinsDSAI
Thank you to all who joined us for yesterday’s “Future of Your Practice: AI Innovations Reshaping Clinical Care” event! Attendees explored how AI is transforming clinical decision-making and shaping the future of patient care. #aiinhealthcare @HopkinsMedicine @HopkinsEngineer
-
Tinglong Dai
@TinglongDai
Chicago
-
Tinglong Dai reposted
Jeffrey Wooldridge
@jmwooldridge
In 2025, I've managed to become unfocused, uninteresting, and unproductive. Lucky me, the trifecta.
-
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
"The global supply chain that we built for the past 30 years is unraveling, and we can no longer afford to take it for granted," said @JHUCarey's @TinglongDai.
Get an in depth look at the issue and how a first-of-its-kind Hopkins dashboard will provide policymakers a pathway to build a more sustainable drug supply chain in the U.S.: washingtondc.jhu.edu/news/hopkins-drug-supply-chain-dashb… -
Tinglong Dai reposted
Hahrie Han 한 하 리
@hahriehan
Totally, completely flabbergasted. #macfellow
hub.jhu.edu/2025/10/08/hahrie-han-macarthur-fellowship-ge… -
Tinglong Dai
@TinglongDai
Dr. Mary Brunkow, a winner of 2025 @NobelPrize in Physiology/Medicine, isn’t a professor or lab head; not even a PI.
She’s a program manager at @isbsci & a member of @ISBLeeHood Lab
Tinglong Dai
@TinglongDai
Her bio says: “Since joining ISB in 2009, I have worked to support a number of different projects: Family Genomics in a wide variety of disease areas, the systems biology of Lyme disease, sepsis biomarkers, and scientific wellness.”
Tinglong Dai
@TinglongDai
According to @Glassdoor, the total salary range for her position is $135K–$184K per year
Tinglong Dai
@TinglongDai
Before that, she worked as a science writer at @isbsci.
Brilliance really can come from anywhere.
@NobelPrize has never felt more inspiring. -
Tinglong Dai
@TinglongDai
Thrilled to share my 2nd @JAMAHealthForum publication, "US Antibiotic Importation and Supply Chain Vulnerabilities": jamanetwork.com/journals/jama-health-forum/fullarticle/10…
Led by the brilliant @dr_msocal, this first-of-its-kind study traced every antibiotic imported into the United States from 1992 to 2024.
The picture that emerged is striking.
The U.S. now sources antibiotics from about 50 countries. Among them, 60+% of the active ingredients come from China and India is the leading supplier of finished medicines. The supply chain may look global, but it depends on an ever-shrinking number of countries and, in many cases, just a few factories.
Efficiency has been the name of the game. Over three decades, U.S. antibiotic imports increased 26-fold while prices fell by more than 90%.
But that efficiency came at a price. Antibiotic shortages are common and persistent. Today, the world’s largest healthcare system depends on a narrow industrial base located halfway around the globe. When a factory stumbles or an export policy changes, hospitals across the U.S. feel the ripple effect.
This work, supported by the @JohnsHopkins Nexus Award and DoD, is part of our ongoing effort to rethink the resilience of essential medicine supply chains. It was a privilege to work with @dr_msocal, Yunxiang Sun, @jeromie_b, Jen Dailey Lambert, and Maqbool Dada on this study.
From @JohnsHopkinsSPH to @JHUCarey to @JHUAPL, our shared goal is to strengthen the nation’s preparedness and ensure reliable access to life-saving drugs. -
Tinglong Dai reposted
JAMA Health Forum
@JAMAHealthForum
Over the past four decades, US #antibiotic imports increased markedly with sourcing for finished dosage forms diversified across many countries, while active #pharmaceutical ingredients relied primarily on China.
jamanetwork.com/journals/jama-health-forum/fullarticle/10… -
Tinglong Dai reposted
Sebastien Bubeck
@SebastienBubeck
Well, this time it's by Terence Tao himself: mathstodon.xyz/@tao/115306424727150237
-
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
Research from the Johns Hopkins Carey Business School finds that despite pressure on clinicians to be early adopters of AI, many face skepticism from peers for using it, raising concerns about the future of AI in health care settings.
carey.jhu.edu/articles/doctors-who-use-ai-are-viewed-nega… -
Tinglong Dai
@TinglongDai
With @davidsimchilevi, Ishai Menache, & Michelle Wu, I am excited to share the 4I Framework for bringing mathematical optimization closer to the world through generative AI.
There is something magical about math. People’s hearts beat faster and their eyes shine when they see math equations, even if they do not fully understand them.
At the same time, that sense of mystery has often kept optimization from being used as widely as it should be. Its potential is enormous, but its real-world adoption has lagged behind.
Back in June, on the historic and magnetic @Cambridge_Uni campus, just after Prof. @STayur's 60th birthday workshop (organized by my wonderful academic sibling, Prof. @FeryalErhun), our team began asking: How can we make optimization more relatable? How can we give frontline decision makers the ability to quickly see what is going on, make sense of model recommendations, and even play around with possibilities?
Those conversations grew into the 4I Framework:
💡 Insight: A trusted and up-to-date view of the system
🔍 Interpretability: Clear explanations of model logic and trade-offs in business terms
💬 Interactivity: The chance to ask what-if questions and explore scenarios in a conversational way
🔄 Improvisation: Flexibility to re-optimize when assumptions or conditions change
Our hope is that these principles will speak to both optimization scholars and practitioners. With the help of generative AI, we can make one of the world’s best kept secrets, which helped win World War II and has powered countless business successes, more accessible to the people who make decisions every day.
We look forward to your thoughts and suggestions!
Tinglong Dai
@TinglongDai
Link to our paper: ssrn.com/abstract=5511218
-
Tinglong Dai
@TinglongDai
“The clearest AI upside? Medicine. From curing cancer to extending lifespan, machines could help us achieve what humanity alone never managed.” — @DarioAmodei
-
Tinglong Dai
@TinglongDai
Attending @AnthropicAI Futures Forum in DC Union Station. @DarioAmodei and @jackclarkSF are having a fireside chat, so far mostly about national security implications
-
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
The nexus for trusted academic experts, global leaders, policymakers, artists, and scholars. 🏛️ #hopkinsbloombergcenter
-
Tinglong Dai reposted
Ethan Mollick
@emollick
In talks with companies about AI they mention: Reddit is the key source of AI data (it isn't & that was written by an SEO company) and the "MIT says 95% of AI pilots fail" thing (seems to be based on 52 interviews at a conference & unclear methodology)
So much bad info out there -
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
Over 90% of recalled AI-enabled medical devices were from public companies, a new study by Carey Professor @TinglongDai shows. Read more from @MorningBrew: www.healthcare-brew.com/stories/2025/09/03/recalled-ai-en…
-
Tinglong Dai reposted
SSRN
@SSRN
This article finds that import #tariffs in complex #supplychains, such as U.S. pharmaceuticals, lead to much smaller consumer price increases than the headline #tariff rate—often around 1% for a 25% tariff.
Read: papers.ssrn.com/sol3/papers.cfm?abstract_id=5237643
Subscribe: www.ssrn.com/index.cfm/en/johns-hopkins-bsr/ -
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
Is investor pressure driving risky AI medical device launches? New study by Carey Professor @TinglongDai finds an association between publicly traded companies and recalls of AI-based medical tools. Read more: carey.jhu.edu/articles/investor-pressure-driving-risky-ai…
-
Tinglong Dai reposted
Kevin Roose
@kevinroose
I'm surprised by how much demand there is for "GPT-5 is a bust/AI is hitting a wall" takes.
Some of it is wishful thinking or motivated reasoning, but some of it feels like a deep yearning for equilibrium -- people really don't want to live in a world that is constantly changing. -
Tinglong Dai
@TinglongDai
Thrilled to be speaking at @WashingtonPost next Tuesday (Sept 9) for @PostLive on how AI is transforming medicine.
The role of AI in medicine feels both clearer and more confusing every day. The real test is how these almost superhuman tools shape diagnosis, treatment, and the fragile web of trust between patients, clinicians, hospitals, and the rest of the healthcare ecosystem.
Prof. @ChrisGMyers and I, both of @JHUCarey, will explore where AI is already changing medicine and what it really takes to get it right.
We’ll be joining @RepBera (D-Calif.), Amir Dan Rubin (CEO, Healthier Capital), Sharmila Makhija (Founding Dean & CEO, @ALWMedSchool), Reid F. Conant, MD (Senior Physician Executive, @AbridgeHQ), and @MattMacVey (Executive VP & CIO, @ChildrensNatl).
Thank you to Christine Collins @JHUCarey for making this event possible and for including us in the conversation. -
Tinglong Dai reposted
Yale CRRIT
@Yale_CRRIT
This team comprised of @JohnsHopkins Branden Lee, Patrick Kramer, Sara Sandri, @Georgetown Ritika Chanda, Crystal Favorito (Johns Hopkins) Olivia Nasef (Johns Hopkins), CRRIT Co-Director @jsross119, @drJoshS (Johns Hopkins), & @TinglongDai (Johns Hopkins) (3/4)
-
Tinglong Dai reposted
Yale CRRIT
@Yale_CRRIT
We are excited to share a recently published study in @JAMAHealthForum, determining whether the lack of clinical validation and publicly traded company status were associated with recalls of AI-enabled medical devices (AIMDs). (1/4)
-
Tinglong Dai
@TinglongDai
Great to catch up with @EliseReuter of @MedTechDive, who covered our @JAMAHealthForum study on AI device recalls: www.medtechdive.com/news/ai-medical-devices-no-validation…
The last time we spoke was back in 2022, when we were talking about how COVID test makers scrambled to keep up with shifting demand.
This time, the urgent issue is the lack of clinical validation for AI-enabled devices. Our study shows the vast majority of recalled devices never went through any such validation. And the consequences were immediate: 60 devices accounted for 182 recalls, nearly half of which happened within the first year of FDA authorization.
Read more: -
Tinglong Dai
@TinglongDai
A big thank you to Maggie Ward @JHUCarey for covering our @JAMAHealthForum study on AI-enabled medical device recalls and the urgent risks for patient safety: carey.jhu.edu/articles/investor-pressure-driving-risky-ai…
We found publicly traded companies drive nearly all the recalls of AI devices.
Even more alarming, many recalled devices were cleared without any clinical validation.
Investor pressure and gaps in @US_FDA oversight mean patients can become the test subjects. -
Tinglong Dai reposted
JAMA Health Forum
@JAMAHealthForum
FDA-cleared #ai medical devices often experience recalls shortly after clearance, particularly those without clinical validation and produced by publicly traded companies. jamanetwork.com/journals/jama-health-forum/fullarticle/10… @TinglongDai @drJoshS
-
Tinglong Dai
@TinglongDai
Imagine a day when machines exceed us. What are the opportunities, the risks, and what must remain human?
On September 12, 2025 (12–1 PM ET), Dr. Risa Wolf and I will moderate a conversation on Artificial Superintelligence (ASI) and the future of healthcare.
We’re joined by two leading voices:
- Dr. Girish Nadkarni, Chairman of Windreich Department of Artificial Intelligence and Human Health at @MountSinaiNYC
- Dr. Bryant Lin, Clinical Professor of Medicine of @StanfordMed & Co-founder of @SoundHealth
This webinar is part of @JHU_HBHI's "Conversations on the Business of Health."
Tinglong Dai
@TinglongDai
RSVP for the webinar on "Artificial Superintelligence (ASI) and the Future of Healthcare" here: hbhi.jhu.edu/events/artificial-superintelligence-asi-and-…
-
Tinglong Dai reposted
Dr. Giovanni E Cacciamani MD MSc FEBU 🇮🇹🇺🇸🇪🇺
@Cacciamani_MD
With 1,000+ FDA-cleared AI devices, the real test is #trust: Can we ensure consent, privacy & adaptability when deletion isn’t enough? Great perspective by
@TinglongDai et al.
Unlearning In Medical AI: A New Frontier For Privacy, Regulation, And Trust | Health Affairs www.healthaffairs.org/content/forefront/unlearning-medica… -
Tinglong Dai
@TinglongDai
Delighted to share my @Health_Affairs Forefront article about "To sustain trust, machines may need to learn how to forget," with my @JohnsHopkins colleagues, Drs. Risa Wolf and Haiyang Yang.
AI is transforming medicine, but trust depends not just on what machines learn, but also on what they can unlearn.
Even when data are removed from a system, an AI model may still be retaining the lessons it learned from that data. This makes unlearning far harder than simply deleting rows in a database.
The idea of machine unlearning raises hard questions: What happens when deletion isn’t enough? What if retraining costs millions? What if privacy laws and regulatory approvals collide?
The @US_FDA has cleared more than 1,000 AI-enabled medical devices for clinical use. It has also begun to address post-deployment modifications, proposing a framework for “adaptive AI” systems that evolve over time. However, most discussions to date have focused on learning (the addition of new data) rather than unlearning.
At stake is not just technology, but trust. Consent, once given, is not necessarily final when the context shifts substantially. And ultimately: Medical AI will only gain ground if it earns and maintains the public’s trust.
Tinglong Dai
@TinglongDai
Link to our piece: www.healthaffairs.org/content/forefront/unlearning-medica…
-
Tinglong Dai reposted
Health Affairs
@Health_Affairs
In their new Forefront article, @TinglongDai, Risa Wolf, and Haiyang Yang from @JohnsHopkins discuss how, even if data are removed from a system, it’s possible that an AI model may still be retaining the ‘lessons’ it learned from training on that data in the past.
-
Tinglong Dai reposted
JAMA Health Forum
@JAMAHealthForum
FDA-cleared #ai medical devices often experience recalls shortly after clearance, particularly those without clinical validation and produced by publicly traded companies.
jamanetwork.com/journals/jama-health-forum/fullarticle/10… @TinglongDai @drJoshS -
Tinglong Dai
@TinglongDai
Chicago
-
Tinglong Dai
@TinglongDai
Delighted to share my 1st publication in @JAMAHealthForum, exploring the patient safety implications of @US_FDA-regulated, AI-enabled medical devices: jamanetwork.com/journals/jama-health-forum/fullarticle/28…
Our study finds AI devices, especially those developed by publicly traded companies, are significantly more likely to be recalled. Many of these tools were never tested on patients before entering clinical use.
Publicly traded companies, the big fish in this still-small pond, developed just over half of the devices but were responsible for nearly all of the recalls.
We were stunned to find that almost half of all AI device recalls occurred within the first year after approval. Even more concerning: 95% of these recalls involved tools with no reported clinical validation.
If AI hasn’t been tested on people, then people become the test.
Our team, spanning @HopkinsMedicine, @JHUCarey, @JohnsHopkinsSPH, @YaleMed, and @GUMedicine, examined nearly 1,000 @US_FDA-cleared AI devices.
Grateful to Branden Lee, who brilliantly led our phenomenal group of @JohnsHopkins medical students—Patrick Kramer, Sara Sandri, Ritika Chanda, @CrystalFavorito, and Olivia Nasef—and to co-authors Drs. @jsross119 and @drJoshS. -
Tinglong Dai
@TinglongDai
Champaign–Urbana
-
Tinglong Dai reposted
Temerty Centre for AI in Medicine (T-CAIREM)
@UofT_TCAIREM
Peer perceptions of clinicians using generative #ai in #medical decision-making | @npjDigitalMed
www.nature.com/articles/s41746-025-01901-x
#medtwitter #aimedicine #ai #healthai #aiinhealthcare #medtech -
Tinglong Dai
@TinglongDai
Would you trust a doctor who uses ChatGPT during your visit?
Today, we published a study in @npjDigitalMed. It’s the first survey of practicing clinicians on how they view peers who use generative AI in medical decision-making: www.nature.com/articles/s41746-025-01901-x
The numbers are striking.
In our randomized experiment of 276 clinicians, a physician who used GenAI as the primary tool was rated far lower in clinical skill (3.79 vs 5.93) and overall competence (3.71 vs 5.99) compared to one who did not use GenAI. That’s on a 7-point scale.
Ratings of the care experience also dropped, from 4.48 to 3.08 stars. On a 5-star scale.
Framing GenAI as a tool to verify decisions helped (clinical skill 4.99, competence 4.94) but didn’t close the gap.
Independent of peer perceptions, clinicians saw the potential of using GenAI to improve healthcare: They believed GenAI improves accuracy (4.30) and viewed institution-customized GenAI even more favorably (4.96).
Grateful to be working with a dream team, including Dr. Haiyang Yang @JHUCarey, Drs. Risa Wolf, Nestoras Mathioudakis, and Amy Knight, MD of @HopkinsMedicine, and Dr. @_yuna_ny (my former @JHUCarey @JohnsHopkinsSPH MBA/MPH student, now at @McKinsey).
This work was supported by @JohnsHopkins Discovery Award. A huge thank you to @JHU_HBHI for a pilot grant that led to our Discovery Award. We also appreciate Drs. @jkvedar and @adamgdunn for such a constructive review process.
Read the study (full-text open access): www.nature.com/articles/s41746-025-01901-x -
Tinglong Dai
@TinglongDai
Outer Banks
-
Tinglong Dai reposted
François Chollet
@fchollet
GenAI isn't just a technology; it's an informational pollutant—a pervasive cognitive smog that touches and corrupts every aspect of the Internet. It's not just a productivity tool; it's a kind of digital acid rain, silently eroding the value of all information.
Every image is no longer a glimpse of reality, but a potential vector for synthetic deception. Every article is no longer a unique voice, but a soulless permutation of data, a hollow echo in the digital chamber. This isn't just content creation; it's the flattening of the entire vibrant ecosystem of human expression, transforming a rich tapestry of ideas into a uniform, gray slurry of derivative, algorithmically optimized outputs.
This isn't just innovation; it's the systematic contamination of our data streams, a semantic sludge that clogs the channels of genuine communication and cheapens the value of human thought—leaving us to sift through a digital landfill for a single original idea. -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Paul Graham
@paulg
The Trump administration has suspended the funding of Terence Tao and the Institute for Pure and Applied Mathematics at UCLA.
-
Tinglong Dai reposted
SSRN
@SSRN
Check out the latest from the SSRN #blog which includes a selection of recent #research on #internationaltrade.
Read more: blog.ssrn.com/2025/07/31/the-latest-research-on-internati…
#academictwitter #econtwitter -
Tinglong Dai
@TinglongDai
Osaka at night
-
Tinglong Dai
@TinglongDai
Nara
-
Tinglong Dai
@TinglongDai
Kyoto
-
Tinglong Dai
@TinglongDai
Tokyo
-
Tinglong Dai
@TinglongDai
New York
-
Tinglong Dai reposted
Sam Altman
@sama
agree with lots of what jensen has been saying about ai and jobs; there is a ton of stuff to do in the world.
people will
1) do a lot more than they could do before; ability and expectation will both go up
2) still care very much about other people and what they do
3) still be very driven by creating and being useful to others
for sure jobs will be very different, and maybe the jobs of the future will look like playing games to us today while still being very meaningful to those people of the future. (people of the past might say that about us.)
betting against human's ability to want more stuff, find new ways to play status games, ability to find new methods for creative expression, etc is always a bad bet. maybe human money and machine money will be totally different things, who knows, but we have a LOT of main character energy.
more to come. -
Tinglong Dai reposted
Matthew E. Kahn
@mattkahn1966
A very interesting picture. Europe's low rate of air conditioning adoption has a second major effect. When 500 million affluent people have low aggregate demand for a product, this reduces the innovation incentive for firms to engage in risky R&D (1/2).
-
Tinglong Dai
@TinglongDai
“Instead of differentiating between OR and AI, we need to confuse, blend, and synthesize them as much as possible.”
So urged Herbert A. Simon, the only human being to have received both The @NobelPrize in Economic Sciences and The Turing Award.
That was nearly 40 years ago, during the second AI winter. At the time, Simon used the @INFORMS platform, then known as TIMS/ORSA, to call on the AI and OR communities to work together to advance decision making.
Today, in the midst of an AI boom, we revisited Simon’s vision through a 2024 survey of INFORMS Fellows.
Led by @holly_wiberg (@CarnegieMellon @HeinzCollege), Henry Lam (@Columbia IEOR),@RadhikaSAS (INFORMS President '22), and myself, the project set out to understand how AI and OR/MS intersect in the current landscape, and how the collaboration Simon envisioned can be deepened as both fields continue to evolve.
We are excited to share our findings in INFORMS Journal on Data Science.
Our article, titled "Synergizing Artificial Intelligence and Operations Research: Perspectives from INFORMS Fellows on the Next Frontier," reflects on how Simon’s call for integration remains not only relevant but increasingly urgent. As a Senior Editor of the journal, I am proud to help bring this conversation forward.
We are grateful to Prof. Yu Ding (@GeorgiaTechISyE), Editor-in-Chief, for inviting us to contribute to the journal. We also appreciate @jmc068 and Matt Walls for their help in administering the survey, and @egerstmann for facilitating its distribution to the Fellows. The design and interpretation of the survey were informed by our colleagues on the INFORMS 2024 AI Roadmap Ad Hoc Committee.
Most importantly, we thank the INFORMS Fellows who generously participated in both the pilot and main survey phases. Their insights and experience made this study possible.
In the spirit of Simon: “We need to build our professional institutions and organizations to use [OR and AI] together, supporting, reinforcing, and extending each other.”
"Mix them up!"
Read the full paper here: tinglongdai.com/wp-content/uploads/wiberg-et-al-2025-syne… -
Tinglong Dai reposted
derek guy
@dieworkwear
"So then, why can't those companies produce things in the US?" Well, they can. But the quality will change.
Let's take a few examples. I'll start with South Korean tailoring.
Assisi is a South Korean tailoring shop that's doing some of the best work in the world right now. -
Tinglong Dai
@TinglongDai
Happy 4th of July from Baltimore
-
Tinglong Dai
@TinglongDai
In @Health_Affairs Scholar, @dr_msocal, Maqbool Dada, and I analyze whether tariffs will help move drug manufacturing back to the USA.
Our piece (academic.oup.com/healthaffairsscholar/article/doi/10.1093…) is titled: "Prescription for Made in America? Tariffs and U.S. Drug Manufacturing."
Our view: probably not, or at least not in the way policymakers hope.
Generic drugs make up more than 90% of U.S. prescriptions but account for just 20% of total spending.
This is a high-volume, low-margin industry that relies heavily on global suppliers. More than 92% of generic API facilities serving the U.S. are located overseas. Tariffs are unlikely to reverse that. If anything, they’re more likely to raise prices and worsen the drug shortages we’re already facing.
Branded drugs are a different story. These manufacturers have pricing power and may be better positioned to absorb the costs of relocating production, especially for APIs.
But even here, the incentives are shaped by trade technicalities. A company could produce APIs in the U.S., ship them abroad for final processing, and re-import the finished product tariff-free, as long as it still qualifies as “made in America” under current rules of origin.
Then there’s the issue of policy instability. Tariffs issued by executive order can be reversed just as quickly. That kind of uncertainty makes long-term investment difficult to justify.
If we’re serious about building a stronger, more resilient pharmaceutical supply chain, tariffs won’t be enough. We need stable policy, targeted incentives, and real tools—grants, tax credits, and advance purchase commitments—to make domestic manufacturing viable.
Read the full commentary here: academic.oup.com/healthaffairsscholar/article/doi/10.1093… -
Tinglong Dai reposted
Lawson Mansell
@lawsonhmansell
This comes as a surprise to many people because so much of health costs are not out-of-pocket (aside from drugs), and are instead shrouded behind insurance premiums that eat into wages and increase by 5-7% a year.
Bob Herman
@bobjherman
Hospitals, pharma, doctors, medical devices, health insurers, PBMs — the entire health care industry has crushed working households through higher prices and out-of-pocket spending, to the point that health care is a bigger expense than a mortgage.
This post is unavailable. See it on X. -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
François Chollet
@fchollet
We are now closer to the year 2100 than to 1950.
Also closer to 2050 than to 2000.
Time to start acting like it. -
Tinglong Dai reposted
Rowan Cheung
@rowancheung
Sam Altman called Meta’s poaching “distasteful” in a Slack to his team
—Claimed Meta missed top targets, went “far down their list”
—Warned of “cultural problems” from poaching
—Said salaries are being evaluated, and OAI stock has “much, much more upside”
WIRED
@WIRED
SCOOP: OpenAI CEO Sam Altman is hitting back at Meta CEO Mark Zuckerberg’s recent AI talent poaching spree. "Missionaries will beat mercenaries," Altman wrote in an internal memo to OpenAI researchers.
WIRED has obtained it: www.wired.com/story/sam-altman-meta-ai-talent-poaching-sp… -
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
Johns Hopkins Gets the Most Federal Money, but Now Much of It Is at Risk
"The university’s troubles show how a Republican campaign against higher education could decimate the nation’s research enterprise"
www.nytimes.com/2025/06/26/us/johns-hopkins-university-fe… -
Tinglong Dai
@TinglongDai
“From the sublime to the ridiculous and back again at the turn of every corner I am seduced daily by East London — its dirty and seedy and supremely stylish. Perfect.”
-
Tinglong Dai
@TinglongDai
Cambridge
-
Tinglong Dai reposted
Kevin Roose
@kevinroose
I used to think the "fancy autocomplete" stuff was just goofy cope from NLP dead-enders, but maybe it's actually more calculated than that. There's clearly a market for people who want to be told that AI is a fad and that they don't need to worry about it.
Nathan Benaich
@nathanbenaich
it’s truly nonsensical to say things like this when genai is legit super useful for consumers and businesses - i mean, i get the academic shtick, but it is automating a ton of work already
get over it :) -
Tinglong Dai reposted
npj Digital Medicine
@npjDigitalMed
Are you a student or trainee interested in science communication? Apply for our communications fellowship! You will assist with journal outreach activities including developing an exciting new podcast on digital health.
Applications due July 4th.
www.nature.com/npjdigitalmed/communication-fellowship -
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
Carey Professor @TinglongDai weighs in on the possibility of an American-made smartphone after the Trump Organization announces the launch of a mobile-phone service, Trump Mobile.
www.wsj.com/tech/trump-t1-phone-components-features-2415c7cd -
Tinglong Dai
@TinglongDai
I spoke with @BBC's @lilyjamali about the feasibility of Trump Mobile delivering on its promise to design and manufacture a smartphone in the U.S. by August: www.bbc.com/news/articles/cjrld3erq4eo
My take: it’s not impossible, but without a working prototype, without a high-tech supply chain, and without economies of scale, it would take nothing short of a miracle. -
Tinglong Dai
@TinglongDai
Quoted in @WSJ on Trump’s gold smartphone: $499, iPhone-like specs, “Made in USA”:
“There’s absolutely no way you could make the screen, get that memory, camera, battery, everything” in the U.S., said Tinglong Dai, a professor of operations management and business analytics at Johns Hopkins University’s Carey Business School.
Dai estimated it would take “at least five years” for the U.S. to establish the infrastructure necessary to make “Made in USA” smartphones a real possibility.
He also played down the push to move smartphone manufacturing to the U.S. More immediately, the focus is on building technology such as semiconductors and medical devices, Dai said.
The Wall Street Journal
@WSJ
Trump Mobile is offering a smartphone for $499 with specs that rival the iPhone 16 Pro Max. Here’s what we know, and how plausible it seems. www.wsj.com/tech/trump-t1-phone-components-features-2415c7cd
-
Tinglong Dai reposted
AAMC
@AAMCtoday
When it comes to the medical product supply chain, “resilience means rethinking, not retreating,” says @TinglongDai as he discusses the uncertainty of rising tariffs and the need to participate in global supply networks. www.aamc.org/news/health-care-sector-braces-supply-chain-…
-
Tinglong Dai
@TinglongDai
To @JohnsHopkins Undergrads: A Quick Heads-Up About a Scam
Someone’s been impersonating me, emailing students from a fake address (like dai.jhu.edu@gmail.com) and offering a paid research opportunity. The emails promise remote work, flexible hours, and weekly pay.
One student lost $2,500. The scammer sent fake checks, made it seem like payment was on the way, then convinced the student to send money after those checks bounced. Others have received similar messages.
Just to be clear: I don’t cold-email students about paid research jobs. Never, ever. I’ll never ask for your financial info. And you can safely assume none of your Hopkins professors would either.
If you get something like this, even if it sounds official, pause and double check. Reach out through Hopkins channels or ask someone you trust.
Let’s look out for each other. @JHU_HOUR -
Tinglong Dai reposted
David Perell
@david_perell
Ok, the poet Dana Gioia explained the problem better than I ever could.
This rips: "The failure of the public sector in this nation is embodied in thousands of ugly buildings and public spaces.
These places have been built practically. They are practical and functional in every respect except in practice, since they communicate to the average person that the citizen is just a number in a game of cost efficiency and crowd control.
The experience that Americans have with walking up the steps of the Lincoln Memorial is the experience of beauty, the embodiment of our political vision of the beauty of democracy, expressed in great architecture, great sculpture, great landscaping, and great language, carved calligraphically in the very walls of the memorial.
Just look at a Depression-era post office with marble floors, carved wooden counters, brass fixtures and often an original mural. This was a vision of a beautiful society to which any citizen who entered could participate in.
Today the post office is all vinyl and plexiglass. It offers no vision but expediency. We are not citizens, but customers in a cut rate 99 cent store vision of democracy. No wonder the public doesn't believe in the government. The government seems not to believe in them as alert, intelligent, sensory human beings."
@DanaGioiaPoet -
Tinglong Dai
@TinglongDai
"Could Have"
By Wisława Szymborska
It could have happened.
It had to happen.
It happened earlier. Later.
Nearer. Farther off.
It happened, but not to you.
You were saved because you were the first.
You were saved because you were the last.
Alone. With others.
On the right. The left.
Because it was raining. Because of the shade.
Because the day was sunny.
You were in luck—there was a forest.
You were in luck—there were no trees.
You were in luck—a rake, a hook, a beam, a brake,
a jamb, a turn, a quarter inch, an instant.
You were in luck—just then a straw went floating by.
As a result, because, although, despite.
What would have happened if a hand, a foot,
within an inch, a hairsbreadth from
an unfortunate coincidence.
So you’re here? Still dizzy from another dodge, close shave, reprieve?
One hole in the net and you slipped through?
I couldn’t be more shocked or speechless.
Listen,
how your heart pounds inside me.
- "View With a Grain of Sand," trans. S. Baranczak and C. Cavanagh (New York: Harcourt, Brace, 1996) -
Tinglong Dai reposted
Bearly AI
@bearlyai
Hollywood has strict rules around AI usage but one poll found ~50% of industry artists are secretly using the tool.
Example: costume designer AI-generates concept images then asks an illustrator to redraw them and thus “launder” the AI-generated content into filmmaking process. -
Tinglong Dai
@TinglongDai
Honored to lead an incredible team of @HopkinsMedicine MD students on the first full look at how @US_FDA-cleared medical AI devices are developed and commercialized, now in @NEJM_AI.
We tracked 950 AI medical devices. The results may surprise you. (1/7)
Tinglong Dai
@TinglongDai
Who is building what?
69% of @US_FDA-cleared AI device manufacturers are private, but public firms make more devices per company.
General Radiology leads with 32%, followed by cardiovascular (18%) and neuropsychiatry (15%).
A booming yet scattered market. (2/7)
Tinglong Dai
@TinglongDai
AI Clearance curve
Average @US_FDA clearances jumped from 1.4 devices per year in 1995-2014 to 146 per year in 2020-24—a 100-fold surge.
Total count went from 27 in the first 20 years to 729 in the last five.
In-house development drives nearly all growth. (3/7)
Tinglong Dai
@TinglongDai
Free access link to the full text: ai.nejm.org/stoken/default+domain/2JMKUKZHGQBDGAFDYKS3/fu…
Post not available here, view on X
Tinglong Dai
@TinglongDai
Grateful to Branden Lee, Shivam Patel, @CrystalFavorito, Sara Sandri, and Maria Rain Jennings, first- and second-year @HopkinsMedicine MD students already shaping medical AI.
Thanks also to Drs @charlottehaug and @zakkohane of @NEJM_AI for thoughtful guidance. (6/7)
Tinglong Dai
@TinglongDai
Medical AI is racing ahead; @US_FDA oversight must keep pace.
Read the full @NEJM_AI study here ➡️
ai.nejm.org/stoken/default+domain/2JMKUKZHGQBDGAFDYKS3/fu… (7/7) -
Tinglong Dai reposted
Center for Health Systems and Policy Modeling
@JHU_CHSPM
📢New in @Health_Affairs: Our analysis compares 3 policy options for Medicare site-neutral payment reform, estimating effects on federal spending, hospital types & beneficiary groups: www.healthaffairs.org/doi/full/10.1377/hlthaff.2024.01501
Timely findings as policymakers continue to consider the issue. -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
The Eras Tour
@tswifterastour
🚨| Taylor Swift officially owns all her albums!
"To say this is my greatest dream come true" -
Tinglong Dai reposted
Andrew Ng
@AndrewYNg
I am alarmed by the proposed cuts to U.S. funding for basic research, and the impact this would have for U.S. competitiveness in AI and other areas. Funding research that is openly shared benefits the whole world, but the nation it benefits most is the one where the research is done.
If not for funding for my early work in deep learning from the National Science Foundation (NSF) and Defense Advanced Research Projects Agency (DARPA), which disburse a good deal of U.S. research funding, I would not have discovered lessons about scaling that led me to pitch starting Google Brain to scale up deep learning. I am worried that cuts to funding for basic science will lead the U.S. — and also the world — to miss out on the next set of ideas.
In fact, such funding benefits the U.S. more than any other nation. Scientific research brings the greatest benefit to the country where the work happens because (i) the new knowledge diffuses fastest within that country, and (ii) the process of doing research creates new talent for that nation.
Why does most innovation in generative AI still happen in Silicon Valley? Because two teams based in this area — Google Brain, which invented the transformer network, and OpenAI, which scaled it up — did a lot of the early work. Subsequently, team members moved to other nearby businesses, started competitors, or worked with local universities. Further, local social networks rapidly diffused the knowledge through casual coffee meetings, local conferences, and even children’s play dates, where parents of like-aged kids meet and discuss technical ideas. In this way, the knowledge spread faster within Silicon Valley than to other geographies.
In a similar vein, research done in the U.S. diffuses to others in the U.S. much faster than to other geographic areas. This is particularly true when the research is openly shared through papers and/or open source: If researchers have permission to talk about an idea, they can share much more information, such as tips and tricks for how to really make an algorithm work, more quickly. It also lets others figure out faster who can answer their questions. Diffusion of knowledge created in academic environments is especially fast. Academia tends to be completely open, and students and professors, unlike employees of many companies, have full permission to talk about their work.
Thus funding basic research in the U.S. benefits the U.S. most, and also benefits our allies. It is true that openness benefits our adversaries, too. But as a subcommittee of the U.S. House of Representatives committee on science, space, and technology points out, “... open sharing of fundamental research is [not] without risk. Rather, ... openness in research is so important to competitiveness and security that it warrants the risk that adversaries may benefit from scientific openness as well.”
Further, generative AI is evolving so rapidly that staying on the cutting edge is what’s really critical. For example, the fact that many teams can now train a model with GPT-3.5- or even GPT-4-level capability does not seem to be hurting OpenAI much, which is busy growing its business by developing the cutting-edge o4, Codex, GPT-4.1, and so on. Those who invent a technology get to commercialize it first, and in a fast-moving world, the cutting-edge technology is what’s most valuable. Some studies (link in original post, below) also show how knowledge diffuses locally much faster than globally.
China was decisively behind the U.S. in generative AI when ChatGPT was first launched in 2022. However, China’s tech ecosystem is very open internally, and this has helped it to catch up over the past two years:
- There is ample funding for open academic research in China.
- China’s businesses such as DeepSeek and Alibaba have released cutting-edge, open-weights models. This openness at the corporate level accelerates diffusion of knowledge.
- China’s labor laws make non-compete agreements (which stop an employee from jumping ship to a competitor) relatively hard to enforce, and the work culture supports significant idea sharing among employees of different companies; this has made circulation of ideas relatively efficient.
While there’s also much about China that I would not want the U.S. to emulate, the openness of its tech ecosystem has helped it accelerate.
In 1945, Vannevar Bush’s landmark report “Science, The Endless Frontier” laid down key principles for public funding of U.S. research and talent development. Those principles enabled the U.S. to dominate scientific progress for decades. U.S. federal funding for science created numerous breakthroughs that have benefited the U.S. tremendously, and also the world, while training generations of domestic scientists, as well as immigrants who likewise benefit the U.S.
The good news is that this playbook is now well known. I hope many more nations will imitate it and invest heavily in science and talent. And I hope that, having pioneered this very successful model, the U.S. will not pull back from it by enacting drastic cuts to funding scientific research.
[Original post, with links: www.deeplearning.ai/the-batch/issue-303/ ] -
Tinglong Dai reposted
Kevin Roose
@kevinroose
Between Apple and Meta, has there been a more expensive error in the past 10 years than hiring LLM skeptics to run your bigco AI research program?
Andrew Curran
@AndrewCurran_
Axios is reporting that META's AI division is restructuring into two teams:
- AI Products (AI assistant cross platforms)
- AI Foundations (Llama development)
Yann LeCun's FAIR remains separate from both. -
Tinglong Dai reposted
Johns Hopkins Bloomberg School of Public Health
@JohnsHopkinsSPH
Announcing the next dean of @JohnsHopkins Bloomberg School of Public Health, Keshia Pollack Porter!
A champion for public health, Dr. Pollack Porter has distinguished herself as a leader in advancing policy change that promotes safe, healthy environments. -
Tinglong Dai
@TinglongDai
Antietam
Tinglong Dai
@TinglongDai
-
Tinglong Dai
@TinglongDai
Ann Arbor
Tinglong Dai
@TinglongDai
-
Tinglong Dai
@TinglongDai
To Ride a Water Buffalo: Address to @JohnsHopkins Full-Time MBA Class of 2025
Deans, faculty, staff, families, friends—and most importantly, @JHUCarey Full-Time MBA Class of 2025, Congratulations!
Let me begin with a confession: I still can’t believe you invited me to give this speech. You already took two required courses from me—Operations Management and Artificial Intelligence. You’ve suffered enough. And in 2025, when professors are being called the enemy of the people, you still believed there might be wisdom from me?
Thank you. I’m honored. And deeply moved.
Some of you may remember that in last year’s graduation speech, I spoke about death (www.linkedin.com/pulse/tortured-poets-quants-department-t…).
This year let’s talk about something a bit more cheerful and more lighthearted: destruction.
Yes. Destruction. Not a word you expect the day before graduation. But destruction is what we have—we are living in an age of destruction and dismantling.
Just open your inbox and check the alerts popping up on your phone every hour. The past 120 days have felt like a controlled demolition. If you ever have a moment not following the news, thank God, you’re so blessed.
Things are burning. Institutions are crumbling. Certainties are collapsing.
And yet… I’m here to tell you that destruction is not the enemy—that is, unlike professors.
It is a gift.
Destruction clears. It humbles. It reveals. It forces us to let go of what we thought was permanent—our assumptions, our identities, even our voices. That’s terrifying. But it’s also how we make space for something better.
So today, as we celebrate what you’ve built, I want to talk about what you must be willing to let fall.
Because being a leader in this new world means knowing how to build—but also how to burn.
First, let me tell you a story. I promise you. It’s nothing political. It’s certainly not in the news.
It happened about 2,600 years ago, in ancient China.
There was a philosopher named Laozi. For those who don’t know, Laozi literally means “Old Master.” According to legend, he was born old and never got older. Basically, the Benjamin Button of ancient China—except he skipped the Brad Pitt looks and went straight to the wisdom.
Laozi was considered the wisest man in the East. Even Confucius—yes, that Confucius—once sought his advice.
But as time passed, Laozi grew tired. Not just tired—but soul-tired. Tired of the noise, the king’s corruption, incompetence dressed as brilliance, and sheer absurdity of it all—the oligarchies.
So he decided to emigrate—no papers, no documentation.
He simply packed up, got on a water buffalo, and rode west. He was letting go.
He and his water buffalo reached the border—a dusty border outpost—and an immigration officer stopped him.
Fortunately, it was the 6th century BCE. No smartphones. No iMessages or Signal group chats to search.
The officer could have blocked his exit. Instead, he made a request:
“Before you go,” he said, “please write down your thoughts.”
Poor Laozi—what else was he supposed to do? So he got off the water buffalo, sat down, and in a single sitting wrote the Tao Te Ching—just 5,000 words that would become the very beginning of Taoist philosophy. A masterpiece read and reread by millions to this day, myself included.
And then he vanished.
I love this story for many reasons. But mostly because it reminds us that one of the most important books in philosophy wasn’t created by AI or productivity software—it was made possible by an immigration officer. And you thought your experiences at the DMV were life-changing.
Many people in this room, myself included, are immigrants. If you are not, chances are, your parents or grandparents or spouses were.
For immigrants, borders are where things end—and where things begin. Laozi’s old world was over. So was ours. Every story of destruction begins with a crossing.
As Taylor Swift once said—well, sang: "Everybody here was someone else before."
This, too, is the essence of the United States of America. You can be anybody. Even in 2025. You’re allowed to become someone new. And that began with the quiet destruction of who you were before.
Now let me tell you another story. This one’s more personal.
I crossed many borders in my life. Some literal. Some linguistic. Some inside myself.
As many of you know, I grew up in a small village in China. Speaking properly has been a lifelong struggle—starting from the moment I began to talk.
For reasons I still don’t fully understand—maybe I spent too much time reading or watching news?—I never quite mastered the local dialect. Villagers would laugh and say I must be trying to speak Mandarin—that I sounded like a city kid.
Then I moved to Shanghai—and discovered I didn’t speak Mandarin properly either. People could tell where I was from the instant I opened my mouth.
Basically, I was linguistically homeless: too foreign for the village, and too village for the city.
A few people didn’t mind my accent. One of them became my best friend. And my wife. The best gift God ever gave me.
Later, I moved to Pittsburgh for graduate school. I was surrounded by native speakers and international peers who spoke English more fluently than I ever could. I felt small. Really small.
And when I was nearing my PhD, a very wise professor told me: “Based on your speaking skills, you will never get a job in a business school.”
And he was absolutely right. I was rejected everywhere.
One Ivy League school said, and I quote: “You’re an excellent researcher, but we can’t allow you to interact with our MBA students.”
He was right—until Johns Hopkins and Carey gave me a chance.
The one school that didn’t reject me became the place I was destined to be.
Five months after moving to Maryland, I was asked to teach MBA students—despite having once been labeled, in essence, untouchable.
That first day, I walked into the classroom trembling. Literally trembling. I feared silence. I feared someone raising their hand and asking, “Professor, what do you mean?”
But no one did.
Instead, they laughed at every joke. They leaned in. They treated me like I belonged.
That was the moment.
The old self cracked—the one who feared being misunderstood. The one who thought his accent disqualified him from being heard.
That self fell apart. And something new—and far more dangerous—emerged: a voice.
That destruction made me a teacher. And I began to realize—sometimes what you’re struggling with says more about the world than it does about you.
What seems like rejection is often a reflection of the limits of others, not your own.
And an accent, a blind spot, an obsession might actually be the very thing that filters your path, that points you to the people and places that truly see you.
The one person who accepted me for all I was became my wife. The one school that didn’t reject me became my home.
As Mark Twain said, "Keep away from people who try to belittle your ambitions. Small people always do that, but the really great make you feel that you, too, can become great."
Destruction isn’t a detour. It’s the path. That realization becomes a powerful filter. It lets you see what must be dismantled. And it gives you permission to destroy what never deserved to stay.
We don’t fear destruction because it’s violent.
We fear it because it can sometimes be honest.
Destruction doesn’t lie. It tells you exactly what’s no longer working—what’s already broken, what’s already gone.
But instead of listening, we preserve the illusion. We polish the cracks. We promote the unpromotable.
In business, we say “disruption.”
But let’s be honest: disruption is just destruction that learned how to pitch. It’s destruction in a turtleneck—ask Steve Jobs. Or these days, a leather jacket—ask Jensen Huang.
Every industry collapse you’ve studied began with someone clinging to an old story:
“This is how it’s always been.”
“This is who we are.”
“This is what people want.”
Remember elevator operators? That used to be a whole profession. A person—sometimes two—whose job was to press the button you just told them to press.
Dignified, uniformed, essential. Until they weren’t.
Joseph Schumpeter called it “creative destruction:” the idea that capitalism renews itself by tearing down the old to make way for the new.
And you are the new.
But most people hear “creative” and forget about the “destruction” part. Because that’s the part that’s messy. Uncomfortable. Sometimes deeply personal.
But here’s the truth: destruction is not the villain. It’s the turning point. The hard edit. The plot twist that makes room for the next act.
And that includes your own story.
Your old self, the one who took safe jobs, stayed quiet in meetings, updated LinkedIn like it was a legal obligation—that version will fall away. Not because the world is cruel. But because you are becoming someone else.
Because you are more powerful than you let yourself imagine.
Marianne Williamson—regardless of your feelings about her or her politics—said something that’s truly worth reading, and worth remembering:
“Our deepest fear is not that we are inadequate. / Our deepest fear is that we are powerful beyond measure. / It is our light, not our darkness, that most frightens us. / We ask ourselves, ‘Who am I to be brilliant, gorgeous, talented, fabulous?’ / Actually, who are you not to be? / You are a child of God. / Your playing small does not serve the world. / There’s nothing enlightened about shrinking.”
We are afraid to destroy our smaller selves.
But we must.
Please—destroy them. Make room.
I want to end this speech by offering a few reflections on how to live in dark times.
This is a dark moment. The tide is low. There is absolutely no doubt about that.
Martin Luther King Jr. said: "Only when it’s dark enough can you see the stars."
And as Warren Buffett put it: "Only when the tide goes out do you discover who’s been swimming naked."
Now is the time to choose the shape of your fire.
Because make no mistake—you will set something ablaze.
Some of you will dismantle the forces that trap our healthcare system in inertia.
Some will rewrite how AI is integrated into human workflows.
Some will reject false trade-offs between profit and justice.
Some will say no to yesterday’s idols—Elon Musk, Peter Thiel, Marc Andreessen, and alike—and choose to lead with a different mindset, and a different moral compass.
Some will take down public health supply chains that treat safety and resilience as an afterthought.
And some of you will take aim at the distractions of our age: The false prophets of progress. The noisy visionaries. The constant whisper that you’re not enough.
You are not here to preserve what is.
You are here to dismantle what no longer works.
That is the calling of every generation of builders.
And that is what you are now: builders.
But building begins with burning.
Let what needs to fall—fall.
And from the ashes: speak. Build. Rise.
Class of 2025:
You have fire in your bellies—now use it.
Go change the world.
Thank you. Congratulations! -
Tinglong Dai reposted
Antonio Regalado
@antonioregalado
Per Boston Globe, ALL Harvard Medical School grants from federal government are cancelled.
Here's the pie chart of their budget.
The cancellations affect everything, from biostatistics department to longevity researcher David Sinclair.
Story from Boston Globe in reply. -
Tinglong Dai
@TinglongDai
Greetings from Banff
-
Tinglong Dai
@TinglongDai
U.S. dominance of Nobel Prizes peaked in the 1990s (via @auren’s five links):
-
Tinglong Dai
@TinglongDai
Landing in Boston
-
Tinglong Dai
@TinglongDai
Thrilled to share our new working paper: Tariffs as a Hidden Tax: Price Pass-Through in Multi-Stage Supply Chains, now on @SSRN.
What happens when a tariff hits the first step in a multi-stage supply chain—like in pharmaceuticals? How much of that cost reaches the consumer?
Working with Maqbool Dada, @dr_msocal, and Yuxiang Sun, we built a tariff impact model to follow the path of a tariff through a multi-stage chain. We found that even large tariffs often lead to only small price increases.
That’s because costs get absorbed along the way—by manufacturers, distributors, and retailers—and because insurance and government programs often shield the final buyer.
We also ask: Can tariffs bring production back home? Often, they raise costs without shifting incentives:
Tinglong Dai
@TinglongDai
Link to the full-text: doi.org/10.2139/ssrn.5237643
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai
@TinglongDai
At the start of the pandemic, Germany was still using fax to send emergency health data.
Five years later, it’s now supporting AI research at national scale—while keeping data decentralized and fully privacy-compliant. Researchers can work with records from 75 million people without needing individual consent, because the data stays local and secure.
In @npjDigitalMed, @colitiscrohn and @jkvedar explain how Germany and the EU are making federated, privacy-compliant health data access a reality.
This shift matters far beyond Europe. As Drs. Baumgart and Kvedar wrote, it offers “unprecedented opportunities to advance digital health innovation and research with global impact.”
Medical AI is one of the few areas in tech that isn’t dominated by Silicon Valley—and isn’t U.S.-centric. From Canada’s national AI strategy to the EU’s Health Data Space, we’re seeing more balanced, global development.
If the U.S. wants to stay competitive, it needs better access to representative data—and a serious plan for trust, governance, and interoperability.
Read the editorial: www.nature.com/articles/s41746-025-01631-0 -
Tinglong Dai
@TinglongDai
Honestly feel lucky to be living in the Larry David era:
Tinglong Dai
@TinglongDai
Link: www.nytimes.com/2025/04/21/opinion/larry-david-hitler-din…
-
Tinglong Dai reposted
Andrew Curran
@AndrewCurran_
Interesting legal argument from META; the use of a single book for pretraining boosts model performance by 'less than 0.06%.' Therefore, taken individually, a work has no economic value as training data.
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Anthony Pompliano 🌪
@APompliano
12% of all private sector jobs in NYC are people staying home to take care of their own relatives.
That is insane. -
Tinglong Dai reposted
Yisong Yue
@yisongyue
One of my PhD students got their visa revoked. I know of other cases amongst my AI colleagues. This is not what investing in US leadership in AI looks like.
Antoine Levy
@LevyAntoine
This is flying a bit under the radar.
But in terms of damage to America’s innovation and knowledge supremacy, the chilling effect of these revocations on the country’s ability to attract and retain scientific talent likely dwarfs the impact of tariffs or other policies. -
Tinglong Dai reposted
Sahil Kapur
@sahilkapur
“We are all afraid,” Murkowski said, taking a long pause. “It’s quite a statement. But we are in a time and a place where I certainly have not been here before. And I’ll tell ya, I’m oftentimes very anxious myself about using my voice, because retaliation is real. And that’s not right.” www.adn.com/politics/2025/04/14/we-are-all-afraid-speakin…
-
Tinglong Dai reposted
Lawrence H. Summers
@LHSummers
The @realDonaldTrump Administration, with its almost dictatorial demands, has brought more unity to @Harvard faculty than there has been during the 21st century. Harvard was contributing to America before Trump and will be contributing long after Trump.
-
Tinglong Dai reposted
Lawrence H. Summers
@LHSummers
Very glad to see President Garber leading Harvard and I hope all universities in resisting extralegal and unreasonable demands from the federal government.
www.thecrimson.com/article/2025/4/15/harvard-denies-trump… -
Tinglong Dai
@TinglongDai
Kudos to @Harvard.
Harvard stood up. Now others must too.
To every American college and university:
Silence is no longer an option.
Kowtowing is no longer enough.
As Ecclesiastes says:
“There is a time to be silent, and a time to speak.”
This is the time to speak.
This is the time to show that humanity matters—
when empathy overcomes cruelty, authoritarian instinct, and willful ignorance.
The time to show that higher education will not be bullied into submission.
The time to protect human dignity.
Because if they can strong-arm Harvard—
what hope is there for anyone else?
Harvard University
@Harvard
The university will not surrender its independence or relinquish its constitutional rights. Neither Harvard nor any other private university can allow itself to be taken over by the federal government. hrvd.me/ResearchFunding25t
-
Tinglong Dai reposted
Saturday Night Live
@nbcsnl
The White Potus
This media couldn't be displayed. See it on the original post. -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Pamela Johnson
@PamelaJohnsonMD
The Academy for Transformative Health Care Leadership is a groundbreaking new leadership development program combining industry-leading business education with world class health care leadership expertise. A collaboration between the @HopkinsMedicine & @JHUCarey, this six-month academy is aimed at transforming our approach towards the health care industry by examining systems-level thinking, policy and finance, change leadership, and executive communication. If you are in the health care space and are looking to develop skills needed to excel in today’s environment while building for tomorrow’s advancements, join us in the inaugural class! More information: carey.jhu.edu/programs/executive-education/programs-indiv…
@JohnsHopkins @healthecon_dan @ChrisGMyers @TinglongDai @JHU_HBHI -
Tinglong Dai
@TinglongDai
Pittsburgh
-
Tinglong Dai
@TinglongDai
.@awmcmu on why AI can't replace OR in his keynote @PittTweet:
"It’s very hard to make sense of real progress today. There’s so much hype. General-purpose LLMs, like GPT-4, are still very stupid when it comes to tasks like 'replan evacuation routes after a disaster.'
"You don’t want a chatbot here. You want a constraint optimization engine. But when executives get on stage and showcase AI products, people get dazzled. It’s easy to forget that 97% accuracy is not good enough. You might need 99.9999%." @INFORMS -
Tinglong Dai reposted
Ryan Petersen
@typesfast
Ocean carriers have announced that to reduce the fees they will skip the smaller ports like Seattle, Oakland, Boston, Mobile, Baltimore, New Orleans, etc.
Some carriers have said they'll just move the capacity serving the U.S. to other trade lanes altogether. /2 -
Tinglong Dai reposted
John Lettieri
@LettieriDC
Peter Navarro said something quite revealing on CNBC today about BMW's factory in South Carolina: "That doesn't work for America. It's bad for our economics, it's bad for our national security."
This piqued my interest. I grew up very close to that BMW facility -- the company's largest in the world -- and I watched it transform my state and region with more than $13 billion invested in since 1992, 11,000 jobs onsite at an 8 million square foot campus, $10+ billion in exports in 2023 alone, etc. One study put the annual economic impact of the plant at $27 billion. In a small state, these are absolutely game-changing numbers.
For all these reasons, BMW's investment has been hailed as a turning point in the state's economic history. Earlier this year, for example, Gov. Henry McMaster said this: "BMW's arrival in South Carolina over 30 years ago transformed our economy and global reputation." He's right. The investment was instrumental in helping the state transition from a textile-dependent economy for most of the 20th century to one with a vibrant advanced manufacturing industry today.
In light of all this, how in the world does Navarro conclude that BMW's investment "doesn't work for America"? I mean, even trade restrictionists have long praised this kind of inbound foreign direct investment in U.S. auto manufacturing. Not Navarro. He's setting a new bar: whatever benefits BMW has delivered are more than canceled out by the fact that it sources parts from a global supply chain. And this is the vision behind the new tariff agenda. -
Tinglong Dai
@TinglongDai
State-by-state Trump approval ratings @Civiqs
-
Tinglong Dai
@TinglongDai
Quoted in @Nature on how Trump’s tariffs impact the supply chain for science: www.nature.com/articles/d41586-025-01060-9
As I told @celestebiever, these tariffs shape all corners and all directions of global trade — and, with that, the U.S.’s standing as both a top importer and exporter.
At stake is everything from basic labware to advanced instruments, landing just as research institutions are already under acute financial stress.
We're talking about the core infrastructure of modern science.
This isn’t just about belt-tightening. It could be the last straw — and risks causing lasting damage.
The U.S. imports billions of dollars in lab equipment and reagents each year — and many of those products, or the components inside them, now face tariffs of 10–54%. Even a DNA sequencer built in California might still depend on optics from Germany or semiconductors from China. Some high-end tools, like precision microscopes, aren’t made in the U.S. at all.
We’re also a top exporter. The U.S. leads the world in exports of diagnostic and lab reagents. If this turns into a prolonged tit-for-tat, the entire scientific supply chain — here and abroad — becomes more fragile, more expensive, and less reliable.
We’re really just at the front end of what could become a tariffs-induced period of chaos.
@JohnsHopkins @INFORMS @JHUCarey @HopkinsMedicine @Jillrosen -
Tinglong Dai reposted
Celeste Biever
@celestebiever
Microscopes, test tubes and other research supplies are about to get more expensive thanks to Trump tariffs. “We're already doing quotes today that are 20% more than they were yesterday” www.nature.com/articles/d41586-025-01060-9
-
Tinglong Dai
@TinglongDai
A Letter of Apology from the Penguins
April 3, 2025
To His Majesty, the Tarriff King:
We apologize
for never making cars,
for not growing your roses,
for failing to ship things
you might wish to tax.
Forgive us:
our exports are mostly innocence and awe.
Unworthy of a place on your impeccable Chart.
We regret to inform you,
our economy runs on wild-caught fish
and the occasional cringe-worthy courtship display.
We did not mean to evade
your holy tariffs.
It’s just—we have nothing you’d want,
except perhaps
this continent’s last good silence.
Apologies for crossing your trade lines
without paperwork.
Our passports were lost in the last storm.
Our volcano—
he erupts when he pleases,
takes no Zoom calls,
won’t join your Signal group.
We have no lobbyists.
Woefully unable to negotiate.
Only this letter,
beside the sea.
We remain,
your Penguins, faithful
to whatever empire
inconvenienced
by our beauty. -
Tinglong Dai reposted
Michael D Abramoff
@MichaelAbramoff
On behalf of the leadership of the Collaborative Community for Ophthalmic Innovation - CCOI, I am excited to announce that Dr Malvina Eydelman, until recently Director of FDA's Office of Health Technology-1 has accepted our offer of becoming the CEO of CCOI.
This is awesome news for medical innovation, for AI, for novel therapies, in ophthalmology and beyond, as we continue to expand CCOI's open, transparent, science based consensus building on how we can bring trusted innovation to patients.
Press release below: -
Tinglong Dai reposted
Max Bayer
@maxonwifi
I think BIO CEO John Crowley ultimately had the most accurate portrayal of what today would be like.
Max Bayer
@maxonwifi
a good snippet of my interview style hah
-
Tinglong Dai reposted
Ryan Petersen
@typesfast
The Taiwan tariffs at 32% are massive BUT there is a carve out that they don’t apply to semiconductors.
-
Tinglong Dai reposted
Mark Zandi
@Markzandi
I’m raising my odds that a recession will begin sometime this year to 40%, up from 15% at the start of the year. Last week’s economic data were disconcerting, including the slide in consumer confidence, punk consumer spending, and persistently high inflation. The intensifying trade war and DOGE cuts are behind all this and with last week’s announcement of big tariff increases on vehicle imports and the coming reciprocal tariffs, things are sure to get worse. Recession remains less likely than not only because layoffs remain low and job and income growth positive. This Friday’s jobs report for March will give us a sense of whether this continues. It is premature to expect much fallout from the trade war and DOGE cuts in the jobs data, suggesting a monthly payroll job gain of close to 150,000. Anything south of 100k would be worrisome, and anything north of 200k would be welcome. But whatever the job number, as long as the tariffs and DOGE cuts continue to mount, so too will the odds of recession.
-
Tinglong Dai reposted
SSRN
@SSRN
As of June 2024, the U.S. FDA has approved 950 medical #ai devices, but the current regulatory #framework requires new submissions for any updates, which limits the algorithms' ability to learn from new #data.
Read: papers.ssrn.com/sol3/papers.cfm?abstract_id=5009572
Subscribe: www.ssrn.com/index.cfm/en/singapore-management-u-res/ -
Tinglong Dai reposted
Amy Klobuchar
@amyklobuchar
Unbelievably dangerous and thank God people are safe.
My first call to Department of Defense tomorrow: why are your planes flying 500 feet below passenger jets full of Minnesotans headed from DCA to my state. www.cnn.com/2025/03/28/us/delta-military-jet-close-call-d… -
Tinglong Dai reposted
Catharine
@DrCatharineY
Down over 60%. That’s how far federal research funding has fallen compared to this time last year. Every delay in funding is a delay in delivering for the health of American people.
-
Tinglong Dai
@TinglongDai
Can we trust AI to deliver health care? And can AI trust humans?
So thrilled to share our perspective in @Nature_NPJ Health Systems, written with the amazing @madelinesagona, @Mario_Macis, and Michael Darden.
The idea came from something we kept seeing—every panel on AI and health care eventually turns into a conversation about trust.
But the word “trust” gets thrown around without much clarity.
So we asked: trust by whom, in whom, and for what?
At its core, trust is about navigating uncertainty. It means being vulnerable—relying on others with the hope that they’ll act competently and in good faith.
In the paper, we break it down into several layers.
For patients, trust allows them to accept that vulnerability in exchange for care. For providers, it gives them the confidence to make decisions, often in complex or high-stakes situations. And for institutions, trust shapes how they engage with the communities they serve.
Now, with AI entering the picture, this web of trust gets even more layered. Patients need to trust their doctors to use AI wisely. Doctors need to trust that AI supports—not replaces—their judgment. Both patients and providers have to trust the health system to use these tools in ways that truly benefit care.
Even AI needs to trust us—on good data, thoughtful oversight, and ethical design.
Trust is the glue holding everything together.
We offer a simple framework to help researchers and practitioners think more clearly about trust in an AI-enabled world.
Huge thank you to @JHU_HBHI for bringing all of us together—and to Editor-in-Chief, Dr. @YiyeZhang1 for the kind invitation and constructive peer review process.
And congratulations to @madelinesagon, @JHU_HBHI Graduate Academy Fellow and Genetic Epidemiology ScM Candidate at @JohnsHopkinsSPH, on an outstanding first peer-reviewed publication! So awesome to work with you!
Tinglong Dai
@TinglongDai
Link to our article: www.nature.com/articles/s44401-025-00016-5
-
Tinglong Dai reposted
Johns Hopkins University
@JohnsHopkins
A few weeks ago, hundreds of Johns Hopkins affiliates stood side by side in Washington, D.C. to stand up for science—because research isn’t just knowledge, it’s progress. As federal cuts threaten the future of discovery, JHU voices like Bonnielin Swenor, Allison Agwu, and Andrew Eneim reminded us: science saves lives, and it needs our support. hub.jhu.edu/2025/03/10/stand-up-for-science-2025/ #researchsaveslives
-
Tinglong Dai
@TinglongDai
Baltimore, Studio-Ghibli style
-
Tinglong Dai
@TinglongDai
Get $1 billion in 3 easy steps:
Step 1. Buy Family Dollar for $9 billion.
Step 2. Hold onto it for 10 years.
Step 3. Sell it for $1 billion. -
Tinglong Dai reposted
The Baltimore Banner
@BaltimoreBanner
Baltimore is home to the busiest car port in the world’s top car-importing nation.
Mexico accounts for 29% of Baltimore’s auto imports, more than Germany’s 22% and Japan’s 20%.
From Dec. 2024:
www.thebaltimorebanner.com/economy/donald-trump-tariffs-p…
Hayes Gardner
@HayesGardner
President Donald Trump has announced a 25% tariff on cars. That could have a substantial effect on the Port of Baltimore, which (with the exception of 2024) has long been the nation's top automobile port.
-
Tinglong Dai reposted
Braydon Dymm, MD
@BraydonDymm
I really appreciate these good/evil robot graphics
npj Digital Medicine
@npjDigitalMed
The U.S. is the first country in the world to propose legislation that would allow AI to autonomously prescribe medications.
@TinglongDai
www.nature.com/articles/s41746-025-01540-2 -
Tinglong Dai
@TinglongDai
Tampa
-
Tinglong Dai reposted
Aamir Javaid, MD
@aamirjavaidmd
Interesting article about uncertainty in FDA regulatory environment regarding AI @TinglongDai
npj Digital Medicine
@npjDigitalMed
The U.S. is the first country in the world to propose legislation that would allow AI to autonomously prescribe medications.
@TinglongDai
www.nature.com/articles/s41746-025-01540-2 -
Tinglong Dai
@TinglongDai
Miami
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
npj Digital Medicine
@npjDigitalMed
The U.S. is the first country in the world to propose legislation that would allow AI to autonomously prescribe medications.
@TinglongDai
www.nature.com/articles/s41746-025-01540-2 -
Tinglong Dai reposted
Siyuan (Steven) Wang
@SStevenWang
Is Johns Hopkins University the first university that used its official account to report the rally? My respect to the leadership! These are brave people at JHU.
Johns Hopkins University
@JohnsHopkins
Scientists, researchers, and advocates recently gathered in Washington, D.C. for the Stand Up for Science rally. Voices from Johns Hopkins and across the globe spoke on the importance of continued investment in scientific progress. hub.jhu.edu/2025/03/10/stand-up-for-science-2025/
-
Tinglong Dai
@TinglongDai
Tariffs
A thousand percent on steel, on wheat,
on candles, on fireworks,
on anything that dares to arrive.
A child plays LEGO,
brick by brick,
a castle, a ship, a dream—
still untaxed, for now.
Lovers whisper across borders,
their words slip through unseen.
No duties on longing,
but April 2nd is near.
And what of free souls?
Will there be a price to move,
a surcharge to be?
Will the wind, the tide, the clouds
owe something too?
A thousand percent on everything.
But the moon still rises,
the eagle still soars,
and the mind still wanders.
While the old man sleeps,
his hands huge,
his joy tremendous,
his nose air-stamped for each snore.
Oh, we shall all be so rich. -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai
@TinglongDai
Happy to share my paper "Artificial Intelligence on Call: The Physician’s Decision of Whether to Use AI in Clinical Practice" (with Prof. Shubhranshu Singh, also @JHUCarey) has been accepted for publication in Journal of Marketing Research @ama_journals, a premier journal in marketing.
Back in 2019, we sensed a quiet revolution underway in medicine. We began exploring a question: When and how do people actually use medical AI?
Over the past six years, one thing has become clear: physicians—not hospitals—ultimately make the "when and how" decisions. And liability concerns are central to those decisions.
Current laws may inadvertently discourage physicians from using—and following—AI, especially for medical decisions where AI could be most helpful.
Better AI alone won't solve this. As AI improves, physicians may become even more concerned about liability if they don't follow its recommendations.
We are grateful to the Co-Editors, @brett__gordon and Vikas Mittal, the Associate Editor, P. K. Kannan, and three anonymous reviewers for their valuable feedback, which significantly improved the paper.
A huge thank you to everyone who has made comments and suggestions over the years that have helped make the paper what it is today.
Tinglong Dai
@TinglongDai
Link to full text PDF: tinglongdai.com/wp-content/uploads/Dai-Singh-2021-AI-on-C…
-
Tinglong Dai reposted
Lawrence H. Summers
@LHSummers
I think we've seen a sea change in perception in the almost two months since President Trump was inaugurated.
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Selina Wang
@selinawangtv
More layoffs coming for federal workers-White House ordering all federal agencies to submit plans for large-scale layoffs by March 13
Pentagon preparing for major layoffs – one Pentagon worker who is a veteran says he fears it will impact the military’s readiness @ABCThis media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
G
@stevensongs
Live from the Oval Office
-
Tinglong Dai reposted
Alex Chan
@alex8chan
“Put (in) the hours. Unless you put a lot of effort, hours, sometimes sleepless nights. Because you are passionate about something.”
Love that @DAcemogluMIT emphasized this. As a grad student, I tried to work no less than my own advisors, and then more.
m.youtube.com/watch?v=nqeFL9_0L_c -
Tinglong Dai reposted
Tinglong Dai
@TinglongDai
Peggy Noonan on J.D. Vance, back in Feb. 2022:
-
Tinglong Dai reposted
The Baltimore Sun
@baltimoresun
Tariffs threaten a host of unintended consequences, writes Tinglong Dai and Christopher S. Tang. www.baltimoresun.com/2025/02/24/how-tariffs-can-hurt-amer…
-
Tinglong Dai reposted
nature
@Nature
About a month after Donald Trump took office as the 47th US president, almost all grant-review meetings remain suspended at the US National Institutes of Health (NIH), preventing the world’s largest public funder of biomedical research from spending much of its US$47 billion annual budget.
go.nature.com/4k8Njjx -
Tinglong Dai reposted
Johns Hopkins University
@JohnsHopkins
If you follow @JohnsHopkins, you know that university research is about more than data—it’s about people. Research saves lives by bringing new treatments, cures, and hope to families everywhere. From transformative medical breakthroughs to innovations that shape tomorrow, the impact is real. Every future discovery starts with the work happening now. Share to show your support. Share so we all understand what’s at stake. Share because research saves lives.
-
Tinglong Dai reposted
Susan Simpson
@TheViewFromLL2
They did, yeah.
-
Tinglong Dai reposted
Jamie Dupree
@jamiedupree
Not sure what's going on here - but the documents shown by DOGE and cited earlier for this $8 million in ICE savings has now been *changed* to $8 billion in the Federal Procurement Data System. 🧵
Jamie Dupree
@jamiedupree
The DOGE team is finally posting examples of where they have cut spending. This one claims $8 billion in savings (bottom figure) - but if you look at the fine print upper right, it's a contract that's instead valued at $8 million.
-
Tinglong Dai reposted
Jamie Dupree
@jamiedupree
The DOGE team is finally posting examples of where they have cut spending. This one claims $8 billion in savings (bottom figure) - but if you look at the fine print upper right, it's a contract that's instead valued at $8 million.
-
Tinglong Dai reposted
Farzad Mostashari
@Farzad_MD
16/ When you hear, "the CDC's Epidemic Intelligence Service class was cancelled and the officers fired" I hope this gives you a sense of what has been lost.
Not just the outbreaks that can't be investigated, the surge capacity gone, but our future public health leadership lost -
Tinglong Dai reposted
Steven Salzberg 💙💛
@StevenSalzberg1
98% of NIH grants that should have gone out this month, didn't. Despite court orders 'unfreezing' the funds, they are frozen. Biomedical research will grind to a halt. Clinical trials will end. All for no reason: the total impact on the US budget is miniscule
-
Tinglong Dai reposted
Jason Shepherd
@JasonSynaptic
Let’s be real. These attacks on science and the NIH are not about efficiency in spending. They are about curbing DEI and taking revenge on institutions for conspiracy level blame for COVID origins etc. JD Vance said this. Project 2025 said this. It makes no economic sense.
-
Tinglong Dai reposted
Ashish K. Jha
@ashishkjha
This is a very odd tweet and super disappointing to see from the esteemed @NIH
The top 7 university recipients of nih funding do NOT include Harvard or Yale
The top 7 are:
Johns Hopkins
UCSF
U of Pennsylvania
Duke
Michigan
Pittsburgh
Washington U in St. Louis
NIH
@NIH
Last year, $9B of the $35B that the National Institutes of Health (NIH) granted for research was used for administrative overhead, what is known as “indirect costs.” Today, NIH lowered the maximum indirect cost rate research institutions can charge the government to 15%, above what many major foundations allow and much lower than the 60%+ that some institutions charge the government today. This change will save more than $4B a year effective immediately.
-
Tinglong Dai reposted
Jonathan Cheng
@JChengWSJ
Yun Sun in Foreign Affairs: “Beijing assumes that Washington’s own policies will dismantle the foundations of U.S. global hegemony, even if it creates a lot of turbulence…in the process. China’s top priority, then, is simply to weather the storm.”
www.foreignaffairs.com/china/chinas-trump-strategy -
Tinglong Dai reposted
James Surowiecki
@JamesSurowiecki
NIH is legally prohibited by the Further Continuing Appropriations Act of 2024 from changing indirect-cost rates across the board without the approval of Congress. This will be immediately challenged in court.
NIH
@NIH
Last year, $9B of the $35B that the National Institutes of Health (NIH) granted for research was used for administrative overhead, what is known as “indirect costs.” Today, NIH lowered the maximum indirect cost rate research institutions can charge the government to 15%, above what many major foundations allow and much lower than the 60%+ that some institutions charge the government today. This change will save more than $4B a year effective immediately.
-
Tinglong Dai
@TinglongDai
Save a penny, lose a dollar.
Nothing weakens America more—or strengthens our adversaries faster—than shortsighted moves like this.
The U.S. dominates global research because we invest in it. Cut funding, and we cut our future.
Slashing indirect costs won’t make research cheaper—it will make it impossible.
The world’s best institutions fuel innovation, and we should be proud to invest in them, not undercut them.
POTUS and Co-President, people are scared of you. If world domination is the goal, you’ve succeeded.
But how much strength can this nation afford to squander? -
Tinglong Dai reposted
Scott Kominers
@skominers
In addition to the inanity of blanket-banning research on these topics, if they're really doing this via string matching, they're also going to throw out econometrics ("biased" estimators), biology (blood-brain "barrier") and a whole bunch of math (Jensen's "inequality") 🤦♂️🤦♂️🤦♂️
Benjamin Ryan
@benryanwriter
NEWS: The following words are officially cause for a federal grant to be rejected, according to @DarbySaxbe, a professor at @USC, who learned this from a program officer at the National Science Foundation.
“This is a crisis for academic freedom & science,” said Dr. Saxbe. -
Tinglong Dai reposted
Hanno Lustig
@HannoLustig
Just had a good phone call with Trader Joe's. I'm going to keep shopping there for now.
Hanno Lustig
@HannoLustig
To reduce my bilateral trade deficit with Trader Joe's, I am now going to shop at Whole Foods, which is more expensive and a bit further way. Thanks all for your advice.
-
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
Justin Trudeau
@JustinTrudeau
I just had a good call with President Trump. Canada is implementing our $1.3 billion border plan — reinforcing the border with new choppers, technology and personnel, enhanced coordination with our American partners, and increased resources to stop the flow of fentanyl. Nearly 10,000 frontline personnel are and will be working on protecting the border.
In addition, Canada is making new commitments to appoint a Fentanyl Czar, we will list cartels as terrorists, ensure 24/7 eyes on the border, launch a Canada- U.S. Joint Strike Force to combat organized crime, fentanyl and money laundering. I have also signed a new intelligence directive on organized crime and fentanyl and we will be backing it with $200 million.
Proposed tariffs will be paused for at least 30 days while we work together. -
Tinglong Dai reposted
Lawrence H. Summers
@LHSummers
.@realDonaldTrump's tariffs are a bully strategy. Bullying doesn’t win over time on the playground or in the international arena. This self-inflicted supply shock is a strategic gift to Xi Jinping. 1/4
-
Tinglong Dai reposted
Peggy Wang
@peggy_wang
I'm saddened, shocked, angry, and appalled that this assault happened to my cofounder in SF. Luckily he was discharged from the hospital last week and is expected to recover, although the full extent of the injury and recovery process will take much longer.
It's absolutely unacceptable that the city and government of San Francisco have utterly failed to press charges or arrest the criminal that attacked my cofounder.
I've been trying to get more information about the attack from @SFPD while @simulacronist was in the hospital, but the police have refused to get me any information.
This is not a good look for the city that is at the forefront and center of the innovation in tech and AI.
𓏴𓀿𓁼𓀐𓏴
@404vish
i was recently discharged from the icu at the sf hospital after traumatic brain injury due to a man hitting me on the back of my head with a metal pipe, completely unprovoked, in the mission in san francisco. this man is known to the police since he’s done this before to other people. afaik they will not be pressing charges or prosecuting this individual due to the state of sf politics.
-
Tinglong Dai
@TinglongDai
Bravo, @WSJ editorial board
-
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
Hopkins will be resilient.
We owe it to our students and patients. -
Tinglong Dai reposted
François Chollet
@fchollet
The $500B number is transparently bogus. Masa doesn't have the money, the UAE doesn't have the money. You'd need a $100B/year in revenue just to cover depreciation. We'll see at most $150B.
-
Tinglong Dai
@TinglongDai
The End of AI’s Artificial Scarcity
A small Chinese tech company, little known even in China, has just forced investors to rethink the economics of artificial intelligence.
The release of DeepSeek R1 shook the financial markets, triggering a $1 trillion sell-off in tech stocks. This AI model, developed outside of Silicon Valley and China’s tech giants, showed next-generation AI no longer required billions of dollars—or an insatiable supply of electricity—to train or operate. Investors panicked, and with good reason.
The market’s reaction reflects a deeper realization: The perceived scarcity of AI was never real. DeepSeek R1 rivals OpenAI's GPT-4o1 in performance, yet its API is much cheaper—$0.55 per million input tokens compared to OpenAI's $15 for the same number of tokens. It can run locally on a workstation. And unlike other large AI systems, it has no rate limits. The notion that access to top-tier models would remain tightly controlled has fueled the AI investment boom, but that assumption no longer holds.
The Cracks Were Already Forming
For the past two years, investors and analysts have treated AI as a winner-take-all game, believing that training frontier models requires billions of dollars, massive compute clusters, and high-end Nvidia GPUs restricted by US export controls. AI was expected to remain centralized, with dominance concentrated in a few companies such as OpenAI, Anthropic, and Meta.
DeepSeek R1 is forcing a rethink of these beliefs.
Much attention has been paid to its reported $6 million training cost, a figure that has rightly been met with skepticism. This figure does not seem to include the cost of prior research, extensive model refinement, and access to large-scale computing infrastructure.
More importantly, DeepSeek almost certainly benefited from exposure to American AI models, probably through distillation techniques that allow smaller models to absorb knowledge from larger ones like GPT-4o1.
Even if the true cost was much higher than $6 million, one thing is clear: AI training and inference is becoming dramatically more efficient. The idea that cutting-edge AI always requires massive capital investment is no longer valid.
This is not an industry collapse, but a correction. AI remains a transformative technology, but the idea that only a few American companies can control its future no longer holds.
The Moat Is Shrinking—But Not Gone
DeepSeek R1 exposes weaknesses in the current AI business model, but it does not automatically dethrone AI giants like OpenAI. If anything, the situation is reminiscent of the search engine market: there are many alternatives to Google, but Google continues to dominate because even a small performance advantage is enough to cement user loyalty.
In AI, these small differences matter even more. Users—whether enterprise customers or individual developers—are not just looking for a model that performs well on benchmarks. They want reliability, consistency, ecosystem integration, and long-term support. AI is not just a product; it is a tool that users need to trust.
OpenAI has a formidable technical advantage that continues to grow, which translates into user confidence. The company’s integration with Microsoft provides a built-in distribution channel through Windows, Azure, and enterprise software. While DeepSeek R1 rivals GPT-4o1 in some raw capabilities, OpenAI's brand, infrastructure, and partnerships make it the default choice for most users.
AI is becoming increasingly specialized. While OpenAI may remain the leader in general-purpose AI, other models are likely to dominate in specific domains—whether mathematics, programming, music composition, movie production, or medicine. DeepSeek or future competitors do not need to outperform OpenAI across the board to gain market share in targeted applications.
AI Is the Great Leveler
For much of human history, power, wealth, and opportunity were tied to physical strength. The fastest runner, the strongest warrior, the most resilient worker—these were the people who determined the course of history. But machines wiped out those advantages. A kung-fu master is no match for a gun. A strong worker is no match for a forklift. What was once a defining human trait became secondary to technology.
AI is on the verge of doing the same for intelligence.
For centuries, cognitive ability has been a key determinant of success. Some people were naturally better at solving problems, reasoning, and absorbing knowledge. These differences have shaped corporate hierarchies, scientific advances, and financial markets.
AI changes this equation. Just as machines have compensated for physical differences, AI will compensate for cognitive differences. Complex tasks that once required years of training and elite analytical skills—discovering drugs, designing software, diagnosing diseases—are now within reach of anyone with a good AI model.
This does not mean that human intelligence will become obsolete, just as physical strength did not disappear when machines took over work. But intelligence will not be as important a determinant of success as it has been in the past. The advantage will shift from raw cognitive ability to knowing how to use intelligence wisely and for meaningful purposes.
The Business Model No Longer Holds
The biggest threat to OpenAI, Anthropic, and Meta is not DeepSeek R1 itself, but what it represents. These companies built their businesses on the assumption that AI inference would remain expensive and cloud-based, with users paying a premium for API access.
But what happens when high-quality models can run locally on consumer hardware?
DeepSeek R1 shows that AI is moving away from the cloud. If today’s best models can run on high-end workstations, similar models will soon run seamlessly on smartphones. This transition would undermine the cloud-based AI access model, just as local processing disrupted cloud gaming.
Even if AI services remain in demand, margins will erode. Hosting AI models is quickly becoming a commodity business, and premium pricing for general-purpose AI is unlikely to be sustainable.
The Policy Failure No One Wants to Discuss
Beyond its business implications, DeepSeek R1 exposes fundamental flaws in U.S. efforts to contain China’s AI progress.
For the past two years, the U.S. has restricted exports of Nvidia’s most powerful chips, betting that without access to cutting-edge hardware, China’s AI development would slow. DeepSeek R1 suggests that the bet isn’t paying off.
The model almost certainly benefited from distillation techniques, learning from American AI systems without direct access to restricted GPUs. And as China’s own AI models advance—fed in part by their exposure to Western systems—simply blocking hardware sales won’t be enough to keep China from catching up.
This raises difficult questions. If cutting off Nvidia chips isn’t working, what should U.S. policy focus on instead? Should restrictions target model architectures and training methods, not just hardware?
Is containment already impossible?
For now, Washington doesn’t have an answer.
AI Isn’t Collapsing—It’s Growing Up
The market’s reaction to DeepSeek R1 was dramatic, but it was also inevitable. Investors had been pricing AI as if the scarcity would last forever. That was never realistic.
None of this means that AI is slowing down. Quite the opposite—AI is accelerating, just in a way that financial markets have not priced in.
The competitive landscape is shifting. AI is getting cheaper, more efficient, and more accessible. The moat around large-scale AI training is shrinking, and the real power now lies in applications, distribution, and deployment.
OpenAI remains dominant, just as Google still leads search despite credible alternatives. But as AI becomes more specialized, new leaders will emerge in targeted areas. The days of assuming that one or two companies will control the entire industry are over.
AI isn’t collapsing. It’s entering its next phase. -
Tinglong Dai reposted
François Chollet
@fchollet
Any AI technique that is less data-efficient and energy-efficient than human intelligence will inevitably be replaced by something better in due time. Data-intensive and compute-intensive techniques can only be temporary, like the room-size computers of the 1950s.
-
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
More than ever, researchers need to explain how they use taxpayer money to provide tangible benefits.
My lab has developed a completely novel way to make cellular therapies a lot more effective. Our approach shows promising results in pancreatic, ovarian and lung cancer. -
Tinglong Dai reposted
FedEx Institute of Technology
@FedexInstitute
We are excited to welcome Dr. Dai, @JHUCarey, for a talk titled "Johns Hopkins Drug Supply Chain Data Dashboard: Improving Data Transparency and Increasing Resiliency in the U.S. Pharmaceutical #supplychain."
Feb. 6 | 3 PM
MPT (103)
Register: memphis.co1.qualtrics.com/jfe/form/SV_7ae2h4jpf0IztgW -
Tinglong Dai
@TinglongDai
16 degrees in Baltimore
-
Tinglong Dai
@TinglongDai
Lunar New Year pre-celebration in Baltimore 🧧
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
Deseret News Opinion
@DeseretOpinion
AI is no longer a niche field pursued by academics and Silicon Valley scientists. — Tinglong Dai www.deseret.com/opinion/2025/01/18/ai-artificial-intellig…
-
Tinglong Dai
@TinglongDai
-
Tinglong Dai
@TinglongDai
Happy 2025! Join us this Friday (Jan. 10) for a fireside chat, “AI for Medical Diagnosis and Prognosis: Reflections and Vision for 2025,” hosted by @JHU_HBHI as part of our Conversations on the Business of Health Webinar Series.
We’ll be joined by Dr. @oziadias, Blue Cross of California Distinguished Associate Professor at @UCBerkeley, whose seminal work has shaped how we think about the role of AI in medicine. The conversation will be moderated by Dr. @healthecon_dan, @JHU_BDPs of Health Economics and Director of @JHU_HBHI.
As co-hosts of this event, Dr. Risa Wolf and I look forward to welcoming you to this important conversation. Together, we’ll explore key topics such as how AI can improve diagnosis and prognosis, the challenges of bias and adoption, and what the future holds for healthcare powered by AI.
This is an exciting way to begin 2025, and we hope you’ll join us!
Tinglong Dai
@TinglongDai
The webinar will take place on Friday, January 10, 2025, from 12 to 1 PM Eastern Time. You can register for the webinar here: jh.zoom.us/webinar/register/WN_1o5X6NL9QkygPhANTSBh4Q.
-
Tinglong Dai
@TinglongDai
Even with its search function, @ChatGPTapp 4o is rarely reliable for any serious information collection task. Its results should be marked: CAUTION, the following may be complete nonsense.
-
Tinglong Dai
@TinglongDai
Supply chain
-
Tinglong Dai
@TinglongDai
Will AI make healthcare more affordable? Will AI replace physicians? These are the questions we explore in our new paper, published today in @npjDigitalMed.
Our incredible research team—including some of the nation’s medical AI heroes, Mahnoor Ahmed, @RoomasaC, @MichaelAbramoff, Harold Lehmann, and Risa Wolf—explored how autonomous AI can address diabetic retinopathy, the leading cause of blindness among working-age adults.
As treatable as diabetic retinopathy is, far too many patients remain unscreened each year. Thankfully, the first @US_FDA-cleared autonomous AI was created to help diagnose it.
In this study, we show autonomous AI saves money. In a large health system, it helps screen up to 1,724 more patients per year and improves follow-up rates by a factor of 11. Even small clinics see net savings with as few as 241 patients screened annually.
AI doesn’t replace specialists. Using AI improves follow-up rates to 65%, compared to just 29% with traditional methods. In large health systems, this means more work for specialists as they treat the extra patients identified by AI. In smaller systems, specialists may shift from routine screenings to focusing on complex cases flagged by AI.
Tinglong Dai
@TinglongDai
Link to the full-text: www.nature.com/articles/s41746-024-01382-4
-
Tinglong Dai
@TinglongDai
“So you work your life away just to pay for a timeshare down in Destin.” — @taylorswift13
-
Tinglong Dai
@TinglongDai
.@Amtrak is a never-ending source of national embarrassment.
Alan Cole
@AlanMCole
An Amtrak train left DC without any of its passengers because of their famously dumb boarding procedure where they don’t let you wait on the platform.
cc @mattyglesias @jbhenchman -
Tinglong Dai
@TinglongDai
New Orleans
-
Tinglong Dai
@TinglongDai
.@nytimes totally nailed it:
-
Tinglong Dai reposted
Arjun (Raj) Manrai
@arjunmanrai
Our conversation with Larry Summers is now out on @NEJM_AI Grand Rounds. There were so many memorable things Larry said, here's just one quote.
Full 39 min. interview: ai-podcast.nejm.org/e/the-economics-of-ai-a-conversation-… -
Tinglong Dai reposted
Mark Dredze
@mdredze
The next phase of our hiring is now live! These 30 @JHU_BDPs faculty slots are part of @HopkinsDSAI initiative to grow AI at @johnshopkins @HopkinsEngineer
Areas: medicine, foundational ML, natural intelligence, neuroscience, genomics, cancer, and computational social sciences
Bloomberg Distinguished Professors
@JHU_BDPs
Today @JohnsHopkins announced seven Bloomberg Distinguished Professorships research clusters as a key part of the university’s new @HopkinsDSAI initiative. These clusters aim to advance interdisciplinary research in data science and AI.
hub.jhu.edu/2024/12/17/data-science-and-ai-institute-bdp-… -
Tinglong Dai
@TinglongDai
.@ylecun on his disagreements with Hinton and Bengio @JHUBloombergCtr:
“We agree on a lot of things, but we also disagree on a few. For example, Jeff [Geoffrey Hinton] believes that current large language models (LLMs) have subjective experiences. I completely disagree—I think he’s completely wrong. Yasha comes at it from a slightly different angle; he’s more concerned about bad actors using AI for harmful purposes, like developing bioweapons or chemical weapons.
“Honestly, I think these fears have been exaggerated and distorted over the years. They’ve become incredibly inflated to the point where they no longer make sense. Delusional is the word I’d use. Some people pushing for extreme regulations—like SB 942—are, frankly, delusional. A year ago, they said AGI (artificial general intelligence) would arrive in a few months, and clearly, they were wrong.
“This is what happens when people talk about AGI—they hear about it and immediately think of some science fiction plot. The reality is far more complex, and we’re not even close to AGI.” -
Tinglong Dai
@TinglongDai
.@ylecun speaking at @JHUBloombergCtr: “Some people used to believe that if you just scaled up models like ChatGPT—trained them on more data with more compute—human-level intelligence would eventually emerge. I never believed in that idea. Ilya was famously a believer, but apparently not anymore. We’ve hit a ceiling: there’s no more natural data. We’ve basically exhausted all publicly available text on the internet. It’s clear that we’re reaching the limits of performance for these systems.”
-
Tinglong Dai
@TinglongDai
.@ylecun on AI regulation: “I’ve always been against regulating AI research and development because it’s counterproductive and based on false fears about its dangers. For example, people worry that AI will take over the world—but we’re nowhere near that. At this stage, any attempt to regulate research would only lead to regulatory capture by a few dominant companies. Instead, the focus should be on freeing up and increasing government funding to support innovation.” @JHUBloombergCtr
-
Tinglong Dai
@TinglongDai
LIVE: .@ylecun at @JHUBloombergCtr doing live podcast recording with @karaswisher
-
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
Hopkins is recruiting 30 Bloomberg Distinguished Professors in AI + X
(X = astrophysics, cancer, neuroscience, genomics, social sciences, etc.)
hub.jhu.edu/2024/12/17/data-science-and-ai-institute-bdp-…
22 of these 30 will be organized in clusters: -
Tinglong Dai
@TinglongDai
This morning at @Enterprise:
Employee: We have your reservation, but we don’t have a car.
Me: So… is this a Seinfeld episode?
Tinglong Dai
@TinglongDai
For those who don’t know:
youtu.be/A-brgkkjnHc -
Tinglong Dai
@TinglongDai
.@US_FDA is not a serious federal agency
-
Tinglong Dai
@TinglongDai
Montreal
-
Tinglong Dai reposted
Neil Renic
@NC_Renic
Academia is mostly apologising for the late reply
-
Tinglong Dai
@TinglongDai
We live in a simulation
-
Tinglong Dai reposted
Crémieux
@cremieuxrecueil
France's national railway company pulled out of bidding for California's high-speed rail contract in 2011.
They concluded that the state was terrible to work in and they preferred to go somewhere that worked better.
They went to North Africa.
Department of Government Efficiency
@DOGE
Summary of California High Speed Rail Project:
-Originally projected (in 2008) to cost $33 billion; now projected to cost between $88.5 and $127.9 billion
-Estimated completion date was 2020; as of 2024, zero passengers have been transported and the majority has not even been fully designed
-Received $6.8 billion in federal funds
-Requesting $8 billion in additional federal funds
Sources:
hsr.ca.gov/wp-content/uploads/2024/02/2024-Draft-Business…
amp.fresnobee.com/news/local/high-speed-rail/article28606… -
Tinglong Dai
@TinglongDai
Vieques
-
Tinglong Dai
@TinglongDai
Puerto Rico
-
Tinglong Dai
@TinglongDai
We are fortunate to have Dr. @MartyMakary leading the @US_FDA.
Marty is a surgeon, scientist, and professor at @HopkinsMedicine and @JHUCarey.
He is a visionary, a fighter, and a tireless advocate for patients.
As a three-time @NYTimes bestselling author, he brings courage, creativity, humor, and a fearless commitment to truth.
The FDA exists to protect trust—in food, medicine, and the devices that sustain us.
Trust, like the finest glass, cracks quietly—fractured by contaminated baby formula, weak or even non-existent inspections, poorly regulated AI devices.
And a global supply chain stretched beyond its limits. The FDA assures us we can trust every generic drug coming from every corner of the Earth. But can we?
Change won’t come overnight, but Marty’s integrity, fearless leadership, and pursuit of what’s right will steer the FDA toward rebuilding trust.
At the September health policy conference Prof. @GeBaiDC organized at @JHUBloombergCtr, Marty spoke with a sense of urgency:
“Just be obedient.
Keep your head down.
Do your job.
“We collect our badges,
attend medical conferences,
present our PowerPoints—
talking only to ourselves.
“Meanwhile, nothing is changing.
We’re not moving forward;
we’re sliding backward.
as a healthcare system.”
We’re lucky to have Marty leading this fearless new chapter. -
Tinglong Dai reposted
Alessandro Rebucci
@arebucci1
@JohnsHopkins #geoeconomics Conf - Call For Papers 👇
Key Dates and Highlights
-Submissions by 01/07/25
-Conference on May 2-3, 2025
-Keynote @m_maggiori
-Policy panel with @pogourinchas
-Lightning-round session for PhDs
-More details here: www.ssrn.com/index.cfm/en/janda/announcement/?id=16296 -
Tinglong Dai reposted
François Chollet
@fchollet
This is exactly what "memorization" means, to be clear. They memorize procedures (in the form of vector functions) that match their training data, and reapply those procedures at test time. And these procedures tend to generalize poorly, precisely due to being vector functions. We've known this about LLMs since inception, and we've known this about RNNs since 2016.
Rohan Paul
@rohanpaul_ai
Your LLM isn't doing math - it's using clever pattern matching tricks.
LLMs perform arithmetic using neither robust algorithms nor memorization; rather, they rely on a “bag of heuristics”, as proposed in this paper.
🤔 Original Problem:
Do LLMs solve reasoning tasks using actual algorithms or just memorize training data? This fundamental question impacts how these models truly learn and generalize.
-----
🛠️ Solution in this Paper:
Using arithmetic as a test case, researchers analyzed individual neurons in LLMs to understand their reasoning mechanism. They identified a circuit of ~200 neurons per layer (1.5%) that handles arithmetic operations. Each neuron implements simple pattern-matching rules called "heuristics" - like recognizing when operands fall within certain ranges or have specific patterns.
-----
💡 Key Insights:
→ LLMs don't use true mathematical algorithms or pure memorization
→ They combine multiple simple pattern-matching rules (heuristics) to solve arithmetic
→ Different neurons activate for different numerical patterns and boost corresponding answer tokens
→ This explains both their capabilities and limitations in arithmetic reasoning
→ The mechanism appears early in training and remains consistent throughout
-----
📊 Results:
→ The identified circuit achieves 96% faithfulness in reproducing full model behavior
→ Only 200 neurons per layer (1.5%) are needed for accurate arithmetic
→ 91% of top neurons implement identifiable heuristic patterns
→ Each heuristic causes 29% accuracy drop when ablated -
Tinglong Dai
@TinglongDai
An important piece by @PTetlock: link.springer.com/article/10.1007/s12115-014-9850-6
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai
@TinglongDai
.@TheEconomist has nailed it: GenAI is massively useful, and its real-world usage is underestimated, just more so internally and in secret
-
Tinglong Dai reposted
Crémieux
@cremieuxrecueil
The World Bank puts out a lot of reports and nobody reads them.
-
Tinglong Dai
@TinglongDai
Good morning, Houston
-
Tinglong Dai reposted
Alex Chan
@alex8chan
Thank you so much for having me! @JHUCarey @JHU_HBHI has one of the world’s most exciting set of scholars and it’s a pleasure to visit! Can’t wait to see y’all again soon!
-
Tinglong Dai
@TinglongDai
The brilliant @alex8chan from @HarvardHBS is gracing us at @JHUCarey today with not one, but two of his latest papers, on organ donation and racial discrimination against physicians. An absolute treat for the @JHU_HBHI community!
-
Tinglong Dai reposted
Tara Bannow
@TaraBannow
Some health care leaders say they're not concerned about Trump’s tariffs because they expect to be carved out
“Pharmaceuticals are usually not affected by tariffs,” AstraZeneca's CEO told reporters last week.
www.statnews.com/pharmalot/2024/11/21/trump-tariffs-healt… from me & @edsilverman @statnews -
Tinglong Dai
@TinglongDai
Will Trump’s tariffs make drugs and medical devices more expensive? A lot of folks think so, but I have a different take. Had a great exchange with @edsilverman of @statnews, and he captured my thoughts nicely in his column with @TaraBannow: www.statnews.com/pharmalot/2024/11/21/trump-tariffs-healt…
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Tinglong Dai
@TinglongDai
"INTRODUCTORY KOREAN DRAMA"
By the legendary Uwe E. Reinhardt @Princeton
web.archive.org/web/20131228111502/https://www.princeton.… -
Post not available here, view on X
Tinglong Dai
@TinglongDai
"INTRODUCTORY KOREAN DRAMA"
By the legendary Uwe E. Reinhardt @Princeton
web.archive.org/web/20131228111502/https://www.princeton.… -
Tinglong Dai reposted
Tinglong Dai
@TinglongDai
Can't help posting this:
-
Daniel Polsky
@healthecon_dan
We are not in Kansas anymore.
www.politico.com/news/2024/11/19/dr-oz-cms-administrator-…
Tinglong Dai
@TinglongDai
Can't help posting this:
-
Tinglong Dai reposted
Daniel Polsky
@healthecon_dan
We are not in Kansas anymore.
www.politico.com/news/2024/11/19/dr-oz-cms-administrator-… -
Tinglong Dai
@TinglongDai
Possibilities
By Wisława Szymborska (1923–2012)
I prefer movies.
I prefer cats.
I prefer the oaks along the Warta.
I prefer Dickens
to Dostoyevsky.
I prefer myself liking people
to myself loving mankind.
I prefer keeping a needle and thread
on hand, just in case.
I prefer the color green.
I prefer not to maintain
that reason is to blame
for everything.
I prefer exceptions.
I prefer to leave early.
I prefer talking to doctors
about something else.
I prefer the old
fine-lined illustrations.
I prefer the absurdity
of writing poems
to the absurdity
of not writing poems.
I prefer, where love’s concerned,
nonspecific anniversaries
that can be celebrated
every day.
I prefer moralists
who promise me nothing.
I prefer cunning kindness
to the over-trustful kind.
I prefer the earth
in civvies.
I prefer conquered
to conquering countries.
I prefer having
some reservations.
I prefer the hell of chaos
to the hell of order.
I prefer Grimms’ fairy tales
to the newspapers’ front pages.
I prefer leaves without flowers
to flowers without leaves.
I prefer dogs
with uncropped tails.
I prefer light eyes,
since mine are dark.
I prefer desk drawers.
I prefer many things
that I haven’t mentioned here
to many things
I’ve also left unsaid.
I prefer zeroes on the loose
to those lined up
behind a cipher.
I prefer the time of insects
to the time of stars.
I prefer to knock on wood.
I prefer not to ask
how much longer and when.
I prefer keeping in mind
even the possibility
that existence has its own
reason for being. -
Tinglong Dai reposted
Carole Cadwalladr
@carolecadwalla
2. Journalists will be first. But everyone else is next.
How do we know this? Because it always is. It's the authoritarian playbook.
Lawsuits are first. Prosecutions are next. America needs to learn these moves fast. It's already later than you think.
3/ -
Tinglong Dai reposted
Johns Hopkins Data Science and AI Institute
@HopkinsDSAI
We’re thrilled to announce the #hopkinsdsai Postdoctoral Fellowship Program! We’re looking for candidates across all areas of data science and AI, including science, health, medicine, the humanities, engineering, policy, and ethics.
Apply today!
ai.jhu.edu/postdoctoral-fellowship-program/ -
Tinglong Dai reposted
Alex Byers
@byersalex
“And which great efficiency expert will lead this important new effort to streamline?”
“Actually there will be two people in charge” -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai
@TinglongDai
California will have high-speed rail before it has normal-speed vote counting
Nate Silver
@NateSilver538
Since this election is settled this hopefully won't take on too much of a partisan valence but can we just agree it's ridiculous that it takes California several weeks to count its votes? Imagine if it were a swing state.
-
Tinglong Dai reposted
The Kobeissi Letter
@KobeissiLetter
So what happened in 2021 after the S&P 500 hit 70+ all time highs?
It was followed by the bear market of 2022 where many tech stocks fell over -50%.
In fact, the entire 2021 gain was undone in that following year.
What about in 2017 when the S&P 500 saw 60+ all time highs? -
Tinglong Dai reposted
President Biden Archived
@POTUS46Archive
The struggle for the soul of America since our very founding has always been ongoing.
Campaigns are contests of competing visions. And we accept the choice the country made.
You can’t love your country only when you win. -
Tinglong Dai reposted
Alexis Battle
@alexisjbattle
Going on the faculty market this year? JHU is hiring at all levels. #ashg2024 #ashg24 folks in computational genomics, you could fit in here, contact me if interested!
Johns Hopkins Data Science and AI Institute
@HopkinsDSAI
We're hiring! Faculty of all ranks across the fields of data science and AI. Come be a part of this historic investment in data science and AI. Join us today! @JHUCompSci @JHUECE @JohnsHopkinsAMS
Check out all open positions:
ai.jhu.edu/careers-for-faculty/ -
Tinglong Dai
@TinglongDai
From @MikeBloomberg:
-
Tinglong Dai reposted
François Chollet
@fchollet
Not long ago, most AI experts knew for a fact that neural networks were a failed avenue. Even as late as 2014, many of the top researchers in computer vision were saying that deep learning might be just a temporary fad.
-
Tinglong Dai
@TinglongDai
.@jaycaspiankang wasn’t too off in retrospect
Tinglong Dai
@TinglongDai
From @jaycaspiankang in @NewYorker (June 30):
-
Tinglong Dai reposted
Hinrich Foundation
@hinrichfdn
De-risking is a reality we can no longer avoid. While the movement of goods is relatively easy to track, losses on information and human flows are less visible. These flows, however, are vital for #supplychains to work, @TinglongDai & @christangucla write: www.hinrichfoundation.com/research/wp/supply-chain-securi…
-
Tinglong Dai reposted
Joey Politano 🏳️🌈
@JosephPolitano
Just insane that America had the best productivity growth of comparable major nations in the years pre-COVID, then went "hold my beer" post-COVID and had ~double the productivity growth of the next-closest country
-
Tinglong Dai reposted
Center on Japanese Economy and Business (CJEB)
@ColumbiaCJEB
Join us 11/21 for a Live Webinar: The Salesforce Way of Marketing in the AI Era. Featuring Shoko Suzuki, SVP, Chief Marketing Officer, Salesforce Japan Co., Ltd. Read more and RSVP: www.ticketleap.events/tickets/cjeb-columbia/112124 #japan #salesforce #webinar @Columbia @Columbia_Biz
-
Tinglong Dai reposted
Timothy Liu
@timothyhliu5
Tepper was passed over for partner at Goldman three times
The third time because Jon Corzine viewed Tepper as "Rubin's boy", not his guy
In 2010 Tepper bought Corzine's house in the Hamptons from Corzine's ex-wife for $44 million, tore it down, and built a house twice the size
Finding Compounders
@F_Compounders
David Tepper’s presentation to Undergraduate Business Students
-
Tinglong Dai
@TinglongDai
Stunning shift: In the first 3Qs of 2025, Warren Buffett’s Berkshire Hathaway sold nearly 20x more stock than it bought
-
Tinglong Dai reposted
Mike Bloomberg
@MikeBloomberg
Glad to celebrate one year of the Hopkins Bloomberg Center that is bringing JHU's world-class research and knowledge to the center of Washington to help tackle many of the most difficult challenges facing the nation and our world.
-
Tinglong Dai reposted
Laura Albert
@lauraalbertphd
Happy Halloween!
-
Tinglong Dai reposted
Mike Bloomberg
@MikeBloomberg
I voted for @KamalaHarris without hesitation. When it comes to policy and personal integrity, the contrast could not be clearer, and I hope undecided voters with a history of supporting candidates in both parties join me. bloombg.org/4ecs0cX
-
Tinglong Dai
@TinglongDai
From founder & CEO of @perplexity_ai:
-
Tinglong Dai
@TinglongDai
Duluth, Minnesota:
-
Tinglong Dai reposted
The Lancet Haematology
@TheLancetHaem
2- Jacqueline del Castillo @jdelc: "this blood bracelet is dedicated to the people encircling around every haematology patient, bringing care, healing, laughter, and love. It traces the lines of our haematology community" #ourcommunity @TheEBMT_PAC
www.thelancet.com/journals/lanhae/article/PIIS2352-3026(2… -
Tinglong Dai reposted
Roxana Daneshjou MD/PhD
@RoxanaDaneshjou
The specificity was terrible - 20.6%. Also, these are lesions that were ALREADY deemed by the PCPs to be suspicious. It's not a screening tool, and it's not catching anything the PCP isn't suspicious of.
-
Tinglong Dai reposted
The National Bureau of Asian Research
@NBRnews
@TinglongDai and @christangucla examine #supplychain #derisking driven by concerns of the U.S. and its allies about supply chain resilience and #china’s strategy to redefine the global order. Free #asiapolicy essay made possible by @hinrichfdn support. www.nbr.org/publication/de-risking-global-supply-chains-l…
-
Tinglong Dai reposted
Lawson Mansell
@lawsonhmansell
What’s driving our high and ineffective spending? It’s the prices of hospital and physician services!
While over utilization is certainly a contributor, about two-thirds of the spending growth between 2015 and 2019 came from higher prices. -
Tinglong Dai reposted
Francis Fukuyama
@FukuyamaFrancis
Last time I checked, I’m still alive.
-
Tinglong Dai
@TinglongDai
The secret to happiness
is freedom…
and the secret to freedom
is courage.
— Thucydides (c. 460 – c. 400 BC) -
Tinglong Dai reposted
Neil Renic
@NC_Renic
An academic’s academic
-
Tinglong Dai reposted
kate
@whoiskatrin
I’ve never told how much @Apple did to help evacuate my grandmother from the hell that was happening in Kharkiv in 2023. They hired security contractors to get people out. The level of coordination on the ground and online was insane—translators, drivers, everything. Grateful.
-
Tinglong Dai
@TinglongDai
Delighted to publish "De-risking Global Supply Chains: Looking Beyond Material Flows" (with @christangucla) in Asia Policy, a leading journal of @NBRnews.
In this 24-page article, published today, we challenge the notion that "supply chain de-risking" is a unilateral effort by the West. In fact, China has been steadily de-risking its own supply chains since 2005, as part of its long-standing strategy to redefine the global order.
Supply chains are flows. While most discussions of de-risking focus on material flows — the goods we can see moving across borders — there is a very different story beneath the surface.
Disruptions to information, financial, and human flows often fly under the radar, but these hidden currents can have an even greater impact in the years to come.
A big thank you to Jessica Keough for her tireless work in bringing this essay to life!
Download the full text PDF here: www.nbr.org/wp-content/uploads/pdfs/publications/asiapoli… -
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
The new @JohnsHopkins School of Government and Policy located at the #hopkinsbloombergcenter has a new dean - William G. Howell!
He is a scholar of issues related to separation of powers, political institutions, democratic reform, and the U.S. presidency. hub.jhu.edu/2024/10/24/johns-hopkins-school-of-government… -
Tinglong Dai reposted
Jeromie Ballreich
@jeromie_b
special thanks to @DrTraceyK and @USUhealthsci for coordinating this effort. And shoutout to colleagues who are tackling this effort @MWosinska @dr_msocal @TinglongDai @jgreene2 @drJoshS
@BSPH_HPM -
Tinglong Dai reposted
Jeromie Ballreich
@jeromie_b
I had a wonderful time at the @WhiteHouse discussing pharmaceutical supply chains. It's great to see the administration's interest in this crucial policy area. Our pharmaceutical supply chain is a matter of national security!
-
Tinglong Dai reposted
Laura Albert
@lauraalbertphd
.@INFORMS President Julie Swann kicks off #informs2024 by announcing a major publication deal that will expand the number of libraries that access our journals by more than 10X and that the conference set an attendance record with more than 7000 registrations ‼️#orms
-
Tinglong Dai reposted
Matthew E. Kahn
@mattkahn1966
Congratulations! Back in 1993, the Columbia IAB building was a nasty, old building. Flash forward to 2024 and it is even nastier but the might Economics Department keeps hiring great faculty. Perhaps capital and human capital are substitutes?
Trade Diversion (Jonathan Dingel)
@TradeDiversion
In January 2025, I will join Columbia University as a tenured Associate Professor of Economics. I’m happy to return to my alma mater.
tradediversion.net/2024/10/18/luck-along-the-way/ -
Tinglong Dai
@TinglongDai
The @JohnsHopkins Annual VPR Awards Celebration has become the Oscars of our research community — and there is much to celebrate. For starters, for the 44th year in a row, Hopkins leads the nation with a record-breaking $3.4 billion in research funding.
Here is the inimitable @deniswirtz sharing his reflections with the incredible community at the iconic @george_peabody Library: -
Tinglong Dai reposted
Mario Macis
@Mario_Macis
The one and only @deniswirtz at the annual @JohnsHopkins VPR Awards Celebration. Very grateful for the support! Great sharing this evening with amazing colleagues in the stunning Peabody Library.
@JHUCarey @JHU_HBHI -
Tinglong Dai reposted
Daron Acemoglu
@DAcemogluMIT
Herkese sonsuz teşekkürler
Hem arkadaşlarım hem de aynı zamanda meslektaşlarım olan Simon Johnson ve James Robinson ile beraber, 2024 Nobel Ekonomi Ödülü’ne layık görülmüş olmaktan son derece mutlu ve gururluyum.
Aynı zamanda bu ödülün, ekonomik büyümenin anlaşılmasında kurumların rolünün önemini ve demokrasi ile siyasi değişim meselelerini vurgulamasıyla, sadece Simon, Jim ve benim çalışmalarımı değil, aynı zamanda bu konuda emek sarf eden bir çok harika genç ekonomist ve sosyal bilim insanının da çalışmalarını takdir etmiş olmasından dolayı da çok mutluyum.
Demokrasi olgusunun saldırı altında olduğu zamanlardan geçiyoruz. Yapay zeka ile ilgili bir çok konu, ekonomik eşitsizlikler, çalışma hayatının geleceği, yaşlanan dünya ve iklim değişimi gibi hepsi birbirinden önemli konularla baş edebilmek için gerekli olan toplumsal mutabakatı sağlamanın yolu her zamankinden daha güçlü kurumlara sahip olmaktan geçiyor. Temennim, bu ödülün vurgusunun dünyada kurumları ve demokrasiyi güçlendirmenin ne kadar zorlu bir mücadele gerektirdiğine ışık tutacak olması.
Beni gerek kişisel olarak gerekse mesajlarla ve sosyal medya üzerinden tebrik eden ve iyi dileklerini gönderen herkese minnettarım. -
Tinglong Dai
@TinglongDai
2024 Nobel Laureate in Economics @DAcemogluMIT on reviewing and editing:
Tinglong Dai
@TinglongDai
Source: "The Art and Practice of Economics Research: Lessons from Leading Minds" edited by @SimonBowmaker: www.elgaronline.com/monobook/9781849808460.xml
-
Tinglong Dai
@TinglongDai
Wow, what an incredible honor to have spoken in the same speaker series as @DAcemogluMIT, thanks to @Supernetworks' brilliant organization and kind invitation!
With today's @NobelPrize announcement, the speaker series will be even more in demand 👏👏👏
Anna Nagurney
@Supernetworks
Extraordinary news this morning! We have hosted @DAcemogluMIT in our @umassinforms speaker series @umassamherst and his wife @AsuOzdaglar as well. Congratulations to all the @NobelPrize laureates!
-
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
Interesting quote: "As an active researcher himself, Wirtz is aware of the burden on scientists when it comes to drafting proposals for external grants; most proposals consist of “20 or 15 pages of actual science, but then hundreds of pages of crap. Quote me on that.”
-
Tinglong Dai reposted
Johns Hopkins University
@JohnsHopkins
Three years ago, after learning about Costa Rica's renowned public health system, Sarah Szanton, dean of @JHUNursing proposed a bold solution to address U.S. health care challenges. Now, a pilot program in Baltimore connects residents with nurses and health workers to reduce disparities, with plans to expand nationwide. hub.jhu.edu/magazine/2024/fall/neighborhood-nursing-pilot…
-
Tinglong Dai
@TinglongDai
Nobel Prize prediction: Taylor Swift for the Nobel in Economics?
-
Tinglong Dai reposted
Michael Baym
@baym
I feel like the critique that the physics Nobel is going to computer science misunderstands the culture of physics. They aren’t ceding the prize to cs, they’re claiming it as a branch of physics
-
Tinglong Dai reposted
The Nobel Prize
@NobelPrize
“How could I be sure it wasn’t a spoof call?”
2024 physics laureate Geoffrey Hinton received a phone call from Stockholm in the early hours in a hotel room in California. Multiple Swedish accents helped reassure him that his #nobelprize in Physics, awarded today, was real.This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Scott Kominers
@skominers
Apparently not LLMs 😅
Tinglong Dai
@TinglongDai
Congratulations to @geoffreyhinton and @HopfieldJohn, an inspiration to all students of deep learning and AI — and to @NobelPrize committee for this unconventional choice that took @ChatGPTapp and @AnthropicAI by surprise!
-
Tinglong Dai reposted
The Nobel Prize
@NobelPrize
"I'm in a cheap hotel in California which doesn't have a good internet or phone connection. I was going to have an MRI scan today but I'll have to cancel that!"
- New physics laureate Geoffrey Hinton speaking at today’s press conference where his #nobelprize was announced. -
Tinglong Dai reposted
Alex Chan
@alex8chan
This is why the @NobelPrize remains so relevant 123 years after its creation! A daring and innovative choice! Bravo!
The Nobel Prize
@NobelPrize
BREAKING NEWS
The Royal Swedish Academy of Sciences has decided to award the 2024 #nobelprize in Physics to John J. Hopfield and Geoffrey E. Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.” -
Tinglong Dai
@TinglongDai
Congratulations to @geoffreyhinton and @HopfieldJohn, an inspiration to all students of deep learning and AI — and to @NobelPrize committee for this unconventional choice that took @ChatGPTapp and @AnthropicAI by surprise!
-
Tinglong Dai
@TinglongDai
No better place in the world to discuss AI ethics and governance than @JHUBloombergCtr!
-
Tinglong Dai
@TinglongDai
“We have focused so much on talking about bias and equity, all of which are very important, and safety. We sort of stopped talking about the need for our systems to actually be effective,” said Dr. Jane Pinelis @JHUAPL at the @JohnsHopkins AI Ethics and Governance Symposium @JHUBloombergCtr
-
Tinglong Dai
@TinglongDai
“When you see an impressive robot, look for the human.” — @jleonardmit said at the AI Ethics and Governance Symposium @JohnsHopkins @JHUBloombergCtr
Tinglong Dai
@TinglongDai
“It's not an accident that Silicon Valley is 3000 miles away from DC.” - @jleonardmit @MIT
-
Tinglong Dai
@TinglongDai
Focus on how AI transforms the medical profession and improves operational efficiency, not whether AI beats humans, said @VictorDzau, president of @theNAMedicine at @JHU_BDPs @JohnsHopkins Symposium on AI Ethics and Governance
Tinglong Dai
@TinglongDai
Regarding the role of industry, @VictorDzau stated, “Now, I have large conversations about industry. We can talk about this, but I think the first part is not excluded to include them.” @theNAMedicine @JHUBloombergCtr @JohnsHopkins
-
Tinglong Dai
@TinglongDai
“I don't think the national academies are very different than your own organizations in that it seems like we're doing all AI all the time. It's really hard to get in a conversation with someone anywhere that it doesn't somehow turn to AI.” Said, @Marcia4Science, president of @theNASciences @JHUBloombergCtr @JohnsHopkins Symposium on AI Ethics and Governance
Tinglong Dai
@TinglongDai
“And the first principle we came up with was human accountability. Humans, anyone who uses AI has to take personal responsibility. You can't say, oh, the AI did that, and so you can't retract my paper because of what AI. No, you have to take responsibility.” @Marcia4Science of @theNASciences
Tinglong Dai
@TinglongDai
“To give advice based on AI is going to be a huge challenge because of the fact that, first of all, no one really knows how it works. And secondly, we're constantly seeing new sorts of issues and behaviors in a very fast-moving area,” said @Marcia4Science, president of @theNASciences
-
Tinglong Dai reposted
npj Digital Medicine
@npjDigitalMed
How accurate & ethical are claims of #artificialintelligence outperforming clinicians?
"Constructive ethical guidance that can benefit authors, journal editors, & peer reviewers when reporting and evaluating findings in studies comparing AI to physician performance" is offered in this commentary from @JojannekeDrogt, @JongsmaKr et al. from @UMCUtrecht.
www.nature.com/articles/s41746-024-01255-w -
Tinglong Dai reposted
John Burn-Murdoch
@jburnmurdoch
NEW: we may have passed peak obesity 🎉📈📉🙏
In what might be one of the most significant trends I have ever charted, the US obesity rate fell last year. -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
William Wang
@WilliamWangNLP
BREAKING: Taylor Swift's Eras Tour just did what AI couldn’t—pushed NeurIPS by a whole day! 🤖 🤣🤣🤣
#neurips 2024 Conference Date Change
The conference start date has been changed to Tuesday December 10 in order to support delegates arriving on Monday. neurips.cc -
Tinglong Dai
@TinglongDai
Automation has emerged as the biggest winner of the port strike. Never witnessed such a surge in support for automation before.
-
Tinglong Dai reposted
François Chollet
@fchollet
Lots of what is called "an LLM" is actually a complex symbolic software system involving numerous sub-calls to multiple LLMs
-
Tinglong Dai
@TinglongDai
According to @Polymarket, the East Coast port strike has a 9% chance of ending by this Friday, a 70% chance of ending by next Friday, and a 92% chance of ending by the end of October
Consumers have been incredibly calm. I don't know anyone who is rushing to stock up on anything, not even bananas -
Tinglong Dai reposted
Negin Golrezaei
@NeginGolrezaei
I'm sad to share that John D.C. Little, a long-time member of the MIT Sloan community, passed away on September 27, 2024. He made huge contributions to Operations Research, including the famous Little’s Law. RIP John. www.deefuneralhome.com/obituaries/John-D-C-Little?obId=33…
-
Tinglong Dai reposted
Frank Luntz
@FrankLuntz
We need more Midwestern presidential candidates. 💯
The Recount
@therecount
“Midwest Nice” takes the VP debate stage.
-
Tinglong Dai reposted
Peyman Milanfar
@docmilanfar
A surprising & little-known results in classical statistics:
mean (μ) and median (m) are within one std deviation:
|μ−m| ≤ σ
For unimodal densities, bound is even tighter
|μ−m| ≤ 0.7746 σ
This beautiful results first appeared in a 1932 paper by Hotelling & Solomons
1/3 -
Tinglong Dai reposted
Al Franken
@alfranken
.@danacarvey did his Joe Biden on @nbcsnl this past Saturday. He and I have been discussing it for a while. Here's a conversation we had a few months ago.
This media couldn't be displayed. See it on the original post. -
Tinglong Dai
@TinglongDai
I spoke on @NPR with @stephanie_h of @Marketplace about the looming longshoremen’s strike and its potential to cripple our supply chains, from car parts to fresh produce: www.marketplace.org/2024/09/30/longshoremens-strike-nears…
In the short term, I noted that we may see empty shelves of items like bananas and avocados, though price increases are less likely.
@stephanie_h asked me what else I have to say, so I mentioned U.S. ports are among the least automated globally, consistently ranking below many in the developing world. For example, as Prof. Christopher Tang likes to point out, major U.S. ports like Los Angeles (ranked 336th out of 348) and New York & New Jersey (ranked 306th) consistently fall near the bottom of global port efficiency rankings according to @WorldBank.
We are not even asking for world-beating ports. But what many consider to be the world's greatest nation should at least meet the standards of the developing world when it comes to automation.
Without this, labor-related supply chain disruptions will remain a recurrent concern.
@JohnsHopkins/@Jillrosen @JHUCarey @INFORMS -
Tinglong Dai reposted
Jim Gaffigan
@JimGaffigan
Such an incredible ride. SNL week 1
This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
The New York Times
@nytimes
In six billion years the sun will expand into a red giant, consuming Mercury, maybe Venus and potentially Earth, too.
But astronomers spotted a rocky world orbiting a white dwarf, suggesting a future in which our planet outlives its star. www.nytimes.com/2024/09/26/science/planet-white-dwarf.htm… -
Tinglong Dai
@TinglongDai
Spotted: A LaTeX Beamer slide at an ophthalmology conference @HMSeye
-
Tinglong Dai
@TinglongDai
Boston
-
Tinglong Dai
@TinglongDai
Heading to Boston tomorrow for the 2024 Harvard Data Science & AI Symposium, hosted by the @harvardmed @HMSeye & @MassEyeAndEar. Excited to moderate a panel on "Beyond Diabetic Retinopathy: What Eye Diseases Will AI Be Commercialized for Next?" at the Schepens Eye Research Institute, Friday, 9/27 from 1:50-2:35 pm.
Diabetic retinopathy (DR) is well known in the AI world, because it's the first FDA-approved commercialized autonomous AI system for. But what's next? We'll explore how AI could revolutionize the diagnosis, treatment, and prevention of eye diseases—and even systemic diseases—through AI-powered eye exams.
I'll be joined by incredible panelists leading the way:
- @MichaelAbramoff, MD, PhD, Founder & Exec. Chairman, @AItheRightWay
- @aabajwa, MCA, Co-Founder, CEO & CTO, @ophthalytics
- Petri Huhtinen, PhD, Clinical Director, @OptomedOyj
- Kaushal Solanki, PhD, CEO & Founder, @EyenukInc
For those attending, I look forward to seeing you in Boston for what's going to be an exciting conversation — let's shape the field together! -
Tinglong Dai
@TinglongDai
2025 @USNews Best National Universities🏆
1️⃣ @Princeton
2️⃣ @MIT
3️⃣ @Harvard
4️⃣ @Stanford
5️⃣ @Yale
6️⃣ @Caltech
6️⃣ @DukeU
6️⃣ @JohnsHopkins
6️⃣ @NorthwesternU
🔟 @Penn
Tinglong Dai
@TinglongDai
The 2025 @USNews Best National Universities (11-20):
⑪ @Cornell
⑪ @UChicago
⑬ @BrownU
⑬ @Columbia
⑮ @Dartmouth
⑮ @UCLA
⑰ @UCBerkeley
⑱ @RiceUniversity
⑱ @NotreDame
⑱ @VanderbiltU
Full list: www.usnews.com/best-colleges/rankings/national-universities -
Tinglong Dai
@TinglongDai
Inclusion starts with not intentionally excluding individuals with diverse needs. Conference organizers can and should do better in not excluding attendees traveling with young children. At least they’re not supposed to do so unapologetically and ruthlessly
Eve Hayes on bluesky at evehayes123
@EveHayes123
I’m at a conference today. The organisers just told me that I can’t bring my daughter to the networking session or the dinner tonight. I couldn’t get a babysitter. This is one of those times, as a single parent, that you really miss out. It felt like a gut punch.
-
Tinglong Dai reposted
Amit Jain, MD, MBA
@AmitJainSpine
One of the reasons I love @JohnsHopkins is the vibrant academic community across the university. I’m so excited and honored to have been granted an appointment as an associate professor of business w/ courtesy @JHUCarey!
@JHU_HBHI @healthecon_dan @TinglongDai @HopkinsMedicine -
Tinglong Dai reposted
François Chollet
@fchollet
When you use a tool to do a task you could have done yourself, your skills atrophy and you become dependent on the tool, giving it power over you.
This can be fine... Unless:
The task is vital, and the tool might not always be there.
You don't own the tool, giving someone else power over you.
The task is "thinking" -- surrendering your mind. -
Tinglong Dai reposted
Patricia Davidson
@DrPMDavidson
Lovely to see you soaring @TinglongDai @JohnsHopkins
Tinglong Dai
@TinglongDai
Great leaders may travel far, but their impact stays. Since Prof. @DrPMDavidson left @JohnsHopkins in 2021 to become a university president, her wisdom still resonates in our halls. Thrilled to catch up with her during her visit to Baltimore—her energy was pure joy and power! Welcome home, Trish! #hopkinsfamily
-
Tinglong Dai
@TinglongDai
Great leaders may travel far, but their impact stays. Since Prof. @DrPMDavidson left @JohnsHopkins in 2021 to become a university president, her wisdom still resonates in our halls. Thrilled to catch up with her during her visit to Baltimore—her energy was pure joy and power! Welcome home, Trish! #hopkinsfamily
-
Tinglong Dai reposted
Hopkins Med News
@HopkinsMedNews
Purposeful regulation is key to balancing AI innovations in healthcare with ethical standards. Read more in the recent @MedEconomics article by Johns Hopkins senior neurosurgery resident @tdazad and Dr. Tinglong Dai www.medicaleconomics.com/view/do-no-harm-the-imperative-f…
-
Tinglong Dai
@TinglongDai
Faculty compensation in the 1970s
-
Tinglong Dai
@TinglongDai
Good morning, Baltimore
-
Tinglong Dai reposted
Jim Gaffigan
@JimGaffigan
I post this every year on 9-11. No words. #remebering911
-
Tinglong Dai reposted
Tej D. Azad 🇺🇸
@tdazad
Thanks to @AAAS_STPF for highlighting this piece penned w/ @TinglongDai !
@JHU_HBHI @JHUCarey @HopkinsMedNews
S&TPolicyFellowships
@AAAS_STPF
How do we regulate AI purposefully to ensure that it enhances rather than hinders the practice of medicine? @STPF_Alumni @TDAzad argues for #artificialintelligence and how it can be used ethically in #healthcare.
www.medicaleconomics.com/view/do-no-harm-the-imperative-f… -
Tinglong Dai
@TinglongDai
Apple's event was an unmistakably staged show — polished, but lacking any real human connection, not even a feigned attempt at one. The green screen backdrop and forced enthusiasm only felt more disconnected. Yes, one can’t replicate Jobs’ performance, but an inspiring product release needs more than a bunch of non-acting actors. The pandemic ended last year. It's time to get real again.
-
Tinglong Dai reposted
David Simchi-Levi
@davidsimchilevi
@TinglongDai and I just published a short paper "Embracing the Impact: Understanding the Decline of Impact Factors in INFORMS Journals and the Future of OR/MS Publishing," see pubsonline.informs.org/do/10.1287/orms.2024.03.07/full/
-
Tinglong Dai
@TinglongDai
Prof. @davidsimchilevi of @MIT and I are thrilled to see our new OR/MS Today article published today: pubsonline.informs.org/do/10.1287/orms.2024.03.07/full/
David, who has served as Editor-in-Chief of two INFORMS flagship journals (Operations Research, 2006-2011; Management Science, 2018-2023), is a central figure in the world of OR/MS publishing. His vision and leadership have transformed the field and touched many lives.
We've been following recent changes in the impact factors of @INFORMS journals. In July 2023, we wrote a blog post on INFORMS Connect (connect.informs.org/discussion/unraveling-the-mystery-of-…) explaining the significant drop in impact factors across the board and predicting that the full impact of @Clarivate's methodological changes would be felt this year. The most recent impact factors confirmed our predictions.
Of the 12 INFORMS journals tracked by Clarivate, seven have seen their impact factors drop by more than 25% in the past two years. This motivated us to write a follow-up article looking ahead and discussing the future of OR/MS publishing.
We're especially grateful to Prof. @christangucla, @INFORMS VP of Publications, for his candid and insightful feedback on the need to adapt to these changes and rethink our publishing model. We also benefited greatly from Matt Walls, @INFORMS Director of Publications, whose in-depth research broadened our perspective.
A big thank you to Kara Tucker for publishing our article in the September issue of OR/MS Today and for her excellent editorial guidance.
As David and I wrote, "The world has moved on, and we risk being left behind if we do not follow suit. As a community, we have always embraced change and often benefited from it. There is every reason to believe that this time will be no different."
That's our hope — and our conviction.
Our @INFORMS community will embrace change and emerge stronger than ever! -
Tinglong Dai
@TinglongDai
Come visit @JHUCarey - the new glass wall on the 12th floor adds even more beauty
-
Tinglong Dai reposted
François Chollet
@fchollet
From Deep Learning with Python:
- Deep learning research is an evolution process driven by poorly-understood empirical results
- Math in DL papers is usually worthless and was placed there purely as a signal of seriousness
- The key to good research is understanding what *causes* your results, and that requires ablation studies (which few researchers attempt to do)
Syed Amaan
@syedamaann
@fchollet on ablation studies in deep learning research:
-
Tinglong Dai reposted
Steven Salzberg 💙💛
@StevenSalzberg1
Good piece in @WSJ today by @NidhiSubs, about the unraveling of the cancer microbiome hypothesis, incl quotes from me and others in the field www.wsj.com/health/healthcare/cancer-study-retracted-rese…
-
Tinglong Dai
@TinglongDai
ChatGPT is so inspiring ✨
-
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
A mysterious cartoonist captures the essence of my contribution to my students' research.
(Left: my colleague Prof. Pei-Hsun Wu, right: me) -
Tinglong Dai
@TinglongDai
Chicago
Tinglong Dai
@TinglongDai
“Stay on the upper deck. Because the river is only 17 feet deep. The boats. It's 24 feet high, so the math is in your favor.”
Safety instructions can be fun. -
Tinglong Dai
@TinglongDai
45 minutes ahead of schedule for a 3-hour train ride - @Amtrak is a breath of fresh air
-
Tinglong Dai
@TinglongDai
Champaign-Urbana
-
Tinglong Dai reposted
Ed Livingston
@ehlJAMA
The same thing happened with the University of California. They canceled their Elsevier contract and were expecting a huge backlash. After about a year, one of the librarians told me that they were surprised that almost nobody complained.
When I was at JAMA, we did a survey of young clinicians to find out how they acquire medical information. They told us they got it from UpToDate, Wikipedia and podcasts. They specifically said they don’t read journals or textbooks.
Richard McElreath 🐈⬛
@rlmcelreath
What happened when MIT stopped paying Elsevier? Not much except they are saving a lot of money. "MIT is interested in collaborating with other libraries to reinvest these funds in community-controlled open publishing initiatives..." sparcopen.org/our-work/big-deal-knowledge-base/unbundling…
-
Tinglong Dai reposted
Daniel Polsky
@healthecon_dan
Congratulations to @MattDEisenberg and @JHSPH_CMAP! This team will certainly make a policy impact on the important public health issue of suicide prevention through their exceptional research. Thank you @AmericanHealth!
Matthew Eisenberg
@MattDEisenberg
I am absolutely honored that @JHSPH_CMAP has received this funding from @AmericanHealth. It has been a whirlwind first year as CMAP director and I’m so proud of all we have accomplished and the faith that has been placed in the team. None of this would be possible without the many contributions and leadership from @Lizstuartdc @BrendanSaloner @AleneKH Sachini Bandara @Mark_Meiselbach @CatherineEttman Karen Shen @OliviaKSugarman @andrew_andersn Sabriya Linton, and many others. And of course thank you to our founding and former Center Directors @Lizstuartdc @colleenlbarry & @EmmaBethMcGinty for putting us on this path and the mentorship you provided all of us.
Onward and looking forward to generating evidence and making impact in the upcoming year! -
Tinglong Dai
@TinglongDai
Sports writing at its best
@meganmaurice @guardian: www.theguardian.com/sport/live/2024/aug/10/paris-2024-oly… -
Tinglong Dai
@TinglongDai
Thrilled to be joined by Prof. Tugce Martagan of @TUeindhoven, a leader in using operations research and artificial intelligence to advance biomanufacturing, for a panel discussion on Wed, Sept 4, 2024 at 11 am ET: event.on24.com/wcc/r/4657883/790B986C408F714D663D2DADD5CD…
Tugce's decade of award-winning collaboration with industry has carved out a new space for the OR/AI community to drive efficiency, safety, and innovation in biomanufacturing.
Together, Tugce, Dr. Deenadayalan Bakthavatsalam, and I will discuss the role of AI in shaping biomanufacturing, why OR is essential to making AI really work, and how regulators can help accelerate AI scaling.
The panel will be moderated by @TMenichiello, Editor of @BioProcessOL and @_CellandGene.
Register for the online panel here: event.on24.com/wcc/r/4657883/790B986C408F714D663D2DADD5CD… @INFORMS @JHUCarey @rBIO_pharma @JohnsHopkins -
Tinglong Dai
@TinglongDai
America the Beautiful
Tinglong Dai
@TinglongDai
"There is of course something unresolved and contradictory at the heart of the Biles arc, the Biles catharsis narrative. What hurt Biles was overexposure combined with losing. The way Biles can be fixed, sport tells us, is even more exposure combined with victory." - @barneyronay @TheGuadianNews: www.theguardian.com/sport/article/2024/jul/30/simone-bile…
-
Tinglong Dai reposted
François Chollet
@fchollet
The question of whether LLMs can reason is, in many ways, the wrong question. The more interesting question is whether they are limited to memorization / interpolative retrieval, or whether they can adapt to novelty beyond what they know. (They can't, at least until you start doing active inference, or using them in a search loop, etc.)
There are two distinct things you can call "reasoning", and no benchmark aside from ARC-AGI makes any attempt to distinguish between the two.
First, there is memorizing & retrieving program templates to tackle known tasks, such as "solve ax+b=c" -- you probably memorized the "algorithm" for finding x when you were in school. LLMs *can* do this! In fact, this is *most* of what they do. However, they are notoriously bad at it, because their memorized programs are vector functions fitted to training data, that generalize via interpolation. This is a very suboptimal approach for representing any kind of discrete symbolic program. This is why LLMs on their own still struggle with digit addition, for instance -- they need to be trained on millions of examples of digit addition, but they only achieve ~70% accuracy on new numbers.
This way of doing "reasoning" is not fundamentally different from purely memorizing the answers to a set of questions (e.g. 3x+5=2, 2x+3=6, etc.) -- it's just a higher order version of the same. It's still memorization and retrieval -- applied to templates rather than pointwise answers.
The other way you can define reasoning is as the ability to *synthesize* new programs (from existing parts) in order to solve tasks you've never seen before. Like, solving ax+b=c without having ever learned to do it, while only knowing about addition, subtraction, multiplication and division. That's how you can adapt to novelty. LLMs *cannot* do this, at least not on their own. They can however be incorporated into a program search process capable of this kind of reasoning.
This second definition is by far the more valuable form of reasoning. This is the difference between the smart kids in the back of the class that aren't paying attention but ace tests by improvisation, and the studious kids that spend their time doing homework and get medium-good grades, but are actually complete idiots that can't deviate one bit from what they've memorized. Which one would you hire?
LLMs cannot do this because they are very much limited to retrieval of memorized programs. They're static program stores. However, can display some amount of adaptability, because not only are the stored programs capable of generalization via interpolation, the *program store itself* is interpolative: you can interpolate between programs, or otherwise "move around" in continuous program space. But this only yields local generalization, not any real ability to make sense of new situations.
This is why LLMs need to be trained on enormous amounts of data: the only way to make them somewhat useful is to expose them to a *dense sampling* of absolutely everything there is to know and everything there is to do. Humans don't work like this -- even the really dumb ones are still vastly more intelligent than LLMs, despite having far less knowledge. -
Tinglong Dai reposted
NYTPitchbot
@DougJBalloon
Opinion | I said Biden should drop out. But now that he has, I think it was a mistake.
-
Tinglong Dai
@TinglongDai
From @jaycaspiankang in @NewYorker (June 30):
-
Tinglong Dai reposted
Selina Wang
@selinawangtv
Biden is standing down.
He says "I believe it is in the best interest of my party and the country for me to stand down and to focus solely on fulfilling my duties as President for the remainder of my term" -
Tinglong Dai reposted
Curt Langlotz
@curtlanglotz
Truth: "State of the art LLMs don't accurately diagnose patients (performing worse than physicians), don't follow guidelines, cannot interpret lab results, fail to follow instructions, and are sensitive to both the quantity and order of information." www.nature.com/articles/s41591-024-03097-1
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Nate Silver
@NateSilver538
Fully RETRACT and RESCIND, sometimes it seems like both parties are trying to throw this election.
-
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
We are thrilled to announce that Alex Triantis has been appointed to a second term as dean of Johns Hopkins Carey Business School.
hub.jhu.edu/2024/07/19/carey-dean-alex-triantis-reappoint… -
Tinglong Dai reposted
The Daily Show
@TheDailyShow
BREAKING: Sources close to President Biden say he has begun to consider the possibility of accepting the notion that the likelihood of him becoming aware that he may come around to the conclusion that the prospect of the circumstances surrounding the occurrence of him not being potentially able to consider the feasibility that
-
Tinglong Dai reposted
Edward Luce
@EdwardGLuce
J.D. Vance last night reversed his Hillbilly elegy thesis in which he blamed his blue collar communities for fatalism and bad cultural habits. In Milwaukee he portrayed them only as victims. That's his Trumpian U-turn. My JDV review from eight years ago. www.ft.com/content/614cff5e-7b68-11e6-ae24-f193b105145e
-
Tinglong Dai reposted
Nate Silver
@NateSilver538
It's just so weird living through this real-life Emperor Has No Clothes Moment.
He obviously shouldn't be president for 4 more years. Everyone knows this. It's incredibly revealing which people are willing to lie about it.
AG
@AGHamilton29
This is from Biden’s speech to the NAACP yesterday.
He loses track of what he’s saying constantly. Asking people to ignore what’s right in front of them is absurd. -
Tinglong Dai
@TinglongDai
I'm with @selina_cheng.
You don't fight for press freedom only when it's convenient or safe — that's never the case anyway (or anywhere)
Selina Cheng
@selina_cheng
Today I was fired from the Wall Street Journal.
This is my statement: -
Tinglong Dai
@TinglongDai
"OpenAI and Anthropic’s DIY approach meant that early adopters of LLMs disproportionately tended to be bad actors... Meanwhile the AI-in-your-face approach by Microsoft and Google has led to features that are occasionally useful and more often annoying."
Arvind Narayanan
@random_walker
Here's my thesis on what went wrong with generative AI from a business perspective and how that's changing now.
When ChatGPT launched, people found a thousand unexpected uses for it. This got AI developers overexcited. They completely misunderstood the market, underestimating the huge gap between proofs of concept and reliable products.
This misunderstanding led to two opposing but equally flawed approaches to commercializing LLMs.
OpenAI and Anthropic focused on building models and not worrying about products. For example, it took 6 months for OpenAI to bother to release a ChatGPT iOS app and 8 months for an Android app!
Google and Microsoft shoved AI into everything in a panicked race, without thinking about which products would actually benefit from AI and how they should be integrated.
Both groups of companies forgot the “make something people want” mantra. The generality of LLMs allowed developers to fool themselves into thinking that they were exempt from the need to find a product-market fit, as if prompting is a replacement for carefully designed products or features.
OpenAI and Anthropic’s DIY approach meant that early adopters of LLMs disproportionately tended to be bad actors, since they are more invested in figuring out how to adapt new technologies for their purposes, whereas everyday users want easy-to-use products. This has contributed to a poor public perception of the technology. (A point we make in the AI Snake Oil book: www.amazon.com/Snake-Oil-Artificial-Intelligence-Differen…)
Meanwhile the AI-in-your-face approach by Microsoft and Google has led to features that are occasionally useful and more often annoying. It also led to many self-owns due to inadequate testing like Microsoft's early Sydney chatbot and Google's Gemini image generator. This has also caused a backlash.
But things are changing. OpenAI and Anthropic seem to be transitioning from research labs focused on a speculative future to something resembling regular product companies. If you take all the human-interest elements out of the OpenAI boardroom drama, it was fundamentally about the company's shift from creating gods to building products.
Google and Microsoft are slower to learn, but my guess is that Apple will force them to change. Last year it was seen as a laggard on AI, but it seems obvious in retrospect that the slow and thoughtful approach that Apple showcased at WWDC is more likely to resonate with users.
Still, we shouldn't expect changes overnight. There are unsolved research challenges when it comes to making LLM-based AI assistants that actually work: www.aisnakeoil.com/p/new-paper-ai-agents-that-matter
Enterprise adoption will likely be even slower. the barriers include integrating it into existing products and workflows and training people to use it productively while avoiding its pitfalls. We should expect this to happen on a timescale of a decade rather than a year.
If you want clear analyses of AI that look past the hype, subscribe to the AI Snake Oil newsletter. www.aisnakeoil.com -
Tinglong Dai
@TinglongDai
Peggy Noonan on J.D. Vance, back in Feb. 2022:
-
Tinglong Dai
@TinglongDai
Wall Street isn't impressed :-)
-
Tinglong Dai reposted
Face The Nation
@FaceTheNation
Top Democratic sources believe that Democrats who had thoughts about challenging President Biden are now standing down "because of this fragile political moment," @CostaReports reports.
Concerns about Biden's debate performance "faded almost instantly" after the attempted assassination of Donald Trump, with Democratic sources say the party should stay cool at a turbulent time. -
Tinglong Dai reposted
David Altizer
@dvdaltizer
You can actually see Evan taking that iconic image during the rally coverage below.
Observe Evan's calm demeanor. This is a result of his extensive experience and unwavering confidence. It's as if he was born to frame and compose shots, a skill that has become his second nature.
I'm sure he saw the flag up there, waited for that moment, and initiated the shutter on his Sony.
It wouldn't have looked as good if he had been standing slightly to the right like the other photographers. He was in the perfect spot, and he knew it.
3/5 -
Tinglong Dai
@TinglongDai
.@benwallacewells can write:
-
Tinglong Dai
@TinglongDai
God Bless the USA 🇺🇸 May we use this moment to cool our rhetoric, recognize virtues like grace and courage, and unite as one people.
This media couldn't be displayed. See it on the original post. -
Tinglong Dai
@TinglongDai
Biden insists he is staying in the race
-
Tinglong Dai
@TinglongDai
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai
@TinglongDai
Sizzling🌡️
-
Tinglong Dai reposted
Mitra Taheri
@MitraLTaheri
Truly transformative. ..So proud to be part of the @JohnsHopkins community.
Tinglong Dai
@TinglongDai
In January 2013, @nytimes published an article titled "$1.1 billion in thanks from Bloomberg to Johns Hopkins": www.nytimes.com/2013/01/27/nyregion/at-1-1-billion-bloomb…
More than a decade has passed, @MikeBloomberg ('64) has given several more billions to @JohnsHopkins, making undergraduate admissions need-blind, launching perhaps academia's most ambitious effort to use AI and data science to transform business and society, positioning Johns Hopkins in the most central location in the nation's capital (@JHUBloombergCtr), and elevating Hopkins in countless ways, both visible and subtle.
This morning is another $1 billion thank you from @MikeBloomberg (hub.jhu.edu/2024/07/08/johns-hopkins-investment-financial…).
Through this gift, medical education at @HopkinsMedicine will be free for families earning less than $300,000, and graduate programs across the university, including @JHUCarey , will be able to attract the best and brightest to Johns Hopkins and change the world for the better. -
Tinglong Dai
@TinglongDai
In January 2013, @nytimes published an article titled "$1.1 billion in thanks from Bloomberg to Johns Hopkins": www.nytimes.com/2013/01/27/nyregion/at-1-1-billion-bloomb…
More than a decade has passed, @MikeBloomberg ('64) has given several more billions to @JohnsHopkins, making undergraduate admissions need-blind, launching perhaps academia's most ambitious effort to use AI and data science to transform business and society, positioning Johns Hopkins in the most central location in the nation's capital (@JHUBloombergCtr), and elevating Hopkins in countless ways, both visible and subtle.
This morning is another $1 billion thank you from @MikeBloomberg (hub.jhu.edu/2024/07/08/johns-hopkins-investment-financial…).
Through this gift, medical education at @HopkinsMedicine will be free for families earning less than $300,000, and graduate programs across the university, including @JHUCarey , will be able to attract the best and brightest to Johns Hopkins and change the world for the better. -
Tinglong Dai
@TinglongDai
Baltimore's 4th of July celebration - A drone performance of the USA map
Photos taken on the 24th floor of @JHUCarey -
Tinglong Dai
@TinglongDai
Wow
-
Tinglong Dai reposted
Joyce Carol Oates
@JoyceCarolOates
Q. What do you have to say in response to the fact that your situation is hopeless, Mr. President?
A. I don't feel that my situation is hopeless.
Q. What do you have to say in response to the fact that all the polls have you losing?
A. I don't believe the polls.
Q. What do you have to say in response to the fact that increasing numbers of people are saying that you are "failing" & that you are not the person you were in 2020?
A. I don't believe ....(etc.)
Q. What do you have to say...(etc.)
A. I don't believe...(etc.)
Q. What do...(etc.)
A. I don't... (etc.) -
Tinglong Dai reposted
Marketing Science
@Mrktng_Science
Articles in Advance 06/24 (4 of 7)
"Overdiagnosis and Undertesting for Infectious Diseases"
by Tinglong Dai, Shubhranshu Singh
@TinglongDai
pubsonline.informs.org/doi/abs/10.1287/mksc.2022.0038 -
Tinglong Dai
@TinglongDai
Dear @POTUS team, if Joe Biden really is like Yoda, shouldn't it be "Like Yoda, Biden is"?
Start talking that way, Joe Biden should.
Much more effective, he will become. -
Tinglong Dai reposted
Ashley Winter MD || Urologist
@AshleyGWinter
I am begging people to understand that almost every test on the planet has something called "false positives," which means a risk that a result of the test is WRONG. A WRONG result can lead to necessary fear, worry, and at worst, unnecessary INVASIVE procedures or even unnecessary surgery. The chance of someone getting a false positive goes down when the person is at a higher risk of having an actual problem (ie. A specific SYMPTOM). This is called a higher "pretest probability." This is why doctors don't recommend everyone get every test all the time.
-
Tinglong Dai reposted
Sergey Radchenko
@DrRadchenko
This election is doing more to discredit American democracy than Putin and Xi Jinping could ever hope to. I am worried about the image projected to the outside world. It is not an image of leadership. It is an image of terminal decline.
-
Tinglong Dai
@TinglongDai
Fairy Chimneys
-
Tinglong Dai
@TinglongDai
Rose Valley
-
Tinglong Dai reposted
Loren Balhorn
@fraubalhorn
Had dinner with two retired historians from East Germany yesterday, both of whom expressed supreme confidence that the United States has entered its “Soviet phase”. Hard to disagree.
-
Tinglong Dai
@TinglongDai
Not your usual hotel check-in experience
-
Tinglong Dai
@TinglongDai
Göreme
-
Tinglong Dai
@TinglongDai
Artificial intelligence could have done better in this debate
-
Tinglong Dai
@TinglongDai
Dolmabahçe Palace
-
Tinglong Dai
@TinglongDai
Istanbul Modern
-
Tinglong Dai reposted
Zak Kukoff
@zck
LLMs are commodities. The real AI money is in professional services
-
Tinglong Dai
@TinglongDai
Bosphorus
-
Tinglong Dai
@TinglongDai
Where Asia meets Europe
-
Tinglong Dai
@TinglongDai
Public transportation in Istanbul reminds me of what we could have in major U.S. cities
Tinglong Dai
@TinglongDai
Here's another one from Fatih, one of the busiest areas lined with restaurants and shops
-
Tinglong Dai
@TinglongDai
You don't have to be a clock lover to fall in love with clockwork at Topkapı Palace
-
Tinglong Dai
@TinglongDai
Istanbul
-
Tinglong Dai
@TinglongDai
Manchester
-
Tinglong Dai
@TinglongDai
York
-
Tinglong Dai
@TinglongDai
A @ChatsworthHouse bookcase inspired by the Fibonacci sequence
-
Tinglong Dai
@TinglongDai
Chatsworth House
-
Tinglong Dai
@TinglongDai
London
-
Tinglong Dai
@TinglongDai
Lake District
-
Tinglong Dai
@TinglongDai
Liverpool
-
Tinglong Dai
@TinglongDai
Reykjavík
-
Tinglong Dai
@TinglongDai
Greetings from Iceland 🇮🇸
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
Helena Bottemiller Evich
@hbottemiller
Wow: @OIGatHHS just released its probe of FDA's handling of the infant formula debacle and basically confirmed all the big stories I broke back in 2022 at @politico.
-FDA did not heed multiple safety warnings
-The agency's structure was a major problem
oig.hhs.gov/reports-and-publications/all-reports-and-publ… -
Tinglong Dai reposted
Louise Schiavone
@LouiseSchiavone
Covered the U-S Senate from inside that beautiful building for 20 years. Appreciated it every single day.
-
Tinglong Dai
@TinglongDai
DC
-
Tinglong Dai reposted
Johns Hopkins University
@JohnsHopkins
#icymi Here's a recap of Monday's #discoveryjhbc series event. washingtondc.jhu.edu/upcoming-events/series-forums/discov…
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
. @OpenAI CTO @miramurati and @AZSecretary @GaSecofState weigh in on challenges, opportunities and threats by #ai on at interview with @karaswisher at #hopkinsbloombergcenter.
Dive into what was discussed in the first #discoveryjhbc series event here: washingtondc.jhu.edu/upcoming-events/series-forums/discov… -
Tinglong Dai
@TinglongDai
.@Princeton professor @random_walker speaks at the closing panel of the @WhiteHouse conference on AI Aspirations held at @JHUBloombergCtr @JHUCarey @SAISHopkins: “We just don't see the justification for changing this norm of openness and for seeing these models is so dangerous. There are many other arguments such as national competitiveness, are we allowing other countries to build on our research but not give their research back and turn? These are arguments we've debated for decades, we've debated them during the Cold War, and we came down strongly on the side of openness, nothing about that has really changed today. I think we should keep revisiting this question. If the facts change, we should change our minds. But as things stand today, I think we should default strongly to to continuing this norm of openness especially when we consider public sector uses of AI. It's good for science, it's good for innovation, and democratizes access to AI. It makes the market more competitive. Yes, there are a few risks, but we think those risks are minimal. And there are many other ways to manage those risks, as opposed to putting AI back in the bottle.”
Tinglong Dai
@TinglongDai
On AI and science, @random_walker of @Princeton says, “And I think the root cause here is scientists often having a tendency to treat the AI model itself as the store of knowledge as opposed to a tool for a scientific discovery for human understanding.”
Tinglong Dai
@TinglongDai
Here is the corrected version for better clarity:
"As things stand now, based on our evidence review, we don't see the justification for changing this norm of openness or viewing these models as inherently dangerous. There are other arguments, such as national competitiveness and whether we are allowing other countries to build on our research without reciprocating. These arguments have been debated for decades, including during the Cold War, and we consistently sided with openness. Nothing has fundamentally changed today.
"We should continue revisiting this question and be open to changing our stance if the facts change. However, as things stand today, we should strongly default to maintaining this norm of openness, especially when considering public sector uses of AI. Openness is beneficial for science, innovation, and democratizing access to AI. It also makes the market more competitive. While there are a few risks, we believe they are minimal and manageable through other means, rather than restricting access to AI." -
Tinglong Dai
@TinglongDai
On the value of openness in AI research, and why people have different views on the subject, from @Princeton's insightful @random_walker:
"The aspiration I want to focus on is AI for healthcare, drug development, and scientific discovery more broadly. The value I want to highlight is openness, a topic that hasn't been discussed much today. By openness, I mean what you do with all the data and the models you've invested in and trained. Do you keep them proprietary, or do you make them available to the public, allowing anyone to download, use, and modify them for their purposes?
"This issue has been fiercely controversial in the tech industry over the past year or two. There are many ways to navigate this controversy, but one core reason for differing opinions is how people perceive AI. Do they see it as a product or as a repository of scientific knowledge and a tool for advancing science? AI models can be seen as both.
"As products, AI models are akin to iPhones, which we accept as proprietary. However, AI models can also be viewed as a way to access knowledge, similar to the internet. In the scientific community, there is a strong norm of openness, which has enabled AI to advance significantly through collaborative efforts."
Tinglong Dai
@TinglongDai
.@Princeton professor @random_walker speaks at the closing panel of the @WhiteHouse conference on AI Aspirations held at @JHUBloombergCtr @JHUCarey @SAISHopkins: “We just don't see the justification for changing this norm of openness and for seeing these models is so dangerous. There are many other arguments such as national competitiveness, are we allowing other countries to build on our research but not give their research back and turn? These are arguments we've debated for decades, we've debated them during the Cold War, and we came down strongly on the side of openness, nothing about that has really changed today. I think we should keep revisiting this question. If the facts change, we should change our minds. But as things stand today, I think we should default strongly to to continuing this norm of openness especially when we consider public sector uses of AI. It's good for science, it's good for innovation, and democratizes access to AI. It makes the market more competitive. Yes, there are a few risks, but we think those risks are minimal. And there are many other ways to manage those risks, as opposed to putting AI back in the bottle.”
-
Tinglong Dai
@TinglongDai
Director @WHOSTP thanks many people for making the @WhiteHouse Conference on AI Aspirations possible, including the incredible team at @JHUBloombergCtr @JHUCarey @SAISHopkins
-
Tinglong Dai
@TinglongDai
President of @CenDemTech, @AlexReeveGivens, speaks at the @WhiteHouse conference on AI Aspirations held at @JHUBloombergCtr @JHUCarey @SAISHopkins: “What's important in this conversation [about trust in AI] is that sometimes it's easy to be a little dismissive of that concern, to say, oh, people don't really trust the technology, and it makes you think you just need to explain it a little bit better. And you'll bring them along for the ride. But if you look at history, the lack of trust is often there for a reason.”
Tinglong Dai
@TinglongDai
@CenDemTech’s @AlexReeveGivens argues that being proactive about risk mitigation is also efficient: “So I think there are concrete things that agencies can do now to kind of push this earlier in the stack, give guidance to people across the agency to do this work, and just surface awareness of the need to do this kind of risk mitigation strategy very early on, which is important for your mission. It's important to protecting the American people. And it was also just more efficient because you're thinking about those considerations early on, rather than being compliance issues that you have to think about later in the process.”
-
Tinglong Dai
@TinglongDai
“Supply chain management is completely changed around the world right now because of artificial intelligence.” Said @RepDonBeyer at the @WhiteHouse conference on AI Aspirations held at @JHUBloombergCtr @JHUCarey
-
Tinglong Dai
@TinglongDai
“With the arrival of AI…”
“Imagine a world in which AI makes…”
A lot of religious-ish sentences -
Tinglong Dai
@TinglongDai
U.S. @OMBPress Director @ShalandaYoung46 speaks at @WhiteHouse conference on AI aspirations @JHUBloombergCtr @JHUCarey: "Government actually cares that people like their experience with government ... we have a long way to go to improve a long-standing reputation that people don't like dealing with government."
Tinglong Dai
@TinglongDai
From @ShalandaYoung46: “And to me the worst thing is people losing faith in their government when they have served. They know they have veterans benefits they know they should be getting medical care and they just decided to give up.”
Tinglong Dai
@TinglongDai
“"Are we going to limit the use of AI? This is actually an opportunity for us to embrace the exciting changes and show that this power can be used for good, especially for services. But it also reminds me if we get it wrong and we don't marry the need to improve government services and use AI to do that in a way that is beneficial to people. Make sure it doesn't bring biases into systems.”
-
Tinglong Dai
@TinglongDai
Senator @MarkWarner speaks at the @WhiteHouse conference on AI Aspirations @JHUBloombergCtr @JHUCarey: “I've spent about as much time with AI experts as most people and let me acknowledge, at least for me, there is no linear relationship between the time I spend talking to AI experts and me actually getting smarter on the subject.”
Tinglong Dai
@TinglongDai
“We don't take the the argument that I heard from all of my tech bros in the social media, which just says, Oh, you guys in policy, you'll never understand. Let us go do this. And we'll come back we can figure out the rules after the fact.”
Tinglong Dai
@TinglongDai
Senator @MarkWarner: “I think making sure that those of you who are coming from the research community can have that kind of unfettered access to some of these models without the liability provisions is something that's critically important. as well.”
-
Tinglong Dai
@TinglongDai
Senator @amyklobuchar speaks at @WhiteHouse conference on AI Aspirations @JHUBloombergCtr @JHUCarey. The warmest applause yet from the audience.
Tinglong Dai
@TinglongDai
And laughter too - more than all the others combined - so far
Tinglong Dai
@TinglongDai
Senator @amyklobuchar made an insightful argument that needs listening to carefully: “If a power grid wants to use AI to improve efficiency, the vendor providing the AI system would have to undergo security testing before selling the product to utilities. The second part is that it creates a pathway for informed, thoughtful policymaking in the years to come. And it increases transparency by requiring companies using what we call high-impact AI systems, which affect things like access to health care, housing, and employment, to first submit to NIST.”
-
Tinglong Dai
@TinglongDai
Can AI help @US_FDA better inspect drug manufacturing facilities around the world? @DrCaliff_FDA thinks it's a yes at the @WhiteHouse Conference on AI aspirations @JHUBloombergCtr @JHUCarey
-
Tinglong Dai
@TinglongDai
The most breathtaking bar chart race of all time:
This media couldn't be displayed. See it on the original post. -
Tinglong Dai
@TinglongDai
In my op-ed for @FederalNewsNet on behalf of the @INFORMS community (federalnewsnetwork.com/commentary/2024/06/to-make-effecti…), I spotlight the @WhiteHouse's new AI guidelines. These guidelines call for chief artificial intelligence officers and chief data officers to be part of AI policymaking.
AI has been around for 70 years, but the rise of generative AI — sparked by the launch of ChatGPT in November 2022 — has made it a real thing for many. Naturally, some are now jockeying for control over it. This can lead to experts being crowded out.
Why does this matter? While AI offers real benefits, it also poses real risks. In high-stakes areas like healthcare, unproven or untested AI can cause serious harm to patients.
AI safety comes from many things, but it's safe to say there is no substitute for involving data scientists in AI policy and decision-making.
Read more: federalnewsnetwork.com/commentary/2024/06/to-make-effecti…
@egerstmann @INFORMS @JHUBloombergCtr @JHUCarey @JHU_HBHI -
Tinglong Dai reposted
Port of Baltimore
@portofbalt
“I’ve been waiting a long time to say this…Maryland, the Fort McHenry Channel is cleared and the Port of Baltimore is open for business!” -@GovWesMoore
-
Tinglong Dai reposted
François Chollet
@fchollet
Search engines used to be regarded as "AI" not long ago. It wouldn't be surprising if interpolative databases (foundation models) stop being regarded as AI in the future, too. They're more a static data structure -- an embedded dataset -- than a cognitively capable agent.
-
Tinglong Dai reposted
-
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
Carey Professor @TinglongDai shares his optimism that port traffic levels will recover quickly after closing due to the Key Bridge collapse. "Location, location, location," he said. "Baltimore is central."
www.baltimoresun.com/2024/06/10/port-baltimore-traffic-ke… -
Tinglong Dai
@TinglongDai
Today in @MedEconomics, Dr. @MichaelAbramoff and I write about the need for a new science to address the growing need to scale medical AI.
The article (www.medicaleconomics.com/view/toward-a-science-of-scaling…) draws on two of our recent publications in @npjDigitalMed and @NEJM_AI.
We discuss our first-of-its-kind randomized controlled trial with @OrbisIntl, which shows a 40% increase in clinical productivity using autonomous AI for diabetic eye exams, underscoring AI's potential to transform resource-constrained healthcare.
We also highlight sustainable models that incentivize broader adoption and continuous improvement, ensuring equitable access to the benefits of AI.
Read the full text: www.medicaleconomics.com/view/toward-a-science-of-scaling… -
Tinglong Dai reposted
Shengwu Li
@ShengwuLi
Do what works for you. (Anthony's procedure works for him!)
But during the tenure track I did not work 60-70 hour weeks. 4-5 hours of focused work (proofs, writing) is a great day. The rest is scaffolding to make those hours happen.
alz
@alz_zyd_
- Empirically for me I feel I was working really, really hard the first 3-4 ish years, like 60-70 hour weeks mostly, and now workload has come down somewhat. It's like grad school, with more projects, and more other responsibilities (teaching, conferences, discussions, etc.)
-
Tinglong Dai reposted
The Conversation U.S.
@ConversationUS
While tariffs don’t have the greatest track record, an expert in global supply chain explains why they believe Biden’s recent 100% tariff on Chinese-made electric vehicles could succeed in giving the U.S. EV industry room to grow:
By @TinglongDai:
theconversation.com/history-says-tariffs-rarely-work-but-… -
Tinglong Dai reposted
Melinda B. Buntin
@MelindaBBuntin
Glad to be one of many giving constructive feedback to @CMSGov
about their #datapolicy with @jzayanian @JAMAHealthForum nam02.safelinks.protection.outlook.com?url=https%3A%2F%2F… -
Tinglong Dai reposted
Daniel Polsky
@healthecon_dan
Don't miss the @tradeoffspod on the latest innovation in nurse-led preventive primary care.
I'm proud to be associated with the inspiring leadership of @JHUNursing and @Sarah_Szanton as they push the envelope. Looking forward to engaging on ideas for sustainability. @JHU_HBHI
Tradeoffs
@tradeoffspod
This week, we get an inside look at an ambitious attempt to pioneer a new form of primary care and the barriers that could derail this dream.
Listen or read for more ⬇️ tradeoffs.org/2024/06/06/neighborhood-nursing-costa-rica-… -
Tinglong Dai
@TinglongDai
Baltimore on a rainy afternoon
-
Tinglong Dai
@TinglongDai
The @INFORMS @informshas is now accepting applications for the 2024 Sanjay and Panna Mehrotra Research Excellence Award. I am thrilled to serve on this year's award committee, along with Profs. Susan (Feng) Lu @Purdue and Sridhar Seshadri @UofIllinois
This prestigious award recognizes a mid-career researcher for their impactful contributions to the health applications of Operations Research and Management Sciences.
Applications are due by July 19, 2024.
For more details and to apply, please visit: connect.informs.org/healthapplications/awards/sanjay-pann… -
Tinglong Dai reposted
Liron Shapira
@liron
Helen Toner (@hlntnr), former OpenAI board member, is finally going on record about firing Sam Altman:
😱 “He gave us inaccurate information about the small number of formal safety processes that the company did have in place.”
😱 “When ChatGPT came out, the board was not informed in advance about that. We learned about ChatGPT on Twitter.”
😱 “Sam didn't inform the board that he owned the OpenAI startup fund.”
My two cents:
This was inexcusable behavior on Sam's part, the old board was right to fire him, and the new board should fire him again.This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
US universities ranked by number of "best scientists" (h index>155). Some under- and some overrated unis?
Harvard 69
MIT 21
Johns Hopkins 20
Stanford 20
UCSD 19
Caltech 13
Berkeley 11
Columbia 10
Yale 10
Cornell 8
Penn 8
Princeton 5
U. Michigan 5
U. Wisconsin 5
U. Illinois 2 -
Tinglong Dai
@TinglongDai
The final results are in:
Tinglong Dai
@TinglongDai
Who do you trust more?
-
Tinglong Dai
@TinglongDai
Dissent is central to any democracy. — Harry Belafonte
-
Tinglong Dai
@TinglongDai
New York
-
Tinglong Dai
@TinglongDai
Leaking window 漏窗 - the ancient Chinese version of Instagram filters @metmuseum
-
Tinglong Dai
@TinglongDai
Grand Central
-
Tinglong Dai
@TinglongDai
Jersey City
-
Tinglong Dai reposted
Besmira Nushi 💙💛
@besanushi
Honest question: What is up with the large non-determinism in the output answer of the @OpenAI calls, even when temperature is set to 0, and the seed is fixed? This is particularly true for long generation tasks. Has anyone found a way to deal with this other than running the prompt several times and averaging out?
-
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
Researchers from @JohnsHopkins have received a grant to create a "first-of-its-kind" drug supply chain dashboard.
The grant is part of the Nexus Awards, a university-wide initiative anchored at the #hopkinsbloombergcenter
Learn more: www.beckershospitalreview.com/pharmacy/with-300k-johns-ho… -
Tinglong Dai reposted
Port of Baltimore
@portofbalt
Our vessel schedule is filling up!
More than 30 vessels and barges are due into our public terminals over the next week.
Baltimore is back! @ILAUnion -
Tinglong Dai
@TinglongDai
Who do you trust more?
-
Tinglong Dai reposted
Arjun (Raj) Manrai
@arjunmanrai
2/ Taylor went on The Eras Tour in 2023-2024. James went on the ERAS® tour in 2023-2024.
-
Tinglong Dai
@TinglongDai
Honored and thrilled to receive the prestigious @JohnsHopkins Nexus Award, with $300,000 in funding, to develop a first-of-its-kind Johns Hopkins Drug Supply Chain Data Dashboard, along with @JHU_HBHI friends and colleagues from our business, medical, and public health schools: Profs. @dr_msocal, Gerard Anderson, @jeromie_b, Mac Dada, Jeremy Greene, and @drJoshS.
Our dashboard, one of only four Nexus Awards in the research category (hub.jhu.edu/2024/05/21/nexus-awards-program-hopkins-washi…), will provide timely insights to tackle drug shortages and supply chain disruptions. Using AI/LLMs and data integration techniques, our interdisciplinary team will deliver this tool to support informed public health policy at the dynamic @JHUBloombergCtr, near key global decision-making hubs.
We are grateful for the support of @JHU_HBHI, @jhucarey, @JohnsHopkinsSPH, and @HopkinsMedicine in laying all the groundwork.
Part of this effort will include a series of conversations with stakeholders in Washington DC and beyond in the coming months. Stay tuned! -
Tinglong Dai
@TinglongDai
Had a blast with my wonderful @JHUCarey MBA graduates and their families:
www.linkedin.com/pulse/tortured-poets-quants-department-t…
Thank you to Deepika Satyadev for the picture. -
Tinglong Dai reposted
The Transcript
@TheTranscript_
Jensen walking into the $NVDA Earnings Call at the top of the hour.
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Mariana Socal
@dr_msocal
#graduation is finally here - Celebrating the @JohnsHopkinsSPH class of 2024. So proud of my many master’s and doctoral students , who I’m grateful to now have as colleagues 🎓🎉 go team #publichealth go!
-
Tinglong Dai reposted
Anna Nagurney
@Supernetworks
An OUTSTANDING MBA commencement address by Professor @TinglongDai to the @JohnsHopkins MBA Class of 2024 with Emily Dickinson, great poet from our very own @Townofamherst, quoted. I am printing this one out and framing it.
Tinglong Dai
@TinglongDai
The Tortured Poets (and Quants) Department
Address to the @JohnsHopkins MBA Class of 2024 and Families
Tinglong Dai, Bernard T. Ferrari Professor of Business
May 20, 2024
Dear Class of 2024 Graduates and Families, Faculty, Staff, Deans, and Friends,
Four months ago, I stood on this very podium and accepted the Bernard T. Ferrari Professorship from President Ron Daniels. It was a magical evening, surrounded by so many friends and colleagues — and my family.
The next morning, a snowstorm shut down the entire Baltimore area. I was in a celebratory mood, so the snow was absolutely beautiful and only added to the magic.
Since the Ferrari professorship ceremony, people have kept asking me, "How has your life changed with your Ferrari?" My usual response is, "I LOVE it — my only complaint is that it's just too fast!"
"Can I borrow your Ferrari?" The answer is always, "Of course!" It's easy to be generous when you don't actually own a powerful, blazing, exhilarating race car.
Speaking of racing, let me tell you a story about a race between two strangers.
In 1999, I was a high school senior in China, attending school away from my parents. My tiny village was 15 miles away, and every two weeks or so, I'd hop on my bike and ride home.
On a frigid Saturday in December, just days after a blizzard had turned the world into a frozen wonderland, I set out on my usual route. It was about 10 degrees Fahrenheit, cold enough to freeze your breath in the air. The road was surprisingly clear, and I pedaled along, bundled up but still shivering.
Out of nowhere, this guy on a bike appeared, matching my pace. Then he accelerated, leaving me behind. My teenage pride couldn't take it, so I pedaled harder and passed him.
This impromptu race became an unspoken war. He accelerated, then I accelerated, neither of us willing to be the loser in this spontaneous race.
Twenty minutes later, I was riding like a maniac. The cold was biting, and my legs felt like they were on fire, but sheer stubbornness and the fear of losing face to this random stranger drove me on.
I began to sing to keep myself going — a melody that I'm sure sounded as desperate as I felt.
Just as I was savoring my small victory, he rocketed past me again. His legs were a blur of motion, like a human cyclone on wheels.
In my desperation, I pushed myself beyond my limits, pedaling and singing louder, as if sheer willpower could warm me up. But it didn't.
The next thing I knew, the world went black.
When I came to, it was eight hours later, and I was lying in a clinic just off the highway.
A young couple who ran the roadside clinic had found me unconscious on the highway outside their door. They nursed me back from the brink of severe hypothermia. Without them, my foolish, pride-fueled race might have cost me my life.
Thanks to their kindness, I'm here today to tell you this story — and two important life lessons I learned from it.
Lesson #1: Location. Location. Location. If you collapse while riding your bike in a freezing winter, make sure you collapse in front of a clinic. Distance matters for health outcomes.
Lesson #2: As much as you enjoy competitions, make sure you are in the right race. Not every race is worth doing, and some can cost you your life.
This brings me to a broader point about life and death.
Now, at least 60% of you have a significant interest in healthcare. So, I suppose it’s okay to talk about death, even at this moment of celebration, without making too many people uncomfortable.
You see, my last name is D-A-I. For years, people have playfully and endearingly called me "Dr. Die.” It’s hard to be a medical doctor with a name like that, but it’s okay to be a doctor who doesn’t treat patients.
Emily Dickinson, one of the greatest poets in human history, wrote:
Because I could not stop for Death —
He kindly stopped for me —
The Carriage held but just
Ourselves — And Immortality.
Emily Dickinson was obsessed with death throughout her life. She wasn’t alone. Charlie Munger, one of the greatest business thinkers who passed away last year at the age of 99, once said, “ALL I want to know is where I'm going to die, so I'll never go there.”
Our lives are finite, and this realization should push us to leave a lasting impact. Live like you’re dying — fearlessly and fully.
Too often, we are tortured by the fear of being too good or too sharp or too fast — who am I to shine? Who am I to play big? Who am I to change the world?
The real question is, knowing our life is so short and we only live once, who are you not to shine? Who are you not to play big? Who are you not to change the world?
In the last two years I have lost two friends, both in their 50s. Not to mention the COVID-19 pandemic and the ongoing wars around the world that have normalized death in ways that are more frightening than we realize.
As I grieved for these losses, I remembered losing my grandfather when I was about to start college. He passed away at the age of 62. I still dream about him from time to time, even today. He would have been 86 this year.
When I first came to the United States, I noticed that people, especially in business school, talked about death all the time. “My laptop just died.” “He killed it.” “Wow, she has a killer CV.” But the ultimate proof is in the ubiquitous deadlines. The deadline for this, the deadline for that. What's the drop-dead deadline? Can we extend the deadline?
At first I thought, "Wow, these business people are really morbid! Why are they so obsessed with death?" But then I realized it's actually about keeping everyone on their toes. You see, in business, time is money. But in life, money isn’t time — and money can’t buy time. And money certainly can't buy life. So deadlines are a way of keeping us all alive, even if they sound a bit morbid.
Our lives have a deadline, it’s about 3,000 weeks from now, and it's very easy to forget that. Live fully, live fearlessly, and live as if we're dying — because we are.
Live like water.
@JHUCarey is on the water. Water is how we communicate with people decades, centuries before us who are no longer with us. Water is how we communicate with people centuries after us, after we're no longer with them. Water is history — and the future. Water is the perfect setting for a school like ours that builds for what's next.
And water is diverse and unifying. It’s everywhere and it’s the most fundamental feature of our planet Earth. It flows through every culture and brings people together. Look around — our class embodies this diversity, with representatives from Bangladesh, China, Colombia, the Dominican Republic, Ethiopia, Ghana, India, Indonesia, Jamaica, Japan, Kazakhstan, Nigeria, Pakistan, Rwanda, Saudi Arabia, Sri Lanka, and Türkiye. We even have people from the United States! We’re united in water.
When we live like water, we release our fears and doubts. We find our way. We find clarity and purpose.
We sing. We dance. We hug. We fight. We celebrate. We love. We flow.
We live in the serenity and resilience of water. We change the world with the grace and power of water.
Class of 2024 Johns Hopkins MBAs, you are truly exceptional. I am so proud of you. Let’s face it: this has been a very tough program, and yet you have made it through all the tortures.
Congratulations to you and your families, My Beloved — and Tortured — Poets and Quants! Now, for the next few thousand weeks — Go Change the World! -
Tinglong Dai reposted
Noushin Yahyavi, MD, MBA
@nyahyavi
Beautifully said @TinglongDai.
“You see, in business, time is money. But in life, money isn’t time — and money can’t buy time. And money certainly can't buy life.”
Live your life to the fullest and go change the world! @JHUCarey @JohnsHopkins
Tinglong Dai
@TinglongDai
The Tortured Poets (and Quants) Department
Address to the @JohnsHopkins MBA Class of 2024 and Families
Tinglong Dai, Bernard T. Ferrari Professor of Business
May 20, 2024
Dear Class of 2024 Graduates and Families, Faculty, Staff, Deans, and Friends,
Four months ago, I stood on this very podium and accepted the Bernard T. Ferrari Professorship from President Ron Daniels. It was a magical evening, surrounded by so many friends and colleagues — and my family.
The next morning, a snowstorm shut down the entire Baltimore area. I was in a celebratory mood, so the snow was absolutely beautiful and only added to the magic.
Since the Ferrari professorship ceremony, people have kept asking me, "How has your life changed with your Ferrari?" My usual response is, "I LOVE it — my only complaint is that it's just too fast!"
"Can I borrow your Ferrari?" The answer is always, "Of course!" It's easy to be generous when you don't actually own a powerful, blazing, exhilarating race car.
Speaking of racing, let me tell you a story about a race between two strangers.
In 1999, I was a high school senior in China, attending school away from my parents. My tiny village was 15 miles away, and every two weeks or so, I'd hop on my bike and ride home.
On a frigid Saturday in December, just days after a blizzard had turned the world into a frozen wonderland, I set out on my usual route. It was about 10 degrees Fahrenheit, cold enough to freeze your breath in the air. The road was surprisingly clear, and I pedaled along, bundled up but still shivering.
Out of nowhere, this guy on a bike appeared, matching my pace. Then he accelerated, leaving me behind. My teenage pride couldn't take it, so I pedaled harder and passed him.
This impromptu race became an unspoken war. He accelerated, then I accelerated, neither of us willing to be the loser in this spontaneous race.
Twenty minutes later, I was riding like a maniac. The cold was biting, and my legs felt like they were on fire, but sheer stubbornness and the fear of losing face to this random stranger drove me on.
I began to sing to keep myself going — a melody that I'm sure sounded as desperate as I felt.
Just as I was savoring my small victory, he rocketed past me again. His legs were a blur of motion, like a human cyclone on wheels.
In my desperation, I pushed myself beyond my limits, pedaling and singing louder, as if sheer willpower could warm me up. But it didn't.
The next thing I knew, the world went black.
When I came to, it was eight hours later, and I was lying in a clinic just off the highway.
A young couple who ran the roadside clinic had found me unconscious on the highway outside their door. They nursed me back from the brink of severe hypothermia. Without them, my foolish, pride-fueled race might have cost me my life.
Thanks to their kindness, I'm here today to tell you this story — and two important life lessons I learned from it.
Lesson #1: Location. Location. Location. If you collapse while riding your bike in a freezing winter, make sure you collapse in front of a clinic. Distance matters for health outcomes.
Lesson #2: As much as you enjoy competitions, make sure you are in the right race. Not every race is worth doing, and some can cost you your life.
This brings me to a broader point about life and death.
Now, at least 60% of you have a significant interest in healthcare. So, I suppose it’s okay to talk about death, even at this moment of celebration, without making too many people uncomfortable.
You see, my last name is D-A-I. For years, people have playfully and endearingly called me "Dr. Die.” It’s hard to be a medical doctor with a name like that, but it’s okay to be a doctor who doesn’t treat patients.
Emily Dickinson, one of the greatest poets in human history, wrote:
Because I could not stop for Death —
He kindly stopped for me —
The Carriage held but just
Ourselves — And Immortality.
Emily Dickinson was obsessed with death throughout her life. She wasn’t alone. Charlie Munger, one of the greatest business thinkers who passed away last year at the age of 99, once said, “ALL I want to know is where I'm going to die, so I'll never go there.”
Our lives are finite, and this realization should push us to leave a lasting impact. Live like you’re dying — fearlessly and fully.
Too often, we are tortured by the fear of being too good or too sharp or too fast — who am I to shine? Who am I to play big? Who am I to change the world?
The real question is, knowing our life is so short and we only live once, who are you not to shine? Who are you not to play big? Who are you not to change the world?
In the last two years I have lost two friends, both in their 50s. Not to mention the COVID-19 pandemic and the ongoing wars around the world that have normalized death in ways that are more frightening than we realize.
As I grieved for these losses, I remembered losing my grandfather when I was about to start college. He passed away at the age of 62. I still dream about him from time to time, even today. He would have been 86 this year.
When I first came to the United States, I noticed that people, especially in business school, talked about death all the time. “My laptop just died.” “He killed it.” “Wow, she has a killer CV.” But the ultimate proof is in the ubiquitous deadlines. The deadline for this, the deadline for that. What's the drop-dead deadline? Can we extend the deadline?
At first I thought, "Wow, these business people are really morbid! Why are they so obsessed with death?" But then I realized it's actually about keeping everyone on their toes. You see, in business, time is money. But in life, money isn’t time — and money can’t buy time. And money certainly can't buy life. So deadlines are a way of keeping us all alive, even if they sound a bit morbid.
Our lives have a deadline, it’s about 3,000 weeks from now, and it's very easy to forget that. Live fully, live fearlessly, and live as if we're dying — because we are.
Live like water.
@JHUCarey is on the water. Water is how we communicate with people decades, centuries before us who are no longer with us. Water is how we communicate with people centuries after us, after we're no longer with them. Water is history — and the future. Water is the perfect setting for a school like ours that builds for what's next.
And water is diverse and unifying. It’s everywhere and it’s the most fundamental feature of our planet Earth. It flows through every culture and brings people together. Look around — our class embodies this diversity, with representatives from Bangladesh, China, Colombia, the Dominican Republic, Ethiopia, Ghana, India, Indonesia, Jamaica, Japan, Kazakhstan, Nigeria, Pakistan, Rwanda, Saudi Arabia, Sri Lanka, and Türkiye. We even have people from the United States! We’re united in water.
When we live like water, we release our fears and doubts. We find our way. We find clarity and purpose.
We sing. We dance. We hug. We fight. We celebrate. We love. We flow.
We live in the serenity and resilience of water. We change the world with the grace and power of water.
Class of 2024 Johns Hopkins MBAs, you are truly exceptional. I am so proud of you. Let’s face it: this has been a very tough program, and yet you have made it through all the tortures.
Congratulations to you and your families, My Beloved — and Tortured — Poets and Quants! Now, for the next few thousand weeks — Go Change the World! -
Tinglong Dai
@TinglongDai
The Tortured Poets (and Quants) Department
Address to the @JohnsHopkins MBA Class of 2024 and Families
Tinglong Dai, Bernard T. Ferrari Professor of Business
May 20, 2024
Dear Class of 2024 Graduates and Families, Faculty, Staff, Deans, and Friends,
Four months ago, I stood on this very podium and accepted the Bernard T. Ferrari Professorship from President Ron Daniels. It was a magical evening, surrounded by so many friends and colleagues — and my family.
The next morning, a snowstorm shut down the entire Baltimore area. I was in a celebratory mood, so the snow was absolutely beautiful and only added to the magic.
Since the Ferrari professorship ceremony, people have kept asking me, "How has your life changed with your Ferrari?" My usual response is, "I LOVE it — my only complaint is that it's just too fast!"
"Can I borrow your Ferrari?" The answer is always, "Of course!" It's easy to be generous when you don't actually own a powerful, blazing, exhilarating race car.
Speaking of racing, let me tell you a story about a race between two strangers.
In 1999, I was a high school senior in China, attending school away from my parents. My tiny village was 15 miles away, and every two weeks or so, I'd hop on my bike and ride home.
On a frigid Saturday in December, just days after a blizzard had turned the world into a frozen wonderland, I set out on my usual route. It was about 10 degrees Fahrenheit, cold enough to freeze your breath in the air. The road was surprisingly clear, and I pedaled along, bundled up but still shivering.
Out of nowhere, this guy on a bike appeared, matching my pace. Then he accelerated, leaving me behind. My teenage pride couldn't take it, so I pedaled harder and passed him.
This impromptu race became an unspoken war. He accelerated, then I accelerated, neither of us willing to be the loser in this spontaneous race.
Twenty minutes later, I was riding like a maniac. The cold was biting, and my legs felt like they were on fire, but sheer stubbornness and the fear of losing face to this random stranger drove me on.
I began to sing to keep myself going — a melody that I'm sure sounded as desperate as I felt.
Just as I was savoring my small victory, he rocketed past me again. His legs were a blur of motion, like a human cyclone on wheels.
In my desperation, I pushed myself beyond my limits, pedaling and singing louder, as if sheer willpower could warm me up. But it didn't.
The next thing I knew, the world went black.
When I came to, it was eight hours later, and I was lying in a clinic just off the highway.
A young couple who ran the roadside clinic had found me unconscious on the highway outside their door. They nursed me back from the brink of severe hypothermia. Without them, my foolish, pride-fueled race might have cost me my life.
Thanks to their kindness, I'm here today to tell you this story — and two important life lessons I learned from it.
Lesson #1: Location. Location. Location. If you collapse while riding your bike in a freezing winter, make sure you collapse in front of a clinic. Distance matters for health outcomes.
Lesson #2: As much as you enjoy competitions, make sure you are in the right race. Not every race is worth doing, and some can cost you your life.
This brings me to a broader point about life and death.
Now, at least 60% of you have a significant interest in healthcare. So, I suppose it’s okay to talk about death, even at this moment of celebration, without making too many people uncomfortable.
You see, my last name is D-A-I. For years, people have playfully and endearingly called me "Dr. Die.” It’s hard to be a medical doctor with a name like that, but it’s okay to be a doctor who doesn’t treat patients.
Emily Dickinson, one of the greatest poets in human history, wrote:
Because I could not stop for Death —
He kindly stopped for me —
The Carriage held but just
Ourselves — And Immortality.
Emily Dickinson was obsessed with death throughout her life. She wasn’t alone. Charlie Munger, one of the greatest business thinkers who passed away last year at the age of 99, once said, “ALL I want to know is where I'm going to die, so I'll never go there.”
Our lives are finite, and this realization should push us to leave a lasting impact. Live like you’re dying — fearlessly and fully.
Too often, we are tortured by the fear of being too good or too sharp or too fast — who am I to shine? Who am I to play big? Who am I to change the world?
The real question is, knowing our life is so short and we only live once, who are you not to shine? Who are you not to play big? Who are you not to change the world?
In the last two years I have lost two friends, both in their 50s. Not to mention the COVID-19 pandemic and the ongoing wars around the world that have normalized death in ways that are more frightening than we realize.
As I grieved for these losses, I remembered losing my grandfather when I was about to start college. He passed away at the age of 62. I still dream about him from time to time, even today. He would have been 86 this year.
When I first came to the United States, I noticed that people, especially in business school, talked about death all the time. “My laptop just died.” “He killed it.” “Wow, she has a killer CV.” But the ultimate proof is in the ubiquitous deadlines. The deadline for this, the deadline for that. What's the drop-dead deadline? Can we extend the deadline?
At first I thought, "Wow, these business people are really morbid! Why are they so obsessed with death?" But then I realized it's actually about keeping everyone on their toes. You see, in business, time is money. But in life, money isn’t time — and money can’t buy time. And money certainly can't buy life. So deadlines are a way of keeping us all alive, even if they sound a bit morbid.
Our lives have a deadline, it’s about 3,000 weeks from now, and it's very easy to forget that. Live fully, live fearlessly, and live as if we're dying — because we are.
Live like water.
@JHUCarey is on the water. Water is how we communicate with people decades, centuries before us who are no longer with us. Water is how we communicate with people centuries after us, after we're no longer with them. Water is history — and the future. Water is the perfect setting for a school like ours that builds for what's next.
And water is diverse and unifying. It’s everywhere and it’s the most fundamental feature of our planet Earth. It flows through every culture and brings people together. Look around — our class embodies this diversity, with representatives from Bangladesh, China, Colombia, the Dominican Republic, Ethiopia, Ghana, India, Indonesia, Jamaica, Japan, Kazakhstan, Nigeria, Pakistan, Rwanda, Saudi Arabia, Sri Lanka, and Türkiye. We even have people from the United States! We’re united in water.
When we live like water, we release our fears and doubts. We find our way. We find clarity and purpose.
We sing. We dance. We hug. We fight. We celebrate. We love. We flow.
We live in the serenity and resilience of water. We change the world with the grace and power of water.
Class of 2024 Johns Hopkins MBAs, you are truly exceptional. I am so proud of you. Let’s face it: this has been a very tough program, and yet you have made it through all the tortures.
Congratulations to you and your families, My Beloved — and Tortured — Poets and Quants! Now, for the next few thousand weeks — Go Change the World! -
Tinglong Dai reposted
Yashar Ali 🐘
@yashar
NEWS
Scarlett Johansson has just issued this statement on OpenAI.
I have confirmed its authenticity directly with her publicist.
"Last September, I received an offer from Sam Altman, who wanted to hire me to voice the current ChatGPT 4.0 system. He told me that he felt that by my voicing the system, I could bridge the gap between tech companies and creatives and help consumers to feel comfortable with the seismic shift concerning humans and Al. He said he felt that my voice would be comforting to people.
After much consideration and for personal reasons, declined the offer.
Nine months later, my friends, family and the general public all noted how much the newest system named "Sky" sounded like me.
When I heard the released demo, I was shocked, angered and in disbelief that Mr. Altman would pursue a voice that sounded so eerily similar to mine that my closest friends and news outlets could not tell the difference. Mr. Altman even insinuated that the similarity was intentional, tweeting a single word "her" - a reference to the film in which | voiced a chat system, Samantha, who forms an intimate relationship with a human.
Two days before the ChatGPT 4.0 demo was released, Mr. Altman contacted my agent, asking me to reconsider. Before we could connect, the system was out there.
As a result of their actions, I was forced to hire legal counsel, who wrote two letters to Mr. Altman and OpenAl, setting out what they had done and asking them to detail the exact process by which they created the "Sky" voice. Consequently, OpenAl reluctantly agreed to take down the "Sky" voice.
In a time when we are all grappling with deepfakes and the protection of our own likeness, our own work, our own identities, I believe these are questions that deserve absolute clarity. I look forward to resolution in the form of transparency and the passage of appropriate legislation to help ensure that individual rights are protected.” -
Tinglong Dai reposted
USACE Baltimore
@USACEBaltimore
Sunday update: The refloat/transit of the DALI from its current location to a local marine terminal is tentatively scheduled for Monday morning, with the full sequence estimated to begin roughly 18 hours beforehand, starting midday today, Sunday.
-
Tinglong Dai reposted
zeynep tufekci
@zeynep
Biden's HHS has recommended disallowing EcoHealth —the collaborator of Wuhan Institute of Virology— from receiving any federal funding.
Let's note the bipartisan discomfort about being misled for years that was on full display during the hearing.
Statement from the Democrats:
Katherine Eban
@KatherineEban
Breaking: @HHSGov has recommended that @EcoHealthNYC - the U.S. science nonprofit that helped fund risky coronavirus research at Wuhan Institute of Virology - be debarred from further federal research funding. Why does this matter? A look back at some key moments.🧵👇/1
-
Tinglong Dai reposted
Andrew Nicklin
@technickle
Great thought from @TinglongDai on AI adoption in business [and I think gov't, too]: it's less like adopting another new technology, and more like adopting a child. hub.jhu.edu/johns-hopkins-briefing/
-
Tinglong Dai reposted
Neel Nanda
@NeelNanda5
I think this is absolutely outrageous behaviour from OpenAI, and far outside my understanding of industry norms. I think anyone considering joining OpenAI should think hard about whether they're comfortable with this kind of arrangement and what it implies about how employees are treated
Kelsey Piper
@KelseyTuoc
When you leave OpenAI, you get an unpleasant surprise: a departure deal where if you don't sign a lifelong nondisparagement commitment, you lose all of your vested equity: www.vox.com/future-perfect/2024/5/17/24158478/openai-depa…
-
Tinglong Dai
@TinglongDai
Tariffs don't work — except when they do.
In my latest article for @ConversationUS, I provide a historical perspective on tariffs, comparing Biden's 100% tariffs on Chinese-imported electric vehicles to Obama's solar panel tariffs: theconversation.com/history-says-tariffs-rarely-work-but-…
The Obama administration's 2012 tariffs on solar panels blocked direct imports, but failed to foster a domestic solar panel industry.
Today, the U.S. relies heavily on imports from China-linked companies operating in Southeast Asia — primarily Cambodia, Malaysia, Thailand, and Vietnam.
Will the same thing happen with EVs, especially with giants like BYD investing in Mexico?
I predict no, for three reasons:
1. Timing: By the time Obama imposed tariffs on solar panels in 2012, half of U.S. installations were already using panels imported from China.
The U.S. market currently imports only a few thousand Chinese-made EVs per year.
2. Evolving global supply chains: The COVID-19 pandemic and the war in Ukraine have reshaped global supply chains, leading to an unmistakable shift toward domestic manufacturing and a focus on supply chain resilience.
3. National security concerns: Much more than solar panels, EVs (and EV networks) pose direct national security concerns.
Given the changing geopolitical reality, Biden's 100% EV tariffs are likely the beginning of a broader, coordinated strategy rather than an isolated measure.
And Europe may be next.
Read more: theconversation.com/history-says-tariffs-rarely-work-but-…
@INFORMS @JHUCarey @JohnsHopkins @HubJHU @JHUMediaTeam -
Tinglong Dai reposted
Mario Macis
@Mario_Macis
Had the pleasure of joining @AlarmistThe podcast and chatting with the amazing @smithrebe about
THE ECONOMICS OF WILLY WONKA'S CHOCOLATE FACTORY!
shows.acast.com/f7690f61-80c5-48b0-b55d-9b99a463a956/6645…
I want to add one thing: Upon reflection, I think the parents/guardians are to blame for their kids' ... shortcomings—except Grandpa Joe, of course, he is awesome.
Anyway, I had a lot of fun, and I hope you find the episode interesting :)
🍫✨ #thealarmist #economics #willywonka
@JohnsHopkins @JHUCarey -
Tinglong Dai
@TinglongDai
The decision to use medical AI is generally made by providers, so figuring out the payment model for providers is key to scaling medical AI.
In a conversation with Hannah Daniel of @Medtech_Insight, I proposed a subscription model that gives healthcare providers unlimited usage rights for a fixed monthly fee.
The insurance company should reimburse providers for these subscription fees plus a margin based on reasonable market rates.
What's your ideal AI payment model?
Read more (behind a paywall -- full text available upon request): medtech.citeline.com/MT154654/Expert-Proposes-AI-Device-R… -
Tinglong Dai reposted
Nikhil Krishnan
@nikillinit
To understand the state of AI in healthcare, here's a back and forth between Medicare and the committee arguing how much they should get reimbursed for a given service
Services get reimbursed for how much work they're put in, the number of minutes, complexit,y etc.
Here they're arguing over the specific minutes it takes to do something and basically talking about how even using the AI...they still need to manually input data into two different interfaces which takes 3 minutes so it should be included in the reimbursement
lol -
Tinglong Dai reposted
Liz Szabo
@LizSzabo
Google's new AI feature is going to further rob newspapers of the readers they need to survive. When I searched "Are women and minorities more likely to be misdiagnosed?", Google quoted facts from my story but linked to a secondary site that ran my article. 1/2
-
Tinglong Dai reposted
Meredith Whittaker
@mer__edith
This is incredibly dangerous. It lays the path for centralized, device-level client side scanning.
From detecting 'scams' it's a short step to "detecting patterns commonly associated w/ seeking reproductive care" or "commonly associated w/ providing LGBTQ resources" or "commonly associated with tech worker whistleblowing." -
Tinglong Dai reposted
Port of Baltimore
@portofbalt
Significant milestone achieved tonight with controlled explosives successfully detonated helping to break apart the bridge and start to free the Dali vessel.
Incredible work by the entire Unified Command. @USACEBaltimore @USCG -
Tinglong Dai reposted
Port of Baltimore
@portofbalt
Another first! The first RoRo cargo ship has arrived at Dundalk Terminal since the Francis Scott Key Bridge incident & for good measure we also had a container ship today at Seagirt. Huge welcome to ACL's Atlantic Sun & @MSCCargo's Shanghai V!
Baltimore is back! @portsamericahq -
Tinglong Dai
@TinglongDai
Today, the Dali ship is expected to be finally freed from the remains of the collapsed Francis Scott Key Bridge in Baltimore, 48 days after the incident: www.thebaltimorebanner.com/politics-power/state-governmen…
Tinglong Dai
@TinglongDai
I recently spoke with @SCMPNews about the impact of the bridge collapse on global trade
Pleased to see that SCMP has dedicated a full-page feature to the bridge, the Port of Baltimore, and the city (www.scmp.com/news/china/article/3260371/baltimore-bridge-…), on its page A4:
Tinglong Dai
@TinglongDai
The port is on the road to recovery and is expected to be fully operational within two weeks: www.supplychainbrain.com/articles/39537-port-of-baltimore…
Tinglong Dai
@TinglongDai
On my flight back from #poms2024, I sat next to Prof. Mac Dada of @JHUCarey for 3 hours
We shared a moment to capture the collapsed bridge in the glow of the sunset
The view was breathtaking — if a little heartbreaking — and we both agreed that the future of the port is bright. -
Tinglong Dai reposted
Mario Macis
@Mario_Macis
“We're embarrassed about things we should be proud of and proud of things we should be embarrassed about.”
What a wonderful speech. I love you Jerry Seinfeld!
Ali A. Rizvi 🇨🇦
@aliamjadrizvi
This is the problem with the media. The focus on a small handful of students walking out on Jerry Seinfeld’s speech has completely obscured any coverage of the actual speech, which is excellent, completely apolitical, and full of gems all young people need to hear.
Here it is in full.This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Patricia Davidson
@DrPMDavidson
Today on #internationalnursesday2024 I pay tribute to my colleagues, past, present and those of the future. #nursing is a fabulous #profession, a wonderful #career and good for society
-
Tinglong Dai
@TinglongDai
Super proud of @_yuna_ny who just finished her MPH Capstone Thesis presentation at @JohnsHopkinsSPH.
Yuna's work on using GPT-4 Vision to advance Emergency Medicine is timely and insightful -
Tinglong Dai reposted
Emily Nussbaum
@emilynussbaum
Enjoying Seinfeld, a show about New Yorkers who own cars
-
Tinglong Dai
@TinglongDai
One of my favorite buildings @JohnsHopkins
-
Tinglong Dai reposted
Daniel
@growing_daniel
Nobody at Apple cares anymore
-
Tinglong Dai reposted
Trung Phan
@TrungTPhan
Apple is getting roasted for its new ad that shows a bunch of musical instruments getting crushed into an iPad.
The ad misunderstands the importance of physical tools in the creative process.
Brian Eno has a great explainer on the advantage of physical instruments over software.
Physical constraints breed creativity and mastery (and, hence, better art):
➡️ “Doesn't it strike you as interesting that we are in the second decade of the 21st century and a lot of the most interesting music is still being made by people playing very primitive instruments like electric guitars and drums.
I mean, what is a drum kit? It could be a bunch of old chairs or cans or anything. It's really quite an arbitrary bunch of junk, but why is it that people can still make interesting stuff?
Not only interesting but [also] innovative stuff using those tools, which — in digital terms — are hopelessly limited compared to all the fabulous possibilities of software synths and a program like Logic?
Well, the reason is because it's hopelessly limited. The reason is that you very quickly can understand what you can do with an electric guitar or a violin or a set of drums and you stop looking for more options and you start grappling with it.
The problem with software-based work is that you never know what it does. You can never exhaust what it does. So you can always cover the fact that you haven't got an idea by trying another option. If you have a lot of options, you don't usually have a lot of rapport with the instrument. If you have a few options, your rapport keeps increasing because you understand the options better and better. And this is why people still make good music with crude instruments and simple instruments. Because they understand them better than our software.” ⬅️ -
Tinglong Dai reposted
Angelica 🌐⚛️🇹🇼🇨🇳🇺🇸
@AngelicaOung
Everybody hated That Apple Ad but the Japanese REALLY hated it. I’ve never seen so many upset Japanese ppl commenting on a single thread:
-
Tinglong Dai reposted
Lulu Cheng Meservey
@lulumeservey
This Apple ad is what happens when you try to be “edgy” with no consideration for strategy.
Consider first principles:
1) Apple is a brand built on giving people a sense of inspiration and joy (eg, the colorful 2000s iPod ads)
2) Their target audience is consumers who are younger and lean progressive
3) Many of these consumers are uneasy about tech companies and fear their jobs/interests being eliminated
Strategically, Apple would’ve been wise to make an ad that:
1) Feels bright, hopeful, and uplifting
2) Celebrates human creativity and flourishing
3) Centers people, not a giant machine
Instead they created a vision of dystopia. The crushing of instruments is viscerally hard to watch and evokes the destruction of art during the Cultural Revolution.
One of Apple’s greatest ads was “1984,” in which color emerged from bleakness and the human spirit vanquished a dictatorial screen.
This latest ad reverses all of that and hands victory back to the screen. -
Tinglong Dai reposted
Paul Graham
@paulg
"If you’re thinking without writing, you only think you’re thinking."
— Leslie Lamport -
Tinglong Dai
@TinglongDai
RIP, Dr. Milind Dawande
"In this short life that only lasts an hour"
By Emily Dickinson: -
Tinglong Dai reposted
Tara Bannow
@TaraBannow
So, so, so unbelievably proud of my @statnews colleagues @bobjherman & @caseymross for their impactful work exposing @UnitedHealthGrp's use of #ai to deny crucial medical care for vulnerable seniors @PulitzerPrizes
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Ziad Obermeyer
@oziadias
Ever wonder why your insurer charges a copay for statins (and other high-value meds)?
Don’t they want to prevent heart attacks, strokes, etc?
Copays never made sense to me.
In this new paper in @QJEHarvard, we show copays kill. 🧵
academic.oup.com/qje/advance-article-abstract/doi/10.1093… -
Tinglong Dai reposted
Evan
@StockMKTNewz
Warren Buffett fills out arenas … here’s what the Berkshire Hathaway annual shareholder meeting looks like roughly 1 hour before Buffett starts speaking
-
Tinglong Dai reposted
The Transcript
@TheTranscript_
Buffett: "Sometimes people would say to me or Charlie...if you could only have lunch with one person from the last 2,000 or so years, who would you want to have it with?". Charlie says, 'I've already met a whole lot of them" because he read all the books. He eliminated all the trouble of going to restaurants to meet them or anything like that"
-
Tinglong Dai reposted
The Transcript
@TheTranscript_
Buffett on a better question to ask:
"What you should probably ask yourself is, "who do you feel that you'd want to spend the last day of your life with them?" And then figure out a way to start meeting them tomorrow. And meet them as often as you can. Because why wait until the last day. Don't bother with the others"
The Transcript
@TheTranscript_
Buffett: "Sometimes people would say to me or Charlie...if you could only have lunch with one person from the last 2,000 or so years, who would you want to have it with?". Charlie says, 'I've already met a whole lot of them" because he read all the books. He eliminated all the trouble of going to restaurants to meet them or anything like that"
-
Tinglong Dai reposted
The Transcript
@TheTranscript_
Buffett: Solar battery tech improvement needs time.
"There are certain things that just take a certain amount of time. My daughter hates it when I use this example, but it's really true that you can't create a baby in 1 month by getting 9 women pregnant. I mean, that's maybe what you want, but there are certain laws of nature that you have to work with"
The Transcript
@TheTranscript_
#buffett responding to a question on why they are not leaning more towards solar energy:
"Solar will never be the only source of electricity..barring some real breakthroughs in storage and that sort of thing"
$BRK.A $BRK.B -
Tinglong Dai reposted
The Transcript
@TheTranscript_
Buffett on AI:
"Scamming has always been part of the American scene. But this would make me, if I was interested in investing in scamming, [AI] is going to be the growth industry of all time....It has enormous potential for good and enormous potential for harm" -
Tinglong Dai reposted
The Transcript
@TheTranscript_
Warren Buffett accidentally calls Greg Abel the name "Charlie": ❤️
"I had actually checked myself a couple of times already. I'll slip. I'll slip again"
$BRK -
Tinglong Dai reposted
Mariana Socal
@dr_msocal
Our paper on the global production of APIs experiencing US #drugshortages found that facilities producing APIs with shortages were very large - some produced 100+ APIs. No diff in year of last FDA inspection or in receiving a warning letter. with @MatthewA_Crane @JAMA_current
-
Tinglong Dai reposted
vicki
@vboykis
I open Insta and there is Llama 3 in my search bar,, I open WhatsApp and there She is again, inviting me to “Ask Meta AI”,, I am wide awake, realizing it was a nightmare, and draw my hand to my face to rub my eyes, I cannot, there is an Ask Meta AI search bar etched into my palm
-
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
CODA day!
The CODA team at @JohnsHopkins maps pancreatic precursor lesions in 3D and at single-cell level. First 3D spatial genomic maps of large tissue samples. A human pancreas contains a staggering ~1000 lesions, some showing multifocality.
Here: www.nature.com/articles/s41586-024-07359-3 -
Tinglong Dai reposted
Paul Graham
@paulg
I met someone helping the British government with AI regulation. When I asked what they were going to regulate, he said he wasn't sure yet, and this seemed the most intelligent thing I've heard anyone say about AI regulation so far.
-
Tinglong Dai
@TinglongDai
1/4 Excited to share our paper, "Overdiagnosis and Undertesting for Infectious Diseases," has been accepted for publication in Marketing Science, a premier @INFORMS journal: tinglongdai.com/wp-content/uploads/DaiSingh-Covid-Testing…
We completed the first version in Nov 2020, inspired by @apoorva_nyc's reportingPost not available here, view on X
Tinglong Dai
@TinglongDai
3/4 @nytimes covered our paper in a Q&A in Jan 2021 during a nationwide test kit shortage: www.nytimes.com/2022/01/06/business/dealbook/business-cap…
Our paper was also selected as the lead article in the online preprint journal COVID Economics by @ceprdc
Tinglong Dai
@TinglongDai
4/4 We're grateful for the feedback and guidance from Senior Editor Prof. Anthony Dukes, Associate Editor, and the anonymous reviewers. Thanks also to Michael Fu, @DrGordonGao, @vgaur01, Erica Gralla, @KathyMcDonald, and Sridhar Seshadri.
Paper link: tinglongdai.com/wp-content/uploads/DaiSingh-Covid-Testing… -
Tinglong Dai
@TinglongDai
So proud to be a part of @JohnsHopkins:
-
Tinglong Dai reposted
Zach Baron
@zbaron
In major win for the Biden Admin, federal judge in NJ rejects multiple constitutional claims by drug companies seeking to block Medicare drug negotiation. First time a court has directly ruled on Takings Clause argument, at heart of lawsuits brought by numerous companies.
-
Tinglong Dai
@TinglongDai
In our @EMS1 op-ed (www.ems1.com/artificial-intelligence/taking-geographical-…), which came out today, @_yuna_ny and I discuss how multimodal AI models like GPT-4 Vision could help expand access to emergency care around the world
Yuna is a physician - and an MBA/MPH student at @JHUCarey & @JohnsHopkinsSPH. I had the honor of working with Yuna on two core courses (Operations Management and Artificial Intelligence) and her MPH Capstone project
This piece is based on our ongoing research to evaluate the potential of GPT4 Vision in emergency medicine - stay tuned!
Read the article: www.ems1.com/artificial-intelligence/taking-geographical-…
@INFORMS @JHU_HBHI @JohnsHopkins @JHUNursing @IDIESJHU -
Tinglong Dai reposted
Anna Nagurney
@Supernetworks
Thanks to my @isenbergumass PhD student @Hassani_Dana for presenting today at the @POM_Society conference #poms2024 in Minneapolis on a paper of ours with Ukrainian @kse_ua colleagues @oniviev1 & Pavlo Martyshev! Our presentations on #foodsecurity & #internationaltrade #network performance under #disruptions are now posted supernet.isenberg.umass.edu/visuals.html
-
Tinglong Dai reposted
Johns Hopkins University
@JohnsHopkins
Hosted by @HopkinsMedicine, tomorrow's Health Policy Forum welcomes Renee Wegrzyn and Kimberley Steele. They will discuss the role of @ARPA_H in advancing health equity and accelerating health outcomes. washingtondc.jhu.edu/upcoming-events/series-forums/health…
-
Tinglong Dai reposted
-
Tinglong Dai reposted
Port of Baltimore
@portofbalt
Another milestone today! First container ship to arrive at Seagirt Terminal since the crisis began. @MSCCargo Passion III came through 35-foot-deep temporary channel. Nearly 1,000 containers handled by about 80 @ILALocal333 workers. We're getting there... @portsamericahq
-
Tinglong Dai reposted
Anna Nagurney
@Supernetworks
I love this photo taken with @TinglongDai @johnshopkins and his colleagues @mayoclinic plus with Suresh Sethi @ut_dallas plus a new friend from the @UofT! The BEST part of scientific conferences is the PEOPLE!
-
Tinglong Dai reposted
Tinglong Dai
@TinglongDai
In this case you should go for the one with the comma.
-
Post not available here, view on X
Tinglong Dai
@TinglongDai
In this case you should go for the one with the comma.
-
Tinglong Dai reposted
-
Tinglong Dai
@TinglongDai
#poms2024 #edebate So far, no one seems to be suggesting that GenAI is a passing fad in the operations management community. More on the constructive side.
It seems that GenAI is a safe research topic and tool -- we just need to be more purposeful, careful, and deliberate in our efforts -
Tinglong Dai reposted
-
Tinglong Dai reposted
POM Society
@POM_Society
Can OM problems, methods and solutions be communicated via stimulating and thought-provoking art produced using Gen AI? - Rema #poms2024 #edebate @TinglongDai
-
Tinglong Dai
@TinglongDai
insights from Professor Rema Padman @HeinzCollege @CarnegieMellon #edebate #poms2024
POM Society
@POM_Society
The humanizing potential of Gen AI (and its continuing evolutions) is likely another area with significant ability to impact the field. Recent reports have shown how Gen AI can produce more empathetic replies to patient queries compared to clinician responses. - Rema #edebate
-
Tinglong Dai reposted
POM Society
@POM_Society
The humanizing potential of Gen AI (and its continuing evolutions) is likely another area with significant ability to impact the field. Recent reports have shown how Gen AI can produce more empathetic replies to patient queries compared to clinician responses. - Rema #edebate
-
Tinglong Dai reposted
Jay Swaminathan, Ph.D.
@jmswaminathan
#poms2023 #edebate - #generativeai could be quite useful in teaching I believe. Particularly for those new topics where one might have limited knowledge, this could be tool that accelerate the learning process. However, one should always cross check the validity. Thoughts?
-
Tinglong Dai reposted
Jay Swaminathan, Ph.D.
@jmswaminathan
#poms2023 #edebate - a fundamental issue with #generativeai tools like #chatgpt at this point of time is a lack of complete clarity about what it is trained and how it generates results. More refinement is needed before it can be adopted in #research
-
Tinglong Dai reposted
-
Tinglong Dai reposted
Anjana Susarla
@asusarla
I am optimistic that collaboration between humans and AI will disrupt how academics assess scholarly manuscripts and disseminate published works as well as closing of gaps among diverse scholars as well as competing scholarly traditions #poms2024 #edebate.
Tinglong Dai
@TinglongDai
Professor @TayurSridhar from @CarnegieMellon on using GenAI in peer review:
#poms2024 #edebate
Sridhar Tayur
@TayurSridhar
#edebate It is an inevitable thing. I used it recently to write a report for a University Professor case at CMU, but disclosed it by say that "Initial draft prepared by ChatGPT'. I color coded my edits, and showed the parts that were struck out. No one on the committee complained
-
Tinglong Dai
@TinglongDai
The first sentence from the piece: "In the rapidly evolving landscape of academic research, artificial intelligence (AI) is poised to revolutionize traditional academic peer review processes and knowledge evaluation systems." #poms2024 #edebate
Anjana Susarla
@asusarla
Yes, this was another review piece my co-authors and I did #poms2024 #edebate
aisel.aisnet.org/jais/vol25/iss1/2/ -
Tinglong Dai
@TinglongDai
Professor @TayurSridhar from @CarnegieMellon on using GenAI in peer review:
#poms2024 #edebate
Sridhar Tayur
@TayurSridhar
#edebate It is an inevitable thing. I used it recently to write a report for a University Professor case at CMU, but disclosed it by say that "Initial draft prepared by ChatGPT'. I color coded my edits, and showed the parts that were struck out. No one on the committee complained
-
Tinglong Dai reposted
Sridhar Tayur
@TayurSridhar
#edebate It is an inevitable thing. I used it recently to write a report for a University Professor case at CMU, but disclosed it by say that "Initial draft prepared by ChatGPT'. I color coded my edits, and showed the parts that were struck out. No one on the committee complained
-
Tinglong Dai reposted
POM Society
@POM_Society
Could a Gen AI agent see patterns between your question and hypothesis to methods from another discipline? Or tell you about the application opportunities of your methods and recommend people who would be good research partners? - Rema #poms2024 #edebate
-
Tinglong Dai
@TinglongDai
Important insight from Professor Rema Padman @HeinzCollege @CarnegieMellon
POM Society
@POM_Society
Some scholars have arbitraged their knowledge of a discipline to transform another (Prof. Alfred Blumstein, for example, used Operations Research to transform the field of Criminal Justice research and practice). - Rema #edebate #poms2024
-
Tinglong Dai
@TinglongDai
Excellent idea - a bit like the idea of using AI to generate draft responses to patient messages to doctors #poms2024 #edebate
Vince Slaugh
@VinceSlaugh
Has anyone discussed the idea of having a standard AI reviewer report with which the other referees could engage? The value of the human referees could be more in weighing the pros and cons rather than in generating that initial list of pros and cons. #poms2024 #edebate
-
Tinglong Dai
@TinglongDai
Insight from @adi_vedant echos what I wrote a while ago: thehill.com/opinion/technology/4037499-generative-ai-is-n… #poms2024 #edebate
Aditya
@adi_vedant
Critical consideration: while ChatGPT is a powerful tool, its reliance on existing data can perpetuate biases and inaccuracies in academic research. Human oversight is essential to ensure nuanced analysis and mitigate potential pitfalls. #poms2024 #edebate
-
Tinglong Dai
@TinglongDai
Professor @asusarla highlights three key downsides/risks of using GenAI for research: content authentication, homogenization of ideas, and hallucinations #poms2024 #edebate
Anjana Susarla
@asusarla
#generativeai offers immense potential in
1.Research design
2. Data collection and analysis
3. Methodological critiques
However, these must be counterbalanced against the challenges posed by content authentication, homogenization of ideas and hallucinations #poms2024 #edebate
Tinglong Dai
@TinglongDai
Professor @TayurSridhar on the value of GenAI in doing Operations Management research: #edebate @NaveenKPhD, @VinceSlaugh, @asusarla, @jmswaminathan, @emremd, @tnetland
Sridhar Tayur
@TayurSridhar
#edebate
I expect two types of research. Positioning GenAI Data Generation as upgrade of 20th Century's Pseudo-random number generator:
1. Figuring out unintended patterns [10%]
2. Users of this synthetic data for simulation/evalation [90%] -
Tinglong Dai
@TinglongDai
Some interesting perspective from @emremd -- ChatGPT-generated reviews are "kinder" but may lack relevance #poms2024 #edebate
Emre M Demirezen
@emremd
Although there are reports that ChatGPT review reports are kinder, it is sad that some reviewers solely rely on ChatGPT for reviews. Gen AI is/can be inherently biased on the data it was trained on, and it likely has a lack of contextual understanding to provide a timely report.
-
Tinglong Dai
@TinglongDai
Professor @asusarla shared guidelines about using ChatGPT and other GenAI tools in peer review: #poms2024 #edebate
Anjana Susarla
@asusarla
my co-authors and I set some guidelines for peer review in this #isr article #poms2024 #edebate
pubsonline.informs.org/doi/10.1287/isre.2023.ed.v34.n2 -
Tinglong Dai
@TinglongDai
Professor @TayurSridhar on the value of GenAI in doing Operations Management research: #edebate @NaveenKPhD, @VinceSlaugh, @asusarla, @jmswaminathan, @emremd, @tnetland
Sridhar Tayur
@TayurSridhar
#edebate
I expect two types of research. Positioning GenAI Data Generation as upgrade of 20th Century's Pseudo-random number generator:
1. Figuring out unintended patterns [10%]
2. Users of this synthetic data for simulation/evalation [90%] -
Tinglong Dai reposted
Sridhar Tayur
@TayurSridhar
#edebate
Next level is creation and use Automated Model Building, especially if it well structured enough, in silico. -
Tinglong Dai
@TinglongDai
I have a question for #poms2024 #edebate panelists (
@TayurSridhar, @NaveenKPhD, @VinceSlaugh, @asusarla, @jmswaminathan, @emremd, @tnetland): What do you think about peer reviews written using ChatGPT? Is it a good or bad thing? Should we require reviewers to disclose their use of GenAI in peer reviews? -
Tinglong Dai
@TinglongDai
The mood about GenAI has been overwhelmingly positive so far ... any downsides? #poms2024 #edebate
-
Tinglong Dai reposted
Jay Swaminathan, Ph.D.
@jmswaminathan
#poms2024 #edebate - @TinglongDai I believe that #generativeai will have a positive and important impact on firm #operations and academic #research
-
Tinglong Dai reposted
Subodha Kumar
@subodha0
Welcome #poms2024 #edebate panelists: @TayurSridhar @NaveenKPhD @VinceSlaugh @asusarla @jmswaminathan @emremd @tnetland Rema Padman
POM Society
@POM_Society
Panelists of #poms2024 #edebate: Prof. Sridhar Tayur @TayurSridhar; Prof. Naveen Kumar @NaveenKPhD; Prof. Rema Padman; Prof. Vince Slaugh @VinceSlaugh, Prof. Anjana Susrala @asusarla; Prof. Tor Netland @tnetland; Prof. Emre Demirezen @emremd; Prof. Jay Swaminathan @jmswaminathan
-
Tinglong Dai reposted
Vince Slaugh
@VinceSlaugh
For some of the most important labor-intense care operations, AI still seems frustratingly far off. Here's some great reporting on this in nursing homes. #poms2024 #edebate
-
Tinglong Dai reposted
POM Society
@POM_Society
Panelists of #poms2024 #edebate: Prof. Sridhar Tayur @TayurSridhar; Prof. Naveen Kumar @NaveenKPhD; Prof. Rema Padman; Prof. Vince Slaugh @VinceSlaugh, Prof. Anjana Susrala @asusarla; Prof. Tor Netland @tnetland; Prof. Emre Demirezen @emremd; Prof. Jay Swaminathan @jmswaminathan
-
Tinglong Dai
@TinglongDai
Very honored to moderate tonight's @POM_Society eDebate. Our topic tonight is: “Will Generative AI become an enduring research tool and topic in Operations Management, or will it be a passing fad that will be forgotten in a few years? ”
Subodha Kumar
@subodha0
We would like to welcome #poms2024 #edebate moderator: @TinglongDai
Tinglong Dai
@TinglongDai
Throughout this conference, GenAI has been the most discussed topic across panels, sessions, and chats. This eDebate will bring out some of the brilliant insights from our @POM_Society #edebate
-
Tinglong Dai reposted
Soumith Chintala
@soumithchintala
apparently Google laid off their entire Python Foundations team, WTF!
( @SkyLi0n who is one of the pybind11 maintainers just informed me, asking what ways they can re-fund pybind11)
The team seems to have done substantial work that seems critical for Google internally as well.
There's a hackernews thread if folks want to read more: news.ycombinator.com/item?id=40171125 -
Tinglong Dai reposted
Matthew E. Kahn
@mattkahn1966
How does AI handle the Lucas Critique? How does the algorithm know how the optimizing agent will respond as the "rules of the game" change? To paraphrase Led Zeppelin, "does the song remain the same"?
Can AI recover time invariant structural attributes of decision makers?
John Cochrane
@JohnHCochrane
Supply, demand, AI, and humans at Grumpy Economist. Will I be displaced by AI? www.grumpy-economist.com/p/supply-demand-ai-and-humans
-
Tinglong Dai
@TinglongDai
Minneapolis
-
Tinglong Dai reposted
Port of Baltimore
@portofbalt
Big day in Baltimore!
The Saimaagracht was the first vessel to successfully move through the temporary 35-foot deep channel from the Port of Baltimore’s public terminals. Great work @USCG sector Baltimore and @USACEBaltimore. -
Tinglong Dai reposted
American Academy of Pain Medicine (AAPM)
@AmerAcadPainMed
Join our latest Pain Matters Podcast as hosts @Shravanid_md and @mbroach4 discuss Artificial Intelligence in healthcare with @Johns Hopkins experts @TinglongDai and @NesMathioudakis.
-
Tinglong Dai
@TinglongDai
Patricia Davidson
@DrPMDavidson
Spring is the most wonderful of seasons. #hope, #renewal and #resilience
-
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
How can the U.S. healthcare system, one of the nation’s largest contributors of waste, mitigate its footprint?
@SecBecerra of @HHSGov joined #hopkinsbloombergcenter for an #earthday @JHU_HBHI @JohnsHopkinsSPH event discussing watts to reduce waste. More: washingtondc.jhu.edu/news/how-the-health-care-system-can-… -
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
We use AI-based FILM to restore damaged images
Here are examples for histology (H&E), EM, MRI, and light-sheet.
Red frames show defects. Green frames show how we eliminate them.
More about our method here: www.biorxiv.org/content/10.1101/2024.03.07.583909v3 -
Tinglong Dai reposted
UCLA Anderson School of Management
@uclaanderson
💙💛💙💛
Let us be grateful to people who make us happy, they are the charming gardeners who make our souls blossom. - Marcel Proust
Tinglong Dai
@TinglongDai
One can have many titles in academia, but being a "Friend of UCLA Anderson" is a very special honor for which I am grateful. A huge thank you to my dear friend, Prof. @christangucla, for inviting me to speak on the distinguished @uclaanderson panel on AI and Healthcare on May 8.
I'll be speaking alongside Aditi Gupta and Yin (Amelia) Hui, two distinguished @uclaanderson alumni and health tech leaders at @Amgen, a biotech giant.
Register for this Wednesday (May 8, 2024) event that will take place 5-6 pm PT (8-9 pm ET): ucla.zoom.us/webinar/register/WN_yvP7fPuCQHui9jePa473SA -
Tinglong Dai reposted
UCLA Anderson School of Management
@uclaanderson
💙💛💙💛
As Ralph Waldo Emerson said, "The only way to have a friend is to be one." Thank you for your support and collaboration - always great to have you join us to share your time and expertise. -
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
Final FY24 budget for US research is a disaster.
Budgets of federal agencies are mostly or cut down
NIH $47.08B flat
NSF $9.06B (-5%)
NASA Sci. $7.33B (-6%)
CDC $9.2B flat
DOE Sci. $8.24B (+1.7%)
ARPA-E 0.46B (-2%)
DOD Sci. $2.6B (-9%)
DARPA $4.1B flat
Global Health $10B (-5%) -
Tinglong Dai reposted
Hopkins Business of Health Initiative
@JHU_HBHI
And @JHU_HBHI
is grateful to @JoanneKenen for organizing and moderating this amazing event.
Now viewable on C-Span here: c-span.org/video/?535075-1/health-human-services-secretar…
Joanne Kenen
@JoanneKenen
was honored to moderate this Earth Day event on climate and health -- and the health system's own considerable carbon footprint - with @SecBecerra - an event that brought together @JohnsHopkinsSPH @IJHNursing @JHUCarey and @JHU_HBHI
Secretary Xavier Becerra
@SecBecerra
I celebrated Earth Day this morning with a fireside chat at @JohnsHopkinsSPH about the Administration’s efforts to help hospitals & health systems move to more sustainable practices. @POTUS’ Inflation Reduction Act is helping us invest in more energy efficient health facilities.
-
Tinglong Dai
@TinglongDai
One can have many titles in academia, but being a "Friend of UCLA Anderson" is a very special honor for which I am grateful. A huge thank you to my dear friend, Prof. @christangucla, for inviting me to speak on the distinguished @uclaanderson panel on AI and Healthcare on May 8.
I'll be speaking alongside Aditi Gupta and Yin (Amelia) Hui, two distinguished @uclaanderson alumni and health tech leaders at @Amgen, a biotech giant.
Register for this Wednesday (May 8, 2024) event that will take place 5-6 pm PT (8-9 pm ET): ucla.zoom.us/webinar/register/WN_yvP7fPuCQHui9jePa473SA -
Tinglong Dai reposted
Daniel Polsky
@healthecon_dan
Thank you @SecBecerra for your visit with us at @JHU_HBHI . There is much work to be done to address the health sector's climate footprint.
Secretary Xavier Becerra
@SecBecerra
I celebrated Earth Day this morning with a fireside chat at @JohnsHopkinsSPH about the Administration’s efforts to help hospitals & health systems move to more sustainable practices. @POTUS’ Inflation Reduction Act is helping us invest in more energy efficient health facilities.
-
Tinglong Dai reposted
San Francisco Chronicle
@sfchronicle
Nurses and supporters protest outside of Kaiser Permanente hospital in S.F. as they demonstrate against the use of AI in treatment of patients. 🎥: @gabriellelurie
-
Tinglong Dai reposted
Curt Langlotz
@curtlanglotz
Wow: "Nurses are all for tech that enhances our skills and the patient care experience. But what we are witnessing in our hospitals is the degradation and devaluation of our nursing practice through the use of these untested technologies." www.beckershospitalreview.com/healthcare-information-tech… via @beckershr
-
Tinglong Dai reposted
Joanne Kenen
@JoanneKenen
was honored to moderate this Earth Day event on climate and health -- and the health system's own considerable carbon footprint - with @SecBecerra - an event that brought together @JohnsHopkinsSPH @IJHNursing @JHUCarey and @JHU_HBHI
Secretary Xavier Becerra
@SecBecerra
I celebrated Earth Day this morning with a fireside chat at @JohnsHopkinsSPH about the Administration’s efforts to help hospitals & health systems move to more sustainable practices. @POTUS’ Inflation Reduction Act is helping us invest in more energy efficient health facilities.
-
Tinglong Dai
@TinglongDai
A room with a view - Classroom 428 @JHUBloombergCtr
-
Tinglong Dai
@TinglongDai
Years from now, we'll look back on 2024 as a pivotal moment when a quiet #generativeai revolution took place in healthcare.
If you're attending #poms2024 in Minneapolis this week, don't miss the special panel on @MayoClinic's pioneering experience integrating #genai into its routine healthcare practices.
Two of @MayoClinic's top AI leaders who have done groundbreaking work in medical AI, Dr. @xiaoxiyao and Dr. @EricWillMD, will share their insights and opportunities for collaboration with the OM community. I'm thrilled to moderate the panel with my esteemed colleague and dear friend, @ProfRachnaShah from the @UMNews.
Click this link to add the event to your calendar: www.addevent.com/event/uJ21303617
Hope to see some of you on Saturday 2:15pm CT at Boundry Waters B! -
Tinglong Dai
@TinglongDai
Wordle 1,038 2/6
⬛🟩⬛🟩🟩
🟩🟩🟩🟩🟩 -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai
@TinglongDai
The five Baltimore women, known as the “Friday Evening Group,” who founded @BrynMawrSchool and @BrynMawrCollege:
- M. Carey Thomas
- Mary Elizabeth Garrett
- Mamie Gwinn
- Bessie King
- Julia Rogers -
Tinglong Dai reposted
Ge Bai
@GeBaiDC
My latest @Forbes: Why Are Cash Prices Lower Than Health Insurance Negotiated Prices?
@JHUCarey @JohnsHopkinsSPH @BSPH_HPM
www.forbes.com/sites/gebai/2024/04/21/why-are-cash-prices… -
Tinglong Dai reposted
-
Tinglong Dai reposted
NEJM AI
@NEJM_AI
Sustainable reimbursement is key for medical AI to scale and benefit patients. A new Policy Corner article by @MichaelAbramoff, @TinglongDai, and @james_y_zou explores fee-for-service and value-based care payment models for the use of medical AI. nejm.ai/445c5JT
-
Tinglong Dai
@TinglongDai
Yes
-
Tinglong Dai reposted
-
Tinglong Dai reposted
NEJM AI
@NEJM_AI
Policy Corner: Scaling Adoption of Medical AI — Reimbursement from Value-Based Care and Fee-for-Service Perspectives nejm.ai/445c5JT
-
Tinglong Dai reposted
Matthew Yglesias
@mattyglesias
As he says, academia’s very heavy emphasis on novelty incentivizes people to do the kind of work that’s not so useful to policymakers who mostly need to get the basics right — better data, a clearer sense of the overall state of the literature.
www.slowboring.com/p/the-economic-research-policymakers -
Tinglong Dai
@TinglongDai
Excited to share our new @NEJM_AI article, which explores payment models for the use of medical AI.
With @MichaelAbramoff (@AItheRightWay) & @james_y_zou (@Stanford), I look at two payment models: fee-for-service and value-based care, and their implications for the sustainability of medical AI businesses
Full text: ai.nejm.org/stoken/default+domain/DUUAKAG7U6RSAWVT3TDD/fu… -
Tinglong Dai reposted
James Zou
@james_y_zou
How to get paid is increasingly the biggest puzzle facing biomedical #ai -- more so than developing the technology. Our new @NEJM_AI article discusses the economics of medical AI + payment landscape ai.nejm.org/stoken/default+domain/DUUAKAG7U6RSAWVT3TDD/fu…
Real pleasure writing this w/ @MichaelAbramoff and @TinglongDai! -
Tinglong Dai reposted
Alex Imas
@alexolegimas
One of the first thing you realize as an immigrant from almost anywhere is that American strawberries look amazing but taste like wet cardboard.
Crémieux
@cremieuxrecueil
Strawberry production in both Europe (gray) and America (blue) has skyrocketed.
Europe produced more strawberries by planting way more acres. America produced more strawberries without planting way more acres.
Instead, America made a better strawberry. -
Tinglong Dai reposted
James Lucas
@JamesLucasIT
Photos that don't look real but are - thread 🧵
1. Cranberry harvest by Kristina Makeeva -
Tinglong Dai
@TinglongDai
There's high-speed rail. And then there's American high-speed rail. But @GoBrightline has been pretty good
-
Tinglong Dai reposted
Kathryn Hulick
@khulick
I couldn't stop thinking about the #totality - the purple-gray sky, the magenta horizon, the surreal ring of glowing light. So I decided to paint it. #solareclipse #eclipse
kathrynhulick.substack.com/p/i-painted-the-totality -
Tinglong Dai
@TinglongDai
Miami
-
Tinglong Dai reposted
Michele Avissar-Whiting
@maw_tweets
Nature now charges $12,290 US to publish a single article open access. Shall we list some things that cost less than that? I'll go first: a 2017 BMW 330i sedan, a 7-day cruise for a family of four, six Louis Vuitton purses. 🤪
-
Tinglong Dai reposted
Hopkins Nursing
@JHUNursing
According to @usnews, the Johns Hopkins School of Nursing is the No. 1 nursing master's program (tied) and No.1 Doctor of Nursing Practice (DNP) program! #gohopnurse
nursing.jhu.edu/newsroom/news/the-johns-hopkins-school-of… -
Tinglong Dai
@TinglongDai
All the AI geniuses stop shining when IT, like a projector or a zoom link, doesn't work :-)
Tinglong Dai
@TinglongDai
I asked @ChatGPTapp 4.0 to illustrate this statement, and this is what I got:
-
Tinglong Dai
@TinglongDai
Legendary health economist Joe Newhouse @Harvard just started his Sam Shapiro Lecture @JohnsHopkinsSPH @JohnsHopkins
-
Tinglong Dai reposted
Andrew Ryan
@Andy_Ryan_dydx
Physician payment reform is in the air.
So I wanted to share our paper about the Medicare Shared Savings Program.
We find that the MSSP has had virtually no budget impact for CMS.
ACOs are not the answer to cost containment in Traditional Medicare.
jamanetwork.com/journals/jama-health-forum/fullarticle/28… -
Tinglong Dai reposted
Paul Graham
@paulg
Someone sent me a cold email proposing a novel project. Then I noticed it used the word "delve."
-
Tinglong Dai reposted
Mario Macis
@Mario_Macis
Totality in Rochester, NY, although this picture doesn’t make it justice. It got dark, sky and all, really fast for 3-4 minutes. Awesome experience.
-
Tinglong Dai reposted
Anna Nagurney
@Supernetworks
Absolutely perfect, written by the one and only Emily Dickinson, and shared by the one and only @tinglongdai! We have to bring him to @townofamherst!
Tinglong Dai
@TinglongDai
🌑🌕🔥
-
Tinglong Dai
@TinglongDai
🌑🌕🔥
-
Tinglong Dai reposted
Eric Topol
@EricTopol
Note the asterisk at the bottom of the Figure—patients without diabetes
Patients with diabetes weight loss was 40% less at the same dose of semaglutide in a new trial. We don't know why.
The longstanding mystery also accounts, in part, for the marked delay of testing GLP-1 drugs for obesity (they were targeted for diabetes and minimal weight loss was seen) -
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Bloomberg Economics
@economics
After a Baltimore bridge came crashing down, dozens of ships scrambled to change course and take their cargo elsewhere. Here’s where they went www.bloomberg.com/graphics/2024-baltimore-bridge-collapse…
-
Tinglong Dai reposted
unusual_whales
@unusual_whales
Larry Summers, an OpenAI board member, thinks AI could replace ‘almost all’ forms of labor, per FORTUNE.
-
Tinglong Dai reposted
Taylor Swift
@taylorswift13
Only a fortnight til The Tortured Poets Department…🤍
-
Tinglong Dai
@TinglongDai
Lehigh
-
Tinglong Dai
@TinglongDai
Visiting @LehighU. Just experienced an earthquake
-
Tinglong Dai reposted
Reuters
@Reuters
The US Army Corps of Engineers said it expects to open a new channel to the Port of Baltimore by the end of April, freeing up commercial shipping blocked by a collapsed bridge, and then restore port access to full capacity by the end of May www.reuters.com/world/us/baltimore-shipping-set-resume-by…
-
Tinglong Dai reposted
Governor Wes Moore
@GovWesMoore
Join us for live updates on our ongoing response to the Key Bridge collapse: x.com/i/broadcasts/1lPJqbwwwylxb
-
Tinglong Dai reposted
Anna Nagurney
@Supernetworks
Why was Taiwan so well prepared for the earthquake? www.reuters.com/world/asia-pacific/resilient-taiwan-respo… I reached out to a colleague in Taiwan, who even spoke in our @umassinforms speaker series, to see if she was safe and she responded today that she and her family are, which is a big relief. #disasters
-
Tinglong Dai reposted
Hopkins Business of Health Initiative
@JHU_HBHI
Join us at @BloombergCenter on Earth Day to hear from @XavierBecerra on steps to address the healthcare sector’s carbon footprint. Followed by an expert panel discussion.
-
Tinglong Dai reposted
✨Mai-Ling Garcia
@Ling_Garcia
Great read about the importance of the Francis Scott Key Bridge. " #baltimore is like Taylor Swift in the universe of U.S. auto ports, and when she's not performing as usual, even for a few weeks, the world feels a great loss." @JohnsHopkins @TinglongDai hub.jhu.edu/2024/03/26/baltimore-key-bridge-collapse-supp…
-
Tinglong Dai reposted
Tracy Alloway
@tracyalloway
Map of the Taiwan earthquake's impact on semiconductor and electronics production, via Barclays: "We believe this could lead to supply disruptions in the tech supply chain"
-
Tinglong Dai reposted
Reuters
@Reuters
Taiwan's biggest earthquake in at least 25 years killed seven people on Wednesday, injuring more than 700, with 77 trapped in tunnels and collapsed buildings, authorities said, as rescuers used ladders to help some people descend to safety www.reuters.com/world/asia-pacific/strong-72-magnitude-ea…
-
Tinglong Dai
@TinglongDai
Jon Stewart **delves into** realities of AI 🤣
This media couldn't be displayed. See it on the original post. -
Tinglong Dai
@TinglongDai
TSMC disrupted by Taiwan's largest earthquake in 25 years
Wen-Yee Lee
@Wenyee_Lee
#tsmc updates: “#taiwan experienced several significant #earthquakes this morning. To ensure the safety of personnel, evacuation procedures have been initiated in some foundry areas according to internal company protocols. The occupational safety system is functioning normally. Detailed information is still being confirmed at the moment.”
-
Tinglong Dai
@TinglongDai
A second temporary alternate channel opened on Tuesday near Hawkins Point, allowing some critical commercial ships to travel around the ruins of the Francis Scott Key Bridge #baltimorebridgecollapse
The Baltimore Banner
@BaltimoreBanner
2nd temporary alternate channel opens near wreckage of Key Bridge www.thebanner.com/community/transportation/francis-scott-…
-
Tinglong Dai reposted
Amanda Kell
@amandamkell
Key Bridge's NTSB investigator also probing implosion of submersible on Titanic trip - by @alex_mann10 @baltimoresun www.baltimoresun.com/2024/04/02/lead-key-bridge-investiga…
-
Tinglong Dai
@TinglongDai
Writing for the March 2024 issue of @cyberdefensemag, I highlight the hidden costs of the now ubiquitous facial recognition technology.
Privacy concerns aside, the technology itself is unreliable, which poses a great danger when it's relied upon by humans.
We need to be open about how models are trained and where the data comes from. These models should also be regularly tested and monitored by outside groups.
Now is the time to act.
Without purposeful governance, every second that passes is a step toward a future where our lives are just data points in a vast surveillance network that is as unchecked as it is flawed.
We cannot put our future on autopilot.
We can and should steer technology toward a future that values and protects people's right to privacy and freedom of choice. -
Tinglong Dai
@TinglongDai
Advice @HillaryClinton gave Maryland Gov. @iamwesmoore in October 2022:
www.thebaltimorebanner.com/politics-power/state-governmen… -
Tinglong Dai
@TinglongDai
First vessel passes through the temporary channel #baltimorebridgecollapse: www.thebaltimorebanner.com/community/transportation/balti…
-
Tinglong Dai
@TinglongDai
GenAI will be a boon to the energy sector in a way many people don't expect—it creates a lot of demand for energy
Jamie Smyth
@JamieSmythF
AI revolution will be boon for natural gas, say fossil fuel bosses, as rapid data center roll out prompts surge in power demand www.ft.com/content/1f93b9b2-b264-44e2-87cc-83c04d8f1e2b via @ft @mylesmccormick_ @amandalanchu
-
Tinglong Dai reposted
Arjun (Raj) Manrai
@arjunmanrai
We know LLMs can ace multiple-choice exams. Taking us deeper, an important new study led by @stephcabral_ and @AdamRodmanMD conducts a nuanced evaluation of the clinical reasoning abilities of GPT-4 wrt physicians. Guess who wins?
Need more of these!
jamanetwork.com/journals/jamainternalmedicine/article-abs… -
Tinglong Dai reposted
Jill Rosen
@Jillrosen
Reporters from Baltimore and around the world worked tirelessly to cover the horrific Key bridge collapse and many relied on @JohnsHopkins bridge, supply chain and deep-water search experts to help explain it. Thank you @BenjaminSchafer Rachel Sangree @TinglongDai @JGBellingham
-
Tinglong Dai reposted
Johns Hopkins University Bloomberg Center
@JHUBloombergCtr
The tragic collapse of Francis Scott Key Bridge sent shockwaves across the nation. Professor of operations management at the @JHUCarey, @TinglongDai explains the potential short-term and long-term impact on the supply chain, economy, and more.
hub.jhu.edu/2024/03/26/baltimore-key-bridge-collapse-supp… -
Tinglong Dai
@TinglongDai
.@nytimes, in its flagship newsletter, "The Morning," quoted me today in the Sunday Debate section
The debate was about whether the Key Bridge collapse will disrupt the Port of Baltimore
I was quoted in the "no disruption" camp:
Tinglong Dai
@TinglongDai
I have since received emails from readers about my position
To be clear, I don't think this collapse is all about the power of images and that the port will recover overnight, because it won't
See my @BaltimoreBanner op-ed: www.thebaltimorebanner.com/opinion/community-voices/key-b…
Tinglong Dai
@TinglongDai
But to go so far as to say that @portofbalt is "dead" or can be easily absorbed by the other East Coast ports would be wrong, for geopolitical reasons, for the operational excellence the port has established in key sectors, and for many other factors
I am optimistic about the future of @portofbalt and remain unconvinced by all the pessimistic views, such as the @washingtonpost article cited in the @nytimes newsletter:
www.washingtonpost.com/opinions/2024/03/26/key-bridge-col… -
Tinglong Dai reposted
The Weeknight on MS NOW
@weeknightmsnow
“We don’t need your favors. We need you to step up and take leadership in this moment.” Maryland Governor @iamwesmoore’s message to Republican leaders in wake of the Baltimore bridge collapse.
-
Tinglong Dai
@TinglongDai
Philadelphia
-
Tinglong Dai reposted
Emily Hofstaedter
@EHofstaedter
“I like to use a phrase that says the Port of Baltimore is the Taylor Swift of U.S. auto parts,” said Tinglong Dai, at Johns Hopkins University. “When she's not performing, the world, it feels a greater loss.”
www.wypr.org/wypr-news/2024-03-27/baltimores-key-bridge-r… -
Tinglong Dai reposted
Johns Hopkins University
@JohnsHopkins
Tinglong Dai, an expert from @JHUCarey, discusses the short-term and long-term impact of the bridge collapse on consumers, suppliers, and the economy. hub.jhu.edu/2024/03/26/baltimore-key-bridge-collapse-supp…
-
Tinglong Dai reposted
James Rogers
@jamesjrogers
Supply chain improvements of recent years will limit impact of Baltimore bridge collapse, says @JohnsHopkins @JHUCarey professor @TinglongDai #baltimore #baltimorebridgecollapse www.marketwatch.com/livecoverage/baltimore-key-bridge-col… via @marketwatch
-
Tinglong Dai reposted
Shravani Durbhakula, MD, MPH, MBA
@ShravaniD_MD
I’ve been overwhelmed with all kinds of letters, emails, DMs, and comments about my recent Guest Essay in the The @nytimes @nytopinion. It’s amazing to hear from each and every one of you. Although I can’t address each of you back, please know that I appreciate your love!
-
Tinglong Dai reposted
Maxime Cohen
@maxccohen
We are excited to announce a new Special Issue of Service Science, an @INFORMS journal, on the Impact of AI on Service Design and Delivery, co-edited by @TinglongDai, @SunnyfaceBeibei, and myself.
-
Tinglong Dai
@TinglongDai
.@WashingtonPost ran a nice article this am on "Cars, Coal and Tofu" and I was quoted on the geopolitical context (thanks @JeanneWhalen): www.washingtonpost.com/business/2024/03/27/baltimore-port…
As U.S. moves to increase trade with friendly nations & reduce geopolitical risk, @portofbalt is uniquely positioned for the supply chain transition
Tinglong Dai
@TinglongDai
The related part:
-
Tinglong Dai
@TinglongDai
Quoted in this informative @WSJ article on the economic impact of the Francis Scott Key Bridge collapse: www.wsj.com/finance/baltimore-bridge-economic-impact-0514…
Thanks to @HarrietTorry for the interview
@JHUCarey @JohnsHopkins @INFORMS @Jillrosen @JeffCohendc @VoicesandWords
Tinglong Dai
@TinglongDai
My discussion focuses on the automotive supply chain:
-
Tinglong Dai
@TinglongDai
Proud to have my paper with @MichaelAbramoff & @OrbisIntl featured in Healthcare AI & Data Science Year in Review
This is probably the 1st time the term "rational queueing theory" has been published in an AI journal, thanks to foundational work by Laurens Debo, @senthil_veer, Refael Hassin, and others @INFORMS
npj Digital Medicine
@npjDigitalMed
We are proud to have seven articles featured in this report of the Healthcare AI & Data Science Year in Review presented by @MeltonMeaux & @yuanhypnosluo. The highest amongst any journal & reflective of the amazing work going on in our #digitalhealth community. See more below...
-
Tinglong Dai
@TinglongDai
-
Tinglong Dai
@TinglongDai
Spoke with @justinterrellho of @NPR's @Marketplace today about the critical role the Port of Baltimore plays in our auto supply chain.
Despite diversions to other ports, Baltimore's unique capabilities remain unmatched.
Listen to the show: www.marketplace.org/2024/03/26/supply-chains-already-feel… -
Tinglong Dai
@TinglongDai
I spoke with @NPR's @camilareads, 4 years after our last conversation, about the #supplychain impact of the Francis Scott Key Bridge collapse, particularly the automotive industry: www.npr.org/2024/03/26/1240974019/mercedes-gm-stellantis-…
Unlike many who see the collapse as accelerating the shift to West Coast ports, I see it as a dent in the @portofbalt continued growth.
The port's upward trajectory is inevitable.
@JHUCarey @JohnsHopkins @INFORMS -
Tinglong Dai reposted
Port of Baltimore
@portofbalt
.@GovWesMoore, @SecretaryPete: Our thoughts are with the families impacted. Our state, local, and federal partners are working closely together.
We will reopen our channel and rebuild the bridge as soon as possible. #keybridgenews #baltimorestrong -
Tinglong Dai
@TinglongDai
Baltimore's Francis Scott Key Bridge has collapsed after being hit by a large cargo ship named Dali
Post not available here, view on X
Tinglong Dai
@TinglongDai
“The vessel notified MD Department of Transportation (MDOT) that they had lost control of the vessel and an allision with the bridge was possible.”
abcnews.go.com/US/live-updates/baltimore-key-bridge-colla…
Tinglong Dai
@TinglongDai
A close-up of Dali and the now collapsed Key Bridge via @ABC
-
Tinglong Dai
@TinglongDai
This is a very sad day for Baltimore, for the US, and for the global shipping industry.
The only consolation is probably the name of the container ship, Dali.
RIP to all the victims. May the Persistence of Memory be with us. -
Tinglong Dai reposted
Tej D. Azad 🇺🇸
@tdazad
A genuine treat to work with @TinglongDai on this OpEd in @MedEconomics on AI & healthcare policy.
www.medicaleconomics.com/view/do-no-harm-the-imperative-f…
@JHU_HBHI @HopkinsNsurg @JHUCarey @AAAS_STPF
Hopkins Med News
@HopkinsMedNews
New @MedEconomics op-ed from @JHUCarey professor Tinglong Dai and @HopkinsNSurg resident Tej D. Azad discusses purposeful AI regulation in health care and policy
www.medicaleconomics.com/view/do-no-harm-the-imperative-f… -
Tinglong Dai
@TinglongDai
DC
-
Tinglong Dai reposted
Shravani Durbhakula, MD, MPH, MBA
@ShravaniD_MD
My ~1000 word Guest Essay in the NYT on how to save #pain #medicine and restore #hope for #patients & #clinicians www.nytimes.com/2024/03/22/opinion/dea-opioids-restrictio… @TinglongDai @drJoshS @TrentDEmerick @Ryan_S_DSouzaMD @JayKarriMD @DrJBhattacharya @JeromeAdamsMD @DoctorJesseMD
-
Tinglong Dai reposted
Hopkins Med News
@HopkinsMedNews
New @MedEconomics op-ed from @JHUCarey professor Tinglong Dai and @HopkinsNSurg resident Tej D. Azad discusses purposeful AI regulation in health care and policy
www.medicaleconomics.com/view/do-no-harm-the-imperative-f… -
Tinglong Dai reposted
Johns Hopkins Neurosurgery
@HopkinsNsurg
An interesting read from @tdazad and @TinglongDai 📰
Hopkins Med News
@HopkinsMedNews
New @MedEconomics op-ed from @JHUCarey professor Tinglong Dai and @HopkinsNSurg resident Tej D. Azad discusses purposeful AI regulation in health care and policy
www.medicaleconomics.com/view/do-no-harm-the-imperative-f… -
Tinglong Dai
@TinglongDai
Super proud to read this compelling @NYTimes op-ed by my former MBA/MPH student @JHUCarey, @ShravaniD_MD, now faculty at @VUmedicine:
nytimes.com/2024/03/22/opi…
What an epic way to start a Friday morning!
@JohnsHopkins @HubJHU @JHU_HBHI
Tinglong Dai
@TinglongDai
Correct link: www.nytimes.com/2024/03/22/opinion/dea-opioids-restrictio…
-
Tinglong Dai reposted
François Chollet
@fchollet
I'd expect an agent to require:
- A sensorimotor space (embodied or virtual)
- An autonomous goal-setting mechanism. Agents follow their own goals.
- A model of self. An agent's goals are centered on the agent itself.
- Decent ability to plan and achieve their goals with no supervision.
The way "agent" is used these days is pretty nonsensical, similar to how "alignment" now just means "fine-tuning a model so it performs better on our benchmarks" (something we've been doing for a long time...) -
Tinglong Dai reposted
Poets&Quants
@PoetsAndQuants
MBA students explore the AI revolution — from basics to real-world applications — in professor Tinglong Dai's AI course at the Carey school.
Read More: poetsandquants.com/2024/03/18/learning-ai-from-its-core-i…
@JHUCarey
#mba #mbadegree #mbastudent #mbaprogram #businessschool #johnhopkins #ai -
Tinglong Dai
@TinglongDai
Delighted to be interviewed by @PoetsAndQuants about the first-of-its-kind core MBA course on Artificial Intelligence that I've been teaching @JHUCarey since 2021: poetsandquants.com/2024/03/18/learning-ai-from-its-core-i…
Tinglong Dai
@TinglongDai
I have been asked about this course more times than I can count:
❓ "You teach #ai at @JohnsHopkins?"
❓"To MBA students?"
❓"And it's mandatory?"
❓ "For three years now?"
❓ "What do you really teach?"
This article sheds some light on these burning questions :-)
Tinglong Dai
@TinglongDai
Many thanks to Maggie Ward, @VoicesandWords, and the marvelous MarComm team at @JHUCarey and Meghan Marrin of @PoetsAndQuants for making this happen!
-
Tinglong Dai reposted
INFORMS
@INFORMS
INFORMS member Tinglong Dai discusses the impact of AI safety on Microsoft's business in this article [TechTarget]
www.techtarget.com/news/Microsoft-whistleblower-OpenAI-th…
#ethicalai #openai #microsoft #informs -
Tinglong Dai
@TinglongDai
Dublin
-
Tinglong Dai
@TinglongDai
Ireland
-
Tinglong Dai
@TinglongDai
Greetings from Dublin
-
Tinglong Dai reposted
The Spectator Index
@spectatorindex
BREAKING: The New York Times reports that Boeing's production process has failed 33 of 89 audits conducted during an FAA probe
-
Tinglong Dai reposted
Helen Toner
@hlntnr
A statement from Helen Toner and Tasha McCauley:
-
Tinglong Dai reposted
Shaun Sutner
@ssutner
@news_techtarget #ai writer @ajao4esther ably dissects the @Microsoft #genai mess including the @nytimes #copyrightlawsuit www.techtarget.com/searchenterpriseai/news/366572699/Micr… thanks @futurebec @TinglongDai #daviottenheimer #aaronrubin
-
Tinglong Dai
@TinglongDai
Join me tomorrow at 2:15pm for our exciting Digital Health panel at the 2024 Hopkins Healthcare Business Conference. 🔥 I'll be discussing big data, big change, and big dreams with
- @strawbrangle, Head of Data & AI at @BlackrockNeuro_;
- @HDDanaLe, CEO & Founder of @wanderhealthco
- @JajaLiao, Principal at @25m_official; and
- @scott_c_taylor, CEO and Co-Founder of @PerxHealth
Register for the event: jhu.campusgroups.com/CBSE/rsvp_boot?id=1947540 -
Tinglong Dai reposted
Austen Allred
@Austen
Somebody told me this was a quote published in Scientific American and I was so skeptical I had to Google it and see.
Unbelievable. -
Tinglong Dai
@TinglongDai
This is absolutely unjustifiable, unacceptable, and unsettling
Denis Wirtz
@deniswirtz
CONGRESS - WAKE UP!
The President's FY 2024 budget request to Congress was $11B for @NSF, an increase of 19.0% from FY23.
Instead, it is slated to decrease by 5%.
This is catastrophic for our nation's pipeline of ideas in science and engineering. -
Tinglong Dai reposted
Historic Vids
@historyinmemes
Steve Jobs on Al in 1985
-
Tinglong Dai reposted
François Chollet
@fchollet
"Accelerating progress" should mean increasing the breadth and depth of our search for meaning & solutions. Increasing the diversity and creativity of our cultural production and consumption.
To a large extent, I feel like generative AI achieves the opposite. It collapses everything into a narrower, standardized, derivative space. It might even result in cultural stagnation if left unchecked.
Not for too long though, since people will simply grow out of it, eventually. Humans have a built-in drive towards interestingness. They get bored quickly. -
Tinglong Dai reposted
Youngjin Yoo
@YoungjinYoo
I like Herbert Simon even better now! #leadingwithai @caseweatherhead
-
Tinglong Dai reposted
Youngjin Yoo
@YoungjinYoo
The What, the Why and the How of Scaling effective and safe medical AI by @TinglongDai at #leadingwithai @caseweatherhead
-
Tinglong Dai reposted
Youngjin Yoo
@YoungjinYoo
What would you do if your doctor uses ChatGPT during your consultation? Would you trust your doctor? 🤔 @TinglongDai #leadingwithai @caseweatherhead
-
Tinglong Dai reposted
Youngjin Yoo
@YoungjinYoo
The most popular CPT code was used a little over 10,000 times over the last 5 years out 11 billion total billings. How can we accelerate the use of medical AI? @TinglongDai #leadingwithai @caseweatherhead
-
Tinglong Dai
@TinglongDai
By Emily Dickinson:
-
Tinglong Dai
@TinglongDai
Cleveland
-
Tinglong Dai reposted
Shengwu Li
@ShengwuLi
While on the tenure track, I had to flee a small authoritarian regime 🇸🇬 because I spoke my mind on social media. Tenure doesn't change anything. You've seen the real me all along.
Luis Garicano 🇪🇺🇺🇦
@lugaricano
Congratulations! Now, are we going to see the real you here ?😉
-
Tinglong Dai reposted
Jerome Adams
@JeromeAdamsMD
Recently had an ER visit for dehydration while out of town. Received some labs and 3 IV bags. Here's the bill- AFTER insurance. It's no wonder medical debt is the top cause of bankruptcy in the U.S. 🏥💸 #healthcarecosts #medicaldebtcrisis
-
Tinglong Dai
@TinglongDai
.@Peggynoonannyc can write
-
Tinglong Dai reposted
Weatherhead School
@caseweatherhead
Do you want to see how business school research can help your organization’s journey with AI? Learn from Prof. Tinglong Dai, a leading voice at the intersection of healthcare and AI, as he speaks at our AI Symposium. #aiedrevolution Sign up: www.eventbrite.com/e/leading-in-the-age-of-artificial-int…
-
Tinglong Dai reposted
Johns Hopkins University
@JohnsHopkins
148 years of innovation and unity. Together, we're creating a legacy of impact and synergy.
Happy birthday, Johns Hopkins!This media couldn't be displayed. See it on the original post. -
Tinglong Dai reposted
Weatherhead School
@caseweatherhead
Is your strategy AI-ready? Engage with leading thinkers at the Weatherhead #ai Symposium on March 1st to navigate new strategic landscape! Discover how data-driven decision-making and operational agility can become your competitive edge. #techinnovation
www.eventbrite.com/e/leading-in-the-age-of-artificial-int… -
Tinglong Dai reposted
Douglas A. Boneparth
@dougboneparth
"Dad, why is my sister's name Rose?"
"Because your mother loves roses."
"Thanks, Dad."
"You're welcome, Nvidia." -
Tinglong Dai
@TinglongDai
Wow
-
Tinglong Dai reposted
Paul Graham
@paulg
The ridiculous images generated by Gemini aren't an anomaly. They're a self-portrait of Google's bureaucratic corporate culture.
-
Tinglong Dai
@TinglongDai
Jensen Huang is the most underrated - or least overrated - motivational speaker of our time
-
Tinglong Dai reposted
Michael D Abramoff
@MichaelAbramoff
Good points, but then will be reactive / defensive. That may not be cutting it anymore in the future as we move more into autonomous AI. In the past we saw total shutdown of gene therapy, etc because of overreaction to unethical behavior; also the issues around children & social media.
Instead, proactively start with an ethical framework, for (autonomous) AI such as
doi.org/10.1080/15265161.2020.1819469
and proactively develop regulatory / bias / reimbursement guardrails for AI from that, as we & others are doing:
pubmed.ncbi.nlm.nih.gov/34478784/
www.nature.com/articles/s41746-023-00913-9
www.nature.com/articles/s41746-022-00621-w -
Tinglong Dai reposted
Yann LeCun
@ylecun
Modeling the world for action by generating pixel is as wasteful and doomed to failure as the largely-abandoned idea of "analysis by synthesis".
Decades ago, there was a big debate in ML about the relative advantages of generative methods vs discriminative methods for classification.
Learning theorists, such as Vapnik, argued against generative methods, pointing out that training a generative modeling was a way more difficult than classification (from the sample complexity standpoint).
Regardless, a whole community in computer vision was arguing that recognition should work by generating pixels from explanatory latent variables. At inference time, one would infer the configuration of latent variables that generated the observed pixels.
The inference method would use optimization: e.g. use a 3D model of an object and try to find the pose parameters that reproduce the image.
This never quite worked, and it was very slow.
Later, some people converted to the Bayesian religion and tried to use Bayesian inference for the latent (e.g. using variational approximations and/or sampling).
At some point, when Non-Parametric Bayes and Latent Dirichlet Allocation became the rage in text modeling, some folks heroically attempted to apply that to object recognition from images.
>>> THIS WAS A COMPLETE AND UTTER FAILURE <<<
If your goal is to train a world model for recognition or planning, using pixel-level prediction is a terrible idea.
Generation happens to work for text because text is discrete with a finite number of symbols. Dealing with uncertainty in the prediction is easy in such settings. Dealing with prediction uncertainty in high-dimension continuous sensory inputs is simply intractable.
That's why generative models for sensory inputs are doomed to failure. -
Tinglong Dai reposted
Q-Cap
@qcapital2020
The faith of the entire US economy
-
Tinglong Dai reposted
François Chollet
@fchollet
Ironically, Transformers are even worse in that regard -- mostly due to their strongly interpolative architecture prior. Multi-head-attention literally hardcodes sample interpolation in latent space. Also, the fact that recurrence is a really helpful prior for symbolic programs.
-
Tinglong Dai
@TinglongDai
Baltimore
-
Tinglong Dai reposted
Matthew E. Kahn
@mattkahn1966
I started my career as a labor economist but was fired by an important University of Chicago labor economist for being an incompetent RA. Now that I'm an older scholar, I am making my comeback!
Adam Grant
@AdamMGrant
A neglected benefit of hybrid work: it improves career options for moms.
As remote work expands, mothers earn more and are more likely to be employed—especially in family-unfriendly fields like finance.
Flexibility doesn't just facilitate balance. It broadens opportunity. -
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
Johns Hopkins Carey Business School
@JHUCarey
The first Bernard T. Ferrari Professorship in Business goes to a renowned expert on AI and supply chain, @TinglongDai.
carey.jhu.edu/articles/new-professorship-honor-dean-emeritus -
Tinglong Dai reposted
Andrea Presbitero
@a_presbitero
🚨🚨Our paper showing that #fdi flows are increasingly concentrated among geopolitically aligned countries is forthcoming in the European Journal of Political Economy @davmalacrino @ShekharAiyar #geoeconomics #friendshoring
ungated draft: www.dropbox.com/scl/fi/1n0hdvmt3ie170oge9tfj/Aiyar-etal_f… -
Tinglong Dai
@TinglongDai
-
Tinglong Dai reposted
Michael D Abramoff
@MichaelAbramoff
At high level, such proof already exists, albeit in diabetes care and not yet in cancer: see our recent paper on an RCT showing how autonomous AI decreased health disparities:
www.nature.com/articles/s41467-023-44676-z
@TinglongDai -
Tinglong Dai reposted
Japan Embassy DC🌸
@JapanEmbDC
🇯🇵 Statement from the Embassy of Japan on Taylor Swift’s Reported Travel from Japan to the United States ✈️🏈 Are you ready for it?
-
Tinglong Dai reposted
Hopkins Business of Health Initiative
@JHU_HBHI
HBHI is so very proud of @TinglongDai, The Ferrari Professor.
Tinglong Dai
@TinglongDai
Friends keep asking me, in disbelief, "What? @JohnsHopkins is giving you a Ferrari?" To which I reply, "Not just a Ferrari. It's the Ferrari. Even better, I'll be Ferrari."
It takes a village to raise a child; in my case, it took an entire university, a family, and a community of friends, colleagues, & mentors to shape the Ferrari Professor of Business I am at @JHUCarey:
www.linkedin.com/pulse/bernard-ferrari-professorship-acce… -
Tinglong Dai
@TinglongDai
Friends keep asking me, in disbelief, "What? @JohnsHopkins is giving you a Ferrari?" To which I reply, "Not just a Ferrari. It's the Ferrari. Even better, I'll be Ferrari."
It takes a village to raise a child; in my case, it took an entire university, a family, and a community of friends, colleagues, & mentors to shape the Ferrari Professor of Business I am at @JHUCarey:
www.linkedin.com/pulse/bernard-ferrari-professorship-acce…
Tinglong Dai
@TinglongDai
Bernard T. Ferrari Professorship Acceptance Speech
Tinglong Dai, Ferrari Professor of Business
January 18, 2024
Thank you for your extremely generous and inspiring introduction, Professor Hau Lee!
President Ron Daniels, Dean and Mrs. Ferrari, Dean Alex Triantis, Dean @EllenJMacKenzie, Dean @Sarah_Szanton, my family, dear friends, colleagues, students, alumni, and distinguished guests,
I am so honored, humbled, and proud to be named Bernard T. Ferrari Professor of Business at @JohnsHopkins. My daughter, Audrey, is here. Audrey was born in 2013, right after I joined Hopkins. When I held her in my arms as a new assistant professor, I never thought that I would be standing here accepting this endowed professorship.
And it’s not just an endowed professorship. It’s the Ferrari professorship.
Friends keep asking me, in disbelief, "What? Johns Hopkins is giving you a Ferrari?"
To which I reply, "Not just a Ferrari. It’s the Ferrari. Even better — I’ll be Ferrari."
This wasn’t even in my wildest dreams.
But that's the thing about this magical place known as @JohnsHopkins.
Hopkins is where medical marvels happen day after day. But what many people don’t know is that we do all kinds of marvels, not just medical.
@JHUCarey, the youngest school in America’s oldest research university, is one such marvel.
I have been so honored to be part of Carey since its early days.
Today, I will talk about three things I marvel at.Post not available here, view on X
Tinglong Dai
@TinglongDai
As I conclude these remarks, my heart is filled with gratitude for the many people who have placed their trust in me, offered their support, and lent a helping hand.
First and foremost is my family sitting here: my wife, Ricci, who has been my lover and best friend since our sophomore year 23 years ago when we did computer science minor together, and our children, Audrey and Carl.
I owe a world of thanks to my parents and siblings for their unwavering support and the autonomy they gave me; to Dean Bernard Ferrari, Dean Alexander Triantis, Vice Deans Goker Aydin, Kevin Frick, Phil Phan, and Valerie Suslow, @JHU_HBHI Director Daniel Polsky, and all my esteemed colleagues at @JHUCarey and across @JohnsHopkins. My Ph.D. advisors, Sridhar Tayur and Katia Sycara, have been absolutely instrumental in my journey.
I extend my deepest gratitude to all my friends, collaborators, mentors — including Prof. Hau Lee, who traveled all way from California to Baltimore to introduce me, plus Prof. Christopher Tang, two supply chain giants born and raised in Hong Kong, my spiritual hometown — and the @INFORMS and @PomsSociety communities, and so many others, not to mention @JHUCarey alumni and supporters, who have made this endowed professorship and more possible.
It takes a village to raise a child; in my case, it took an entire university to shape the Ferrari Professor of Business I am today. As I embark on this new chapter, I’m eager to contribute to the enduring legacy of @JohnsHopkins and the dynamic, thriving culture of @JHUCarey.
Thank you all for this incredible honor. Let’s move forward and move fast together in this race, breaking new ground and reaching new heights. Thank you!
Tinglong Dai
@TinglongDai
Link to the full speech: www.linkedin.com/pulse/bernard-ferrari-professorship-acce…
-
Tinglong Dai reposted
M Eric Johnson
@MEricJohnson
“AI is a fantastic opportunity to address supply chain complexity.” @StanfordGBS Professor Haul Lee on a panel with @Wharton Professor Morris Cohen at the @MSBGU Workshop on Rethinking Global Supply Chains in the Era of AI – hosted by @TinglongDai and Ho-Yin Mak.
-
Tinglong Dai reposted
Curt Langlotz
@curtlanglotz
“The assessment of the algorithms needs to be continuous,” @US_FDA Commissioner Robert Califf, MD, told a packed room at the #ces2024 show in Las Vegas. He added, post-market evaluations of new tools may be more important than pre-market evaluations. shar.es/agfZJ1
-
Tinglong Dai reposted
Denis Wirtz
@deniswirtz
Johns Hopkins leads nation in research spending for 44th consecutive year ($3.4 billion dollars in FY 2022)
This is a testament of the amazing researchers who, through hard work and creativity, make important contributions to many fields of research hub.jhu.edu/2024/01/05/nsf-higher-education-research-spen… -
Tinglong Dai reposted
Mario Macis
@Mario_Macis
❗️Why Can't Americans Have Better and Cheaper Health Care?
Join us next Friday, January 19, at 12 PM ET for a @JHU_HBHI webinar with @JimRebitzer and @Breb1, authors of "Why Not Better and Cheaper? Healthcare and Innovation"
Register here:
hbhi.jhu.edu/events/why-cant-americans-have-better-and-ch…
@JHUCarey -
Tinglong Dai reposted
Patricia Davidson
@DrPMDavidson
Fabulous to catch up with my friend and collaborator @JHUCarey #beautifulbaltimore @JHUNursing @JohnsHopkinsSPH @UOW
-
Tinglong Dai reposted
Crémieux
@cremieuxrecueil
The New York times has just published a story supporting standardized testing.
Thread -
Tinglong Dai reposted
Joyce Carol Oates
@JoyceCarolOates
Socrates conducted the first extension school. see what happened to Socrates.
Oliver Traldi
@olivertraldi
philosophy consists in a series of extension schools to Plato
-
Tinglong Dai
@TinglongDai
Rare good news for the US healthcare system
PhRMA's intention to fight it confirms its goodness
Sheryl Gay Stolberg
@SherylNYT
SCOOP: The FDA has decided to allow Florida to import drugs in bulk from Canada. Other states may follow. It’s a major policy shift, 20 yrs in the making, and POTUS campaign issue. Also a win for Ron DeSantis. PhRMA will fight it. w/@By_CJewett www.nytimes.com/2024/01/05/health/drug-imports-canada-flo…
-
Tinglong Dai
@TinglongDai
Heading to my first #aom2023 in Boston! Join @Soroush_Saghaf (@Harvard @PIASLab) and me for the #ai and #ml Roundtable, part of the "Hot Methods" PDW, Sat 12:15-1:45PM, @SheratonBoston, Independence East.
Thanks @WiljeanaGlover & Soo-Hoon Lee for inviting us! @AOMConnect #pdw -
Tinglong Dai
@TinglongDai
Honored to address @teppercmu '23 PhDs at @CarnegieMellon. Thanks to my mentors Profs Sridhar Tayur & Katia Sycara, and to @RappStar4 for making this happen. Pleasure meeting Dean @Isabelle_Bajeux. Kudos to all graduates! Your PhDs are a superpower, use them to better the world!
-
Tinglong Dai
@TinglongDai
Is it legal for @OpenAI to use copyrighted material to train #chatgpt? My @JHUCarey colleague @KathleenDay just published an insightful piece on this topic for @cqresearcher, and I had the honor of contributing my thoughts alongside @C_Stihler, CEO of @CreativeCommons: library.cqpress.com/cqresearcher/document.php?id=cqresrre…
What is the right balance between fair use and intellectual property rights in AI development? Your thoughts and perspectives are welcome: -
Tinglong Dai
@TinglongDai
This week, @BLaw published an article I co-wrote with @murphy_laur of @ICRW about giving women supply chain workers top priority in #esg initiatives: tinglongdai.com/wp-content/uploads/Prioritizing-Women-Sup…










































































































































































































































































































































































































































































































