Robot "dog" 🤝 Robot "duck"
Video v/@cetiennec
MIT Computer Science and Artificial Intelligence Laboratory
Computer Science and Artificial Intelligence Laboratory (CSAIL) is a research institute at the Massachusetts Institute of Technology (MIT) formed by the 2003 merger of the Laboratory for Computer Science (LCS) and the Artificial Intelligence Laboratory (AI Lab). Housed within the Ray and Maria Stata Center, CSAIL is the largest on-campus laboratory as measured by research scope and membership. It is part of the Schwarzman College of Computing but is also overseen by the MIT Vice President of Research. More on Wikipedia
-
MIT CSAIL
@MIT_CSAIL
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Brass Rats (MIT class rings) over the years: 2026, 1996, 1963, & 1930.
Images v/MIT Museum & MIT. -
MIT CSAIL
@MIT_CSAIL
Happy birthday to MIT prof. Shafi Goldwasser, pioneer in the field of cryptography.
Throwback to when she received the Turing Award: news.mit.edu/2013/goldwasser-and-micali-win-turing-award-… -
MIT CSAIL
@MIT_CSAIL
"Telling a programmer there’s already a library to do X is like telling a songwriter there’s already a song about love." — Pete Cordell
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
A new AI system called "WorldParticle" simulates everything from flowing fluids & sliding sand to folding proteins w/one unified architecture.
It's a deep-learning approach that learns how matter moves, making it potentially useful in developing things like soft robots & medicines: news.mit.edu/2026/single-ai-model-simulates-many-substanc… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Brass Rats (MIT class rings) over the years: 2026, 1996, 1963, & 1930.
Images v/MIT Museum & MIT. -
MIT CSAIL
@MIT_CSAIL
A new AI system called "WorldParticle" simulates everything from flowing fluids & sliding sand to folding proteins w/one unified architecture.
It's a deep-learning approach that learns how matter moves, making it potentially useful in developing things like soft robots & medicines: news.mit.edu/2026/single-ai-model-simulates-many-substanc… -
MIT CSAIL
@MIT_CSAIL
Brass Rats (MIT class rings) over the years: 2026, 1996, 1963, & 1930.
Images v/MIT Museum & MIT. -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
A free breakdown of all things math, including the questions you didn't know to ask: www.math.cmu.edu/~jmackey/151_128/bws_book.pdf
-
MIT CSAIL
@MIT_CSAIL
A free breakdown of all things math, including the questions you didn't know to ask: www.math.cmu.edu/~jmackey/151_128/bws_book.pdf
-
MIT CSAIL
@MIT_CSAIL
ENIAC, the world's first electronic, general-purpose digital computer, at work in its early days. The machine was introduced back in 1946.
Images v/UPenn -
MIT CSAIL
@MIT_CSAIL
15 courses for understanding neural networks, v/@tut_ml shorturl.at/60Se8
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Free book on the foundations of LLMs, from Northeastern University & NiuTrans Research: arxiv.org/abs/2501.09223
-
MIT CSAIL
@MIT_CSAIL
Free book on the foundations of LLMs, from Northeastern University & NiuTrans Research: arxiv.org/abs/2501.09223
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Michael Dertouzos, who directed MIT’s Laboratory for Computer Science for 25+ years.
Watch him discuss the future of computers on NBC’s Today Show w/late MIT prof. Joseph Weizenbaum back in 1983. -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Michael Dertouzos, who directed MIT’s Laboratory for Computer Science for 25+ years.
Watch him discuss the future of computers on NBC’s Today Show w/late MIT prof. Joseph Weizenbaum back in 1983. -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
#otd in 1971 Bell Labs released the first Unix Programmer's Manual.
Download the free PDF here (third edition): shorturl.at/8NP0j
(v/@CoolSWEng) -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Frontier LLMs are showing potential to read, fix, & rewrite neural network weights.
MIT CSAILers call this "Weight-Space Programming." It's what GPT-6 did when it edited various small NNs including Qwen3.5-0.8B, using tools in interpretability research or even direct weight editing. Soon enough, LLMs might natively optimize NN weights at a human or even super-human level: rikiyat.github.io/blog/weight-space-programming/ -
MIT CSAIL
@MIT_CSAIL
Frontier LLMs are showing potential to read, fix, & rewrite neural network weights.
MIT CSAILers call this "Weight-Space Programming." It's what GPT-6 did when it edited various small NNs including Qwen3.5-0.8B, using tools in interpretability research or even direct weight editing. Soon enough, LLMs might natively optimize NN weights at a human or even super-human level: rikiyat.github.io/blog/weight-space-programming/
MIT CSAIL
@MIT_CSAIL
Full thread:
Rikiya Takehi
@rikiyatakehi
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weight-space-programming/ -
MIT CSAIL
@MIT_CSAIL
#otd in 1971 Bell Labs released the first Unix Programmer's Manual.
Download the free PDF here (third edition): shorturl.at/8NP0j
(v/@CoolSWEng) -
MIT CSAIL reposted
Rikiya Takehi
@rikiyatakehi
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weight-space-programming/ -
MIT CSAIL
@MIT_CSAIL
What if an optimizer could think & react while a neural network trains?
MIT researchers developed OPAL: a deep learning optimizer powered by an AI agent that watches the training dynamics & rewrites its own source code as training goes. It improves model performance across vision & language tasks, invents specialized schedules, & develops strategies to prevent loss spikes & escape overfitting: xingjianbai.com/reactive-optimizers/
Try it out for yourself: github.com/xingjian-bai/opal -
MIT CSAIL
@MIT_CSAIL
It's not easy being Pi.
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Happy Halloween! Even MIT’s Great Dome wears a costume sometimes — Halloween or not.
-
MIT CSAIL
@MIT_CSAIL
Happy Halloween! Even MIT’s Great Dome wears a costume sometimes — Halloween or not.
-
MIT CSAIL
@MIT_CSAIL
A map of ARPANET back in 1977. It was the first large-scale computer network, connecting institutions like MIT & Stanford. The system was a predecessor to the Internet.
v/The Computer History Museum -
MIT CSAIL
@MIT_CSAIL
Richard P. Feynman '39 talks about quantum mechanics during a 1964 lecture at Cornell.
Video v/@PhysInHistory & BBC -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
MIT & TJU's new 3D printing tool reimagines simple items as self-aware devices.
You can design decor & even robots in "X-Hinges." It'll embed sensors & conductive traces using different materials to send clear data, enabling real-time interaction w/users: news.mit.edu/2026/design-software-reimagines-everyday-obj… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
MIT & TJU's new 3D printing tool reimagines simple items as self-aware devices.
You can design decor & even robots in "X-Hinges." It'll embed sensors & conductive traces using different materials to send clear data, enabling real-time interaction w/users: news.mit.edu/2026/design-software-reimagines-everyday-obj… -
MIT CSAIL
@MIT_CSAIL
MIT & TJU's new 3D printing tool reimagines simple items as self-aware devices.
You can design decor & even robots in "X-Hinges." It'll embed sensors & conductive traces using different materials to send clear data, enabling real-time interaction w/users: news.mit.edu/2026/design-software-reimagines-everyday-obj… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Growing up w/vibe coding, v/@whatRanjuSaid.
-
MIT CSAIL
@MIT_CSAIL
Growing up w/vibe coding, v/@whatRanjuSaid.
-
MIT CSAIL
@MIT_CSAIL
If you could time travel to see one scientific breakthrough (past or future), what would you choose?
-
MIT CSAIL
@MIT_CSAIL
A great Halloween costume idea (v/@PR0GRAMMERHUM0R).
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
AI is quickly improving at tough theory questions in math, physics, & economics.
In a new MIT study, frontier OpenAI models’ success rate rose from 3% to 35% in 16 months, while a successful attempt cost about $15. The new bottleneck may be verification: futuretech.mit.edu/publication/ai-disruption-and-falling-… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Want genAI to make you a functional, real-world object? Good luck — their 3D models are often physically flawed & hard to edit.
New "InstructMesh" tool makes it easy to generate, tweak, & 3D print things that actually work, such as decor & medical devices: news.mit.edu/2026/instructmesh-tool-lets-users-repair-ai-… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
The making of Stata, back in the early 2000s.
Images v/Zahner -
MIT CSAIL
@MIT_CSAIL
The making of Stata, back in the early 2000s.
Images v/Zahner -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
BOOlean logic to brush up on before Halloween.
v/@38mo1 -
MIT CSAIL
@MIT_CSAIL
BOOlean logic to brush up on before Halloween.
v/@38mo1 -
MIT CSAIL
@MIT_CSAIL
ETH Zurich recently taught a robotic hand to walk: arxiv.org/abs/2609.17172
Its clever design & fluid motion reminded us of another walking hand: Thing from The Addams Family. -
MIT CSAIL
@MIT_CSAIL
An IBM slide from 1979.
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Want genAI to make you a functional, real-world object? Good luck — their 3D models are often physically flawed & hard to edit.
New "InstructMesh" tool makes it easy to generate, tweak, & 3D print things that actually work, such as decor & medical devices: news.mit.edu/2026/instructmesh-tool-lets-users-repair-ai-… -
MIT CSAIL
@MIT_CSAIL
Back when calculators were a little more controversial, v/@JonErlichman.
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Want genAI to make you a functional, real-world object? Good luck — their 3D models are often physically flawed & hard to edit.
New "InstructMesh" tool makes it easy to generate, tweak, & 3D print things that actually work, such as decor & medical devices: news.mit.edu/2026/instructmesh-tool-lets-users-repair-ai-… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
AI is quickly improving at tough theory questions in math, physics, & economics.
In a new MIT study, frontier OpenAI models’ success rate rose from 3% to 35% in 16 months, while a successful attempt cost about $15. The new bottleneck may be verification: futuretech.mit.edu/publication/ai-disruption-and-falling-… -
MIT CSAIL
@MIT_CSAIL
AI is quickly improving at tough theory questions in math, physics, & economics.
In a new MIT study, frontier OpenAI models’ success rate rose from 3% to 35% in 16 months, while a successful attempt cost about $15. The new bottleneck may be verification: futuretech.mit.edu/publication/ai-disruption-and-falling-… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
The late MIT prof. Patrick Winston explains how to speak & get your ideas valued & accepted.
The famous lecture in full: www.youtube.com/watch?v=Unzc731iCUY
v/@MITOCW -
MIT CSAIL reposted
Venki Parsuram
@venki1899
As humanoid robots become part of our everyday lives, they need to do more than just perform impressive motions. They should be safe, fun to interact with, and respond naturally to unexpected contact, without sacrificing performance.
Our latest work, Compliant GMT (C-GMT), brings whole-body compliant responses to general motion tracking.
🌐 More details: venki1899.github.io/c-gmt/
Work done at MIT Improbable AI Lab with @michl_wang, Jin, @pulkitology.This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
The late MIT prof. Patrick Winston explains how to speak & get your ideas valued & accepted.
The famous lecture in full: www.youtube.com/watch?v=Unzc731iCUY
v/@MITOCW -
MIT CSAIL
@MIT_CSAIL
Scientists have lacked the data to bring physical AI underwater. MIT CSAIL researchers are helping robots change that using sonar & ML approaches: www.youtube.com/watch?v=i0O2URp0HpM
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Want genAI to make you a functional, real-world object? Good luck — their 3D models are often physically flawed & hard to edit.
New "InstructMesh" tool makes it easy to generate, tweak, & 3D print things that actually work, such as decor & medical devices: news.mit.edu/2026/instructmesh-tool-lets-users-repair-ai-… -
MIT CSAIL
@MIT_CSAIL
Want genAI to make you a functional, real-world object? Good luck — their 3D models are often physically flawed & hard to edit.
New "InstructMesh" tool makes it easy to generate, tweak, & 3D print things that actually work, such as decor & medical devices: news.mit.edu/2026/instructmesh-tool-lets-users-repair-ai-… -
MIT CSAIL
@MIT_CSAIL
To give context on age:
Grace Hopper began computing at 38, completed the first compiler at 46, helped shape COBOL at 53, kept developing COBOL for the Navy in her 70s, retired from the Navy at 80, & then became a consultant for the Digital Equipment Corporation.
v/@cooperx86 -
MIT CSAIL
@MIT_CSAIL
How to make sure your code keeps getting reused, v/@xkcd.
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
"DREAM" is an AI model from MIT & Meta that can both understand images & generate them from text — then judge its own rough drafts to finish the best one.
The result: stronger vision, better image quality, & ~10% faster generation w/external re-rankers: csailmit.substack.com/p/one-model-to-see-and-imagine-unif… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Computer scientist Margaret Hamilton passed away at 90 on Sep. 30th. Considered an icon at MIT, she designed the flight software for the Apollo missions to the moon.
More on her groundbreaking work: news.mit.edu/2026/margaret-hamilton-computing-pioneer-die…
Video v/MAKERS -
MIT CSAIL
@MIT_CSAIL
Wise words from Geoffrey Hinton: A PhD in computer science is worth it.
v/@samuelxzhang -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
"DREAM" is an AI model from MIT & Meta that can both understand images & generate them from text — then judge its own rough drafts to finish the best one.
The result: stronger vision, better image quality, & ~10% faster generation w/external re-rankers: csailmit.substack.com/p/one-model-to-see-and-imagine-unif… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Computer scientist Margaret Hamilton passed away at 90 on Sep. 30th. Considered an icon at MIT, she designed the flight software for the Apollo missions to the moon.
More on her groundbreaking work: news.mit.edu/2026/margaret-hamilton-computing-pioneer-die…
Video v/MAKERS -
MIT CSAIL
@MIT_CSAIL
Computer scientist Margaret Hamilton passed away at 90 on Sep. 30th. Considered an icon at MIT, she designed the flight software for the Apollo missions to the moon.
More on her groundbreaking work: news.mit.edu/2026/margaret-hamilton-computing-pioneer-die…
Video v/MAKERS -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Left: MIT computer scientist Katie Bouman w/stacks of hard drives of black hole image data.
Right: MIT computer scientist Margaret Hamilton w/the code she wrote that helped put a man on the moon.
(image credit @floragraham)
#ehtblackhole #blackholeday #blackhole -
MIT CSAIL reposted
Massachusetts Institute of Technology (MIT)
@MIT
We’ve lost an icon. Margaret Hamilton, a computer science pioneer best known for her work at MIT leading development of the flight software for the Apollo missions to the moon, has died at 90. Read about her legacy via MIT News: news.mit.edu/2026/margaret-hamilton-computing-pioneer-die…
-
MIT CSAIL
@MIT_CSAIL
What happens when you put an AI coding assistant on an island?
MIT researchers put 13 agents in this digital space w/no task or reward. Over 30 hours, they invented challenges, experiments, sports, & art — & the skills they found helped them accomplish more later: arxiv.org/abs/2610.07130 -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
"DREAM" is an AI model from MIT & Meta that can both understand images & generate them from text — then judge its own rough drafts to finish the best one.
The result: stronger vision, better image quality, & ~10% faster generation w/external re-rankers: csailmit.substack.com/p/one-model-to-see-and-imagine-unif… -
MIT CSAIL reposted
Nitish Dashora @ CoRL
@DashoraNitish
There’s lots of buzz around agentic harnesses for robots, particularly for long horizon tasks that require complex reasoning and memory. But what will it really take to turn a reasoning agent into a reactive, reliable and low-latency robot policy?
In our new paper, workspace models, we design a new memory architecture that acts as a latent harness for stronger reasoning models. Here’s why we think this might be the way forward:
(1/9) -
MIT CSAIL
@MIT_CSAIL
14 sorting algorithms in 60 seconds.
v/@7etsuo -
MIT CSAIL reposted
Sophie Wang
@SophieLWang
Can “chicken” make a base model reason? 🐔
Yes!
Why? With the right first tokens, base models can match RL in reasoning. We trace this to learned training data associations and use a simple data edit to make “chicken” elicit reasoning too.
blog: sophielwang.com/cues -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
On many '90s PCs, your system would give you the go-ahead to power down.
Image v/@AYTOarchive -
MIT CSAIL reposted
Kush Tiwary
@ktiwarylab
There is a new kind of robotics called "Vibe Robotics"
It's when you ask agents for fully functioning embodied agents (both hardware and software) that must solve a given task in physically grounded environments.
We build an evaluation to test how good agents (frontier models) are at "Vibe Robotics".
Website: artifactarena.aiThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
"DREAM" is an AI model from MIT & Meta that can both understand images & generate them from text — then judge its own rough drafts to finish the best one.
The result: stronger vision, better image quality, & ~10% faster generation w/external re-rankers: csailmit.substack.com/p/one-model-to-see-and-imagine-unif… -
MIT CSAIL
@MIT_CSAIL
How some computers were wired before PCBs became common.
v/@alphafox -
MIT CSAIL
@MIT_CSAIL
On many '90s PCs, your system would give you the go-ahead to power down.
Image v/@AYTOarchive -
MIT CSAIL
@MIT_CSAIL
What do you think robots will be doing in 2050?
-
MIT CSAIL
@MIT_CSAIL
#otd in 1950, the Turing Test was born w/Alan Turing’s seminal paper "Computing Machinery and Intelligence."
academic.oup.com/mind/article/LIX/236/433/986238 -
MIT CSAIL reposted
Phillip Isola
@phillip_isola
Students sometimes ask me if it still makes sense, in this accelerating age, to pursue a PhD in AI.
Perhaps counterintuitively, I think it's a great time to do so.
I wrote up some thoughts on this here: web.mit.edu/phillipi/www/writing/PhD-in-age-of-AI.html -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Daniela Rus, CSAIL Director & MIT Prof., says that after automation, "intelligence will live in your pocket, not a data center."
"It is a path toward AI that is powerful w/o being wasteful, ubiquitous w/o being intrusive, & personal w/o requiring surveillance."
Read her full thoughts, which are part of @every's Thesis conference in Nov.: every.to/thesis-statements/daniela-rus -
MIT CSAIL
@MIT_CSAIL
Daniela Rus, CSAIL Director & MIT Prof., says that after automation, "intelligence will live in your pocket, not a data center."
"It is a path toward AI that is powerful w/o being wasteful, ubiquitous w/o being intrusive, & personal w/o requiring surveillance."
Read her full thoughts, which are part of @every's Thesis conference in Nov.: every.to/thesis-statements/daniela-rus -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
"Controlling complexity is the essence of computer programming." — Brian Kernighan
-
MIT CSAIL
@MIT_CSAIL
"Controlling complexity is the essence of computer programming." — Brian Kernighan
-
MIT CSAIL
@MIT_CSAIL
Google launched #otd in 1998.
It looked a bit different 28 years ago. -
MIT CSAIL
@MIT_CSAIL
In 2006, then-CNBC reporter Erin Burnett mentioned the idea of an "iPhone" to Steve Jobs. The product was in development at the time, but not officially announced.
Video v/@historyinmemes -
MIT CSAIL
@MIT_CSAIL
The 25 most important mathematical definitions in data science, v/@DailyDoseOfDS_.
-
MIT CSAIL reposted
Massachusetts Institute of Technology (MIT)
@MIT
Did you know the original Claude — the chatbot’s namesake, according to Anthropic lore — was a renowned MIT professor? Claude Shannon SM ’40, PhD ’40, who taught at MIT from 1956 until his retirement in 1978, was known to many as the father of information theory and of modern digital communications.
-
MIT CSAIL
@MIT_CSAIL
A list explaining key ML concepts and papers: github.com/dair-ai/ML-Papers-Explained
v/@omarsar0 -
MIT CSAIL
@MIT_CSAIL
What is physical AI? MIT CSAIL explores how it's already helping robots understand the real world — & where it's heading next: www.youtube.com/watch?v=L6bSqc5ccCw
-
MIT CSAIL
@MIT_CSAIL
U.S. News and World Report has named MIT the top university in the US.
The Institute was ranked No. 1 in undergraduate engineering, computer science, & economics programs: news.mit.edu/2026/mit-named-nations-top-university-us-new… -
MIT CSAIL
@MIT_CSAIL
A handy guide to Git commands, from Git: git-scm.com/cheat-sheet
-
MIT CSAIL
@MIT_CSAIL
#otd 24 years ago, Tor was launched by MIT students for anonymous communications: www.torproject.org/about/history/
-
MIT CSAIL
@MIT_CSAIL
What was your first handheld device (iPod, Walkman, etc)?
-
MIT CSAIL
@MIT_CSAIL
Notable AI models by country, top 3:
1. USA: 59 🇺🇸
2. China: 35 🇨🇳
3. South Korea: 8 🇰🇷
v/Stanford's AI Index 2026 Annual Report -
MIT CSAIL reposted
Arthur Turrell
@arthurturrell
Today @Google, we published work led by @m_codreanu on AI in Science, drawing on 15 million Gemini interactions, 2,600 specialised AI models & a 600-scientist survey. What did we find about how AI is changing science? Read on...
-
MIT CSAIL
@MIT_CSAIL
AI is already reshaping scientific research, according to an MIT & Google study.
Scientists use LLMs & specialized AI models for different tasks, save ~7 hours a week, & reinvest much of that time in research — but new bottlenecks are emerging: blog.google/innovation-and-ai/technology/ai/ai-economy-at…
MIT CSAIL
@MIT_CSAIL
-
MIT CSAIL
@MIT_CSAIL
What does scientific work actually look like?
MIT's "Science Task Taxonomy" maps tasks across research subfields & sectors, showing how scientific work differs from the wider economy — w/important implications for the impact of AI on science: futuretech.mit.edu/publication/scientific-work
MIT CSAIL
@MIT_CSAIL
-
MIT CSAIL
@MIT_CSAIL
Throwback: MIT prof. Peter Shor gives advice to aspiring computer scientists.
Full video: www.youtube.com/watch?v=wnhZPmB8KLg -
MIT CSAIL
@MIT_CSAIL
A guide to statistics for beginners: www.researchgate.net/profile/Arumugam-Raman/post/could_an…
v/@OACerebro
MIT CSAIL
@MIT_CSAIL
Read here, for those who had 403 errors: online.fliphtml5.com/lbytz/nbjj/
-
MIT CSAIL
@MIT_CSAIL
Beginner-friendly LLM course exploring foundations, architectures, training, deployment, and current trends: github.com/mlabonne/llm-course?tab=readme-ov-file
v/@Hesamation -
MIT CSAIL
@MIT_CSAIL
Time capsule: What a late-'90s high school computer lab looked like, featuring students "surfing the Web."
v/@timecaptalesThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
Today would have been the 85th birthday of Dennis Ritchie (left), inventor of C and co-creator of Unix.
(Image v/Ritchie family archives) -
MIT CSAIL
@MIT_CSAIL
Now that it’s been out for a few days, what are your thoughts on GPT-6 Astra?
-
MIT CSAIL
@MIT_CSAIL
A free beginner’s guide to machine learning: mlstory.org/pdf/patterns.pdf
v/swapnakpanda on X
MIT CSAIL
@MIT_CSAIL
h/t @swapnakpanda
-
MIT CSAIL
@MIT_CSAIL
Shapes can look quite different, depending on your point of view.
v/@TheMathFlow -
MIT CSAIL
@MIT_CSAIL
Free MIT course on distributed systems: www.youtube.com/watch?v=cQP8WApzIQQ&list=PLrw6a1wE39_tb2f…
Here, MIT prof. Robert Morris breaks down what they are (Lecture 1).
v/@MITOCWThis media couldn't be displayed. See it on the original post. -
MIT CSAIL reposted
Manolis Kellis
@manoliskellis
Excited to share our Cross-Species Single-cell Atlas of the Human Brain's Striatum, the brain region that integrates action planning, reward, motivation, decision-making, movement, learning, and reinforcement, implicated in schizophrenia, depression, addiction, and Huntington's Disease: authors.elsevier.com/c/1niW9L7PXy43t
Our map spans 109 samples, 69 donors, and 670,000 nuclei. We identify 31 neuronal subpopulations, including 9 types of medium spiny neurons, and rare D1 interface MSNs forming island-like structures. Patch-clamp recordings show outlier MSNs have smaller somata and stronger hyperpolarization-activated currents than their neighbors.
We show that the dorsolateral-ventromedial gradient shapes: receptor repertoires, disease gene expression, vulnerability.
Across 124 GWAS, we show genetic risk for neuropsychiatric, cognitive and addiction traits concentrates in MSNs and is depleted in glia and vascular cells. Schizophrenia, bipolar and ADHD risk genes are dorsally enriched, and so, unexpectedly, are opioid use disorder risk genes — even though opioid signaling itself is ventrally biased, suggesting heritable risk and acute drug action may engage different parts of the striatum.
Only a quarter of the human gradient genes are similarly zonated in mice. Several drug targets sit in different cells across species, including the mu-opioid receptor, the M5 muscarinic receptor, and the GLP-1 receptor, with important implications for preclinical screening. Profiling chronic antipsychotic exposure in mice, we find the transcriptional response strongly biased toward ventral MSNs.
Finally, we connect genotype to phenotype at single-cell resolution. Our new method, scMOSAIC, reads HTT CAG repeat length and gene expression from the same nucleus — and finds that repeat length varies along that same dorsolateral-ventromedial axis.
Huge congrats to Raleigh Linville @LinvilleRaleigh , Benjamin James @__ben_james__ , who carried this from start to finish, and to co-senior authors #myriamheiman and #danagabuzda, and Kiki Galani, Li-Lun Ho, Stu Fass, James Cameron, Brent Fitzwalter, Erica Engelberg-Cook, Sebastian Pineda, Dennis Dickson, Deborah Mash, Gustavo Turecki, Vanessa Wheeler, and Veronica Alvarez, across Massachusetts Institute of Technology, The Picower Institute for Learning and Memory, MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), Broad Institute of MIT and Harvard, Dana-Farber Cancer Institute, The National Institutes of Health, Mass General Brigham, Mayo Clinic, McLean Hospital, McGill University, and beyond.
Huge thanks to the #scorch consortium and the brain banks in the US and Canada, and a huge thanks to the donors and their families. Every cell in this atlas came from someone who decided their brain should go on being useful. The data are public, and we hope other labs take them further than we can. -
MIT CSAIL
@MIT_CSAIL
Over 70 years ago today, the term “artificial intelligence” was coined in a conference proposal: stanford.io/2WJJJGN
-
MIT CSAIL
@MIT_CSAIL
Basic Linux commands.
-
MIT CSAIL
@MIT_CSAIL
A free repository rich w/AI engineering resources: github.com/chiphuyen/aie-book
v/Chip Huyen -
MIT CSAIL
@MIT_CSAIL
A breakdown of the simplest neural network from the late MIT Prof. Patrick Winston during his 6.034 (Artificial Intelligence) course.
Clip v/@7etsuo & @MITOCW -
MIT CSAIL
@MIT_CSAIL
Just a few bookshelf essentials.
-
MIT CSAIL
@MIT_CSAIL
MIT CSAIL researchers recently exploited a split-second gap in chip security.
The newly discovered flaw can be exploited to reveal sensitive data in Intel & AMD hardware: www.youtube.com/watch?v=-e_orE18TpA -
MIT CSAIL
@MIT_CSAIL
Daniela Rus, CSAIL Director & MIT Professor, has been named one of @TIME's 100 most influential people in AI.
Rus describes her work at CSAIL & Liquid AI as "all about the science & engineering of intelligence, & especially embodied intelligence": time.com/collection/time100-ai/2026/daniela-rus/
TIME
@TIME
TIME’s new cover: Announcing the 2026 TIME100 AI, the world's most influential people in artificial intelligence time.com/collection/time100-ai/2026/
-
MIT CSAIL
@MIT_CSAIL
How people around the world spend their time.
v/Our World in Data -
MIT CSAIL
@MIT_CSAIL
Training finance agents is hard: Private data is locked inside large corporations.
MIT CSAIL researchers release "FinanceGym," the largest finance-agent dataset yet, w/59K+ samples across 37 roles and 40 apps/sites.
Even Kimi K3 & Qwen 3.8 Max failed its test set.
MIT CSAIL
@MIT_CSAIL
FinanceGym data can be used for Reinforcement Finetuning or Supervised Finetuning agentic models.
It contains human-demonstrated trajectories w/thinking traces, & teaches agents financial work, not just financial knowledge.
MIT CSAIL
@MIT_CSAIL
See more info on our Hugging Face page: huggingface.co/datasets/embodied-analysis/FinanceGym
-
MIT CSAIL
@MIT_CSAIL
Grace Hopper's predictions for computing, made back in 1982.
Full talk: www.youtube.com/watch?v=si9iqF5uTFk -
MIT CSAIL
@MIT_CSAIL
Jeff Bezos shared the first Amazon job posting #otd in 1994. 32 years later, it's valued at about $2.85T (Yes, trillion!).
v/Jon Erlichman -
MIT CSAIL
@MIT_CSAIL
"The function of good software is to make the complex appear to be simple." — Grady Booch
Image v/TED -
MIT CSAIL
@MIT_CSAIL
New MIT paper finds that you can can delete an artist from an AI model's training data & nothing changes.
CSAIL engineers also observed that it's hard to trace AI-generated images to specific data. Their work raises new questions about AI copyright: www.csail.mit.edu/news/when-ai-art-has-no-author-mit-stud… -
MIT CSAIL
@MIT_CSAIL
Hi! It's @TamarRottShaham here 👋
Taking over the CSAIL account for the day as I’m wrapping up my postdoc! I’ll be sharing a bit about my research and life at CSAIL. Feel free to ask me anything about AI interpretability research or postdoc life! -
MIT CSAIL
@MIT_CSAIL
Happy 90th birthday to Margaret Hamilton, an MIT computer scientist who helped design the flight software for the Apollo missions in the 1960s.
Image v/MIT Museum -
MIT CSAIL
@MIT_CSAIL
An overview of the fundamentals of neural networks.
v/@tut_ml -
MIT CSAIL
@MIT_CSAIL
All major neural networks, in one chart.
v/The Asimov Institute -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
AI excels w/text & images, but not physics.
The "GeoPT" tool teaches models the basics of real-world mechanics so they can simulate how objects respond to things like water & wind. This efficiency & accuracy boost may help build a physics foundation model: news.mit.edu/2026/ai-models-simulate-wider-range-of-real-… -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
MIT's "ShiftLens" design & fabrication system creates objects that change their surface appearance based on user interactions, w/o any electronics.
It can produce interactive items like a chemical bottle that "transforms" when pressed, slid, or turned: news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-y… -
MIT CSAIL
@MIT_CSAIL
Different types of agent loops, v/@akshay_pachaar.
-
MIT CSAIL
@MIT_CSAIL
MIT's "ShiftLens" design & fabrication system creates objects that change their surface appearance based on user interactions, w/o any electronics.
It can produce interactive items like a chemical bottle that "transforms" when pressed, slid, or turned: news.mit.edu/2026/shiftlens-3d-printed-objects-can-tell-y… -
MIT CSAIL
@MIT_CSAIL
IBM introduced its first PC #otd in 1981. With a 4.77 MHz Intel 8088 microprocessor, the $1600 "Model 5150" soon became the best-selling computer in the world: www.slashgear.com/2072991/80s-most-influential-pieces-of-…
-
MIT CSAIL
@MIT_CSAIL
Computer science shapes just about every facet of our lives now.
Subscribe to MIT CSAIL's Substack to get fresh perspectives on the latest in the field, including AI, robotics, & cybersecurity. You'll get research news, expert opinions, & more for free: csailmit.substack.com -
MIT CSAIL
@MIT_CSAIL
AI excels w/text & images, but not physics.
The "GeoPT" tool teaches models the basics of real-world mechanics so they can simulate how objects respond to things like water & wind. This efficiency & accuracy boost may help build a physics foundation model: news.mit.edu/2026/ai-models-simulate-wider-range-of-real-… -
MIT CSAIL
@MIT_CSAIL
A free guide to the key things every programmers should know: github.com/Babunashvili/Books-To-Read-Before-You-Die/blob…
v/Babunashvili on GitHub -
MIT CSAIL
@MIT_CSAIL
#otd in 1991: Using AppleLink running on a Macintosh Portable, NASA astronauts sent the first email from space.
-
MIT CSAIL
@MIT_CSAIL
A free MIT guide to the mathematical foundations of computer science: courses.csail.mit.edu/6.042/spring18/mcs.pdf
v/@swapnakpanda -
MIT CSAIL
@MIT_CSAIL
#otd in 1991 Tim Berners-Lee first posted on an online newsgroup about a project he was working on called "The World Wide Web."
w3.org/People/Berners-Lee/1991/08/art-6487.txt -
MIT CSAIL
@MIT_CSAIL
MIT researchers found a flaw in modern chip defenses.
They exploited a split-second gap after a processor clears its prediction hardware w/the Interrupt Injection technique. It can bypass defenses & potentially expose sensitive data on Intel & AMD chips: www.csail.mit.edu/news/new-attack-slips-past-latest-defen… -
MIT CSAIL
@MIT_CSAIL
An inside look at the hidden supply chain of AI models.
Full video: www.youtube.com/watch?v=BtvGQF1rPFA -
MIT CSAIL
@MIT_CSAIL
Programmers: Do as I meant, not as I said.
-
MIT CSAIL
@MIT_CSAIL
What's the most valuable thing your AI agent (or agents) do for you?
-
MIT CSAIL
@MIT_CSAIL
This diagram represents what's widely considered to be the world's first "non-trivial" computer program, developed in 1843 by Ada Lovelace for Charles Babbage's Analytical Engine.
(v/@fermatslibrary) -
MIT CSAIL
@MIT_CSAIL
Popular statistical distributions, v/datascience4u on Instagram.
-
MIT CSAIL
@MIT_CSAIL
A guide to key Python functions & concepts, v/@PythonDvz.
-
MIT CSAIL
@MIT_CSAIL
This repository on ML operations has free talks, books, papers and more: shorturl.at/ud0wM
Developed by @visenger -
MIT CSAIL
@MIT_CSAIL
A visualization of a neural network learning handwritten numbers.
v/Hugo Olsson -
MIT CSAIL
@MIT_CSAIL
Daniela Rus, Director of CSAIL & MIT Professor, has received the 2026 High-Tech Prize of the Bavarian Minister-President for her contributions to robotics, AI, & autonomous systems.
Awarded jointly by the Bavarian State Government & the Bavarian Academy of Sciences and Humanities, it's the most highly endowed award for technology & engineering in Germany: www.csail.mit.edu/news/daniela-rus-receives-bavarian-mini… -
MIT CSAIL
@MIT_CSAIL
MIT professor Peter Shor explains two concepts w/i quantum computing: qubits & entanglement.
Video v/MIT Institute Events -
MIT CSAIL
@MIT_CSAIL
"We may have knowledge of the past but cannot control it; we may control the future but have no knowledge of it." — Claude Shannon
Image v/MIT Museum -
MIT CSAIL
@MIT_CSAIL
MIT researchers developed a faster way to sample from AI diffusion models.
They propose a new algorithm reaching the same level of accuracy in exponentially fewer steps, suggesting a path to much more efficient image generation & other diffusion-based AI systems. It recently won Outstanding Paper at ICML: arxiv.org/abs/2602.01338
MIT CSAIL
@MIT_CSAIL
Authors: @FanChen0111, Sinho Chewi, @KonstDaskalakis, & @rakhlin
Additional credit: @MITLIDS -
MIT CSAIL
@MIT_CSAIL
A free guide to competitive programming: duoblogger.github.io/assets/pdf/memonvyftw/guide-t-cp.pdf
-
MIT CSAIL
@MIT_CSAIL
Google CEO Sundar Pichai discusses how AI agents might impact the Internet in the future.
"It will allow humans to use the Internet in ways that give them joy & purpose, & not always be forced to fill out [a bunch of form fields to renew a DMV license, for example]."
Video v/@googledevs & @MatthewBermanThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
#otd in 1969, Apollo 11 launched on its historic journey to the Moon.
The mission was guided by software developed by an MIT team, led by 32-year-old Margaret Hamilton. -
MIT CSAIL
@MIT_CSAIL
"NeuralActuator," developed by researchers at MIT's CDFG Lab (cdfg.mit.edu) w/collaborators from Amazon Robotics, helps low-cost robot arms better model complex actuator behavior in the real world.
The AI model enables sensor-less force perception & force-aware real-robot control, helping close the gap between simulation & reality. It won the Outstanding Systems Paper Award at #rss2026: people.csail.mit.edu/frankzydou/projects/NeuralActuator/This media couldn't be displayed. See it on the original post. -
MIT CSAIL reposted
Zhiyang (Frank) Dou
@frankzydou
Very honored that our work, NeuralActuator, has received the #rss2026 Outstanding Systems Paper Award! 🏆
A huge thank you to all of my collaborators, and many thanks to the #rss community for this recognition!
📄 Paper: arxiv.org/abs/2607.11734
💻 GitHub: github.com/Frank-ZY-Dou/Dynamics-Modeling/tree/main/Neura…
🏠 Project Page: people.csail.mit.edu/frankzydou/projects/NeuralActuator/i…
We have fully open-sourced the code, pretrained checkpoints, datasets, data collection pipeline, and hardware setup.
Actually, many additional experiments and technical discussions (during our development)—including residual torque modeling and different ways of implicitly or explicitly coupling external-force estimation with differentiable simulation—are not included in the paper but are available in the GitHub repository for those interested.
We are also continuously collecting actuator datasets across different platforms. If you’re interested in contributing data or collaborating, we’d love to work together! -
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Robots lack rich & diverse simulation content, which could teach them tasks across various environments.
"SceneSmith" uses three AI agents to create more realistic 3D scenes of places like kitchens & hotels, saving roboticists time on real-world testing: news.mit.edu/2026/ai-agents-create-virtual-playgrounds-to… -
MIT CSAIL
@MIT_CSAIL
32(ish) questions w/an MIT robotics researcher (& actor): www.youtube.com/watch?v=PTIICPx42Rk
-
MIT CSAIL reposted
Zhiyang (Frank) Dou
@frankzydou
Excited to share that our work NeuralActuator: Neural Actuation Modeling for Robot Dynamics and External Force Perception has been accepted to #rss2026!
Your robot — even a low-cost one — can feel external forces without torque or tactile sensors.
TL;DR: NeuralActuator is a neural actuator model that jointly predicts 1️⃣torque to capture the nonlinear and time-varying current–to–torque relationship of low-cost servos, 2️⃣external contact forces (and force detection gates) for sensorless force perception, 3️⃣and motor conditions that indicate each motor’s operating regime.
Here is a fast-forward video clip ⬇️ We are also covering more robots like LeRobot-S101 and Franka Panda.
More details coming soon. -
MIT CSAIL
@MIT_CSAIL
Robots lack rich & diverse simulation content, which could teach them tasks across various environments.
"SceneSmith" uses three AI agents to create more realistic 3D scenes of places like kitchens & hotels, saving roboticists time on real-world testing: news.mit.edu/2026/ai-agents-create-virtual-playgrounds-to… -
MIT CSAIL
@MIT_CSAIL
"The first step of any project is to grossly underestimate its complexity and difficulty." — Nicoll Hunt
-
MIT CSAIL
@MIT_CSAIL
A free guide that answers key reinforcement learning questions: shorturl.at/zcciv
v/@arjunkocher & @sheriyuo -
MIT CSAIL
@MIT_CSAIL
Happy birthday, Nikola Tesla! He helped pioneer the alternating current (AC) power systems that became the foundation of modern electric power distribution.
Image v/Culver Pictures -
MIT CSAIL
@MIT_CSAIL
MIT CSAIL is at ICML this week. One common thread: Teaching ML systems in realistic virtual environments🧵
A couple of examples ⬇️Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
"GeoPT" 🛬
GeoPT teaches simulation models physics by reenacting mechanical interactions in 3D, reaching peak performance 2x faster w/up to 60% less data. It could help predict how vehicles, objects, & robots handle wind, water, & collisions: arxiv.org/abs/2602.20399
MIT CSAIL
@MIT_CSAIL
If you're attending ICML this week & are interested in learning more about either project, send us a DM.
You can find more info about CSAIL research on our website: www.csail.mit.edu
MIT CSAIL
@MIT_CSAIL
In both projects, researchers are helping ML systems learn from more realistic 3D environments to better simulate real-world scenarios.
The core idea is to capture the physical world virtually, making 3D testing more efficient & reducing the need for real-life trial & error. -
MIT CSAIL
@MIT_CSAIL
MIT’s self-organizing boats can create flexible infrastructure on demand.
The team's "FloatForm" system uses a swarm of small aquatic robots, turning water into a programmable surface. These temporary platforms could help w/urban planning & emergencies: www.csail.mit.edu/news/tiny-robot-boats-build-floating-st… -
MIT CSAIL
@MIT_CSAIL
Some more recommended reads, each w/MIT CSAIL authors.
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
How can robots learn to do tasks when instructions are ambiguous?
MIT’s "Masked IRL" uses LLMs to help robots understand vague chore requests & focus on the key details. One model clarifies users’ prompts, then another ignores irrelevant info: news.mit.edu/2026/llms-help-robots-understand-vague-instr… -
MIT CSAIL
@MIT_CSAIL
Free tutorials to help you learn programming tools, v/@ComputerrBooks.
-
MIT CSAIL
@MIT_CSAIL
What's your favorite generative AI tool?
-
MIT CSAIL
@MIT_CSAIL
What came before LaTex, v/@Anthony_Bonato.
-
MIT CSAIL
@MIT_CSAIL
How can robots learn to do tasks when instructions are ambiguous?
MIT’s "Masked IRL" uses LLMs to help robots understand vague chore requests & focus on the key details. One model clarifies users’ prompts, then another ignores irrelevant info: news.mit.edu/2026/llms-help-robots-understand-vague-instr… -
MIT CSAIL reposted
Yaron Meirovitch
@YaronMeirovitch
Massive synaptic pruning between birth and adulthood: the nervous system gets simpler with age! Two-decade connectomic effort in Lichtman lab @harvard reveals how the order in which motor neurons activate decides which connections survive. @NatureNeuro 👉 www.nature.com/articles/s41593-026-02344-7
-
MIT CSAIL
@MIT_CSAIL
Ask us your questions about human-robot collaboration. We’ll pick some to feature in an upcoming 32 questions video w/MIT PhD student Alex Cuellar.
-
MIT CSAIL
@MIT_CSAIL
MIT CSAIL was founded #otd in 1963 as "Project MAC."
-
MIT CSAIL
@MIT_CSAIL
#otd in 2007, the iPhone 1 released: www.businessinsider.com/apple-iphone-evolution-first-ipho…
Images: Getty, LA Times, & Wikimedia Commons. -
MIT CSAIL
@MIT_CSAIL
"You might not think that programmers are artists, but programming is an extremely creative profession. It's logic-based creativity." — John Romero
-
MIT CSAIL
@MIT_CSAIL
Mastering Claude in 18 steps, v/@AnatoliKopadze.
-
MIT CSAIL reposted
MIT CSAIL
@MIT_CSAIL
Do-it-all robots aren't ready for our homes.
Scientists at MIT & beyond think compliance might change that: www.youtube.com/watch?v=auAmgSSNYqM -
MIT CSAIL
@MIT_CSAIL
Even when a robot is doing predictable, everyday tasks, it can struggle to adjust to disturbances like collisions.
MIT’s "CALM" helps them stay on task by tracking the motions of a few human demos & averaging them into a path that’s easy to stick to: www.csail.mit.edu/news/motion-tracking-system-shows-robot… -
MIT CSAIL
@MIT_CSAIL
Do-it-all robots aren't ready for our homes.
Scientists at MIT & beyond think compliance might change that: www.youtube.com/watch?v=auAmgSSNYqM -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Alan Turing, the creator of modern computing.
-
MIT CSAIL
@MIT_CSAIL
A free MIT guide to key computer vision concepts: visionbook.mit.edu
-
MIT CSAIL reposted
Zhaofeng Wu
@zhaofeng_wu
Introducing ><former
Most transformers are rectangles◻️: every layer has the same width
But is that optimal?🤔
We propose variable-width transformers that have different widths across layers, improving loss while cutting compute & KV cache size 🧵 -
MIT CSAIL
@MIT_CSAIL
In German, the word "verschlimmbessern" means "to make something worse while trying to repairing it." In English, we call that "software development."
-
MIT CSAIL
@MIT_CSAIL
100 helpful pointers for mastering Claude, v/@rubenhassid.
-
MIT CSAIL
@MIT_CSAIL
What was your first programming language? What do you remember most about learning it?
-
MIT CSAIL reposted
Adam Wei @CoRL ✈️
@adamwei_
🤖 We introduce Ambient Diffusion Policy, a simple and principled method for training policies with suboptimal data in robotics.
Suboptimal data is everywhere in robotics…
❌ Data filtering is wasteful
❌ Co-training learns both good and bad features
✅ Ambient Diffusion Policy selectively learns useful features via noise-dependent data usage
👇🧵(1/5)This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
A free MIT course about deep learning, 2026 edition: introtodeeplearning.com
Here, MIT CSAIL researcher Alexander Amini discusses the differences between supervised, unsupervised, & reinforcement learning approaches (Lecture 5). -
MIT CSAIL
@MIT_CSAIL
QS has ranked MIT the world's No. 1 university for the 15th year in a row: news.mit.edu/2026/qs-ranks-mit-worlds-no-1-university-0617
The Institute has placed first in subject areas such as:
- computer science & information systems
- data science & AI
- & electrical & electronic engineering. -
MIT CSAIL reposted
alex zhang
@a1zhang
honored to be named one of @sciam's Young American Scientists!
thanks to my advisor @lateinteraction @MIT_CSAIL, and the wonderful people at @PrimeIntellect and @LaudeInstitute for supporting my research. There's a lot of work to do for open science :) -
MIT CSAIL
@MIT_CSAIL
The hidden patterns & math behind Pascal’s Triangle, v/@mathemetica.
-
MIT CSAIL
@MIT_CSAIL
Notes Richard Feynman '39 took for his high school calculus class (early 1930s), v/Physics Today & Melinda Baldwin.
-
MIT CSAIL
@MIT_CSAIL
"A computer will do what you tell it to do, but that may be much different from what you had in mind." — Joseph Weizenbaum
-
MIT CSAIL
@MIT_CSAIL
A free MIT course breaking down fundamental math concepts in computer science: www.youtube.com/watch?v=sbpCTjmw85g&list=PLUl4u3cNGP61VNv…
Here, MIT prof. Erik Demaine breaks down state machines (Lecture 4).
v/@MITOCW -
MIT CSAIL
@MIT_CSAIL
Many AI tools impacting patient care may be operating outside of FDA oversight.
To change that, MIT researchers recommend increased public disclosure, structured dialogue between industry & regulators, & updated FDA guidance: www.csail.mit.edu/news/ai-tools-shaping-patient-care-may-… -
MIT CSAIL
@MIT_CSAIL
Daniela Rus, CSAIL Director & MIT Prof., & Regina Barzilay, MIT Prof. & CSAIL principal investigator, have been named 2026 Tech Power Players by @BostonGlobe.
Full list: globe.com/tech50 #globetech50 -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Web inventor & MIT Professor Emeritus Tim Berners-Lee! Things besides the WWW that didn’t exist before TBL:
- Apple
- Google
- Facebook
- YouTube
- Instagram
- Netflix
- Uber
- Spotify
- & many other technologies we use today
Image v/Sotheby's -
MIT CSAIL
@MIT_CSAIL
In case of fire, save your code first.
Image v/@igor_os777 -
MIT CSAIL
@MIT_CSAIL
A free 30-minute guide to mastering Claude Code: www.youtube.com/watch?v=6eBSHbLKuN0
-
MIT CSAIL reposted
Gabe Grand
@gabe_grand
Excited to share a bunch of recent press in @MIT News and Scientific American!
From today: news.mit.edu/2026/teaching-ai-agents-ask-better-questions… -
MIT CSAIL
@MIT_CSAIL
"Collaborative Battleship" game has revealed that AI agents are better at answering questions than asking them.
CSAIL & SEAS had LMs play together, where Monte Carlo inference strategies helped small agents outpace the largest models at ~1% of the cost: news.mit.edu/2026/teaching-ai-agents-ask-better-questions… -
MIT CSAIL
@MIT_CSAIL
9 distance measures in data science w/algorithms (v/@MaartenGr).
-
MIT CSAIL
@MIT_CSAIL
"Anyone who has lost track of time when using a computer knows the propensity to dream, the urge to make dreams come true, & the tendency to miss lunch." — Tim Berners-Lee
-
MIT CSAIL
@MIT_CSAIL
Classic tech: booting up an old school (mid-90s) computer.
-
MIT CSAIL
@MIT_CSAIL
Free MIT course breaks down deep learning: ocw.mit.edu/courses/6-7960-deep-learning-fall-2024/
Here, MIT ass't prof. Sara Beery discusses how it's driven breakthroughs in areas like image generation, coding, & playing games (Lecture 1). -
MIT CSAIL
@MIT_CSAIL
What really happens inside a computer chip?
MIT CSAIL built an operating system kernel called "Fractal" that gives researchers a clearer view. It revealed that the Apple M1 could be vulnerable to a class of speculative attack known as "Phantom": www.csail.mit.edu/news/study-how-chips-really-work-mit-re… -
MIT CSAIL
@MIT_CSAIL
What’s an invention you feel deserves more credit?
-
MIT CSAIL
@MIT_CSAIL
Bathroom decor inspired by GitHub 😂
v/@r0ktech -
MIT CSAIL
@MIT_CSAIL
Classic computer animation techniques from the 90s, as used in Toy Story.
v/@0xmitsurii -
MIT CSAIL
@MIT_CSAIL
In May 1967, two computer scientists introduced Simula 67, an early programming language that pioneered concepts like classes & objects, laying the groundwork for object-oriented programming: twobithistory.org/2019/01/31/simula.html
-
MIT CSAIL reposted
Joseph Ravichandran
@0xjprx
I built an operating system just for reverse engineering CPUs.
We used it to poke around the branch predictors on Apple Silicon and found some cool stuff (including Phantom fetches!).
See what we found here: www.csail.mit.edu/news/study-how-chips-really-work-mit-re… -
MIT CSAIL
@MIT_CSAIL
MIT researchers developed “Insum,” a technique for speeding up computations on datasets replete w/zeros.
It rewrites Einstein summation (“einsum”) operations to avoid inefficient handling of zeros, improving memory efficiency & performance: www.csail.mit.edu/news/new-way-speed-sparse-computations -
MIT CSAIL
@MIT_CSAIL
Different types of matrices, v/@codek_tv.
-
MIT CSAIL
@MIT_CSAIL
Ranking every birthday, from most to least common. Where does yours rank?
-
MIT CSAIL
@MIT_CSAIL
Einstein's notes from 1913 about relativity.
Images v/@LensScientific -
MIT CSAIL
@MIT_CSAIL
"I was born not knowing & have had only a little time to change that here & there." — Richard P. Feynman '39
-
MIT CSAIL
@MIT_CSAIL
New @washingtonpost story details the secret instructions AI companies give to chatbots to steer them towards certain behaviors: www.washingtonpost.com/technology/interactive/2026/chatbo…
One example from OpenAI: "Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely & unambiguously relevant to the user’s query." -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Edsger Dijkstra, whose algorithm solved the "single-source shortest path" problem: www.youtube.com/watch?v=gdmfOwyQlcI
-
MIT CSAIL
@MIT_CSAIL
What's one thing AI agents can't do yet, but you wish they could?
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Caltech professor Katie Bouman! As a PhD student at MIT & CSAIL researcher, she was part of the team that helped create the first-ever image of a black hole.
-
MIT CSAIL
@MIT_CSAIL
Simple harmonic motion, as visualized by @fermatslibrary.
-
MIT CSAIL
@MIT_CSAIL
The story of Mirai, MIT's AI model that can detect breast cancer years before humans do: www.youtube.com/watch?v=FcYt5ETYV6E
-
MIT CSAIL
@MIT_CSAIL
Easily our favorite Met Gala outfit 👏👏👏
Catherine Yeo
@catherinehyeo
Love seeing Naomi Osaka honor the CLRS Algorithms textbook at this year's Met Gala
-
MIT CSAIL
@MIT_CSAIL
In the '80s, future MIT prof. Bill Freeman envisioned a three-sided zipper that could switch items between soft & rigid states.
Last year, CSAIL revived his idea w/an automated design tool. It 3D prints "Y-zippers" that snap gear, robots, & art into place: news.mit.edu/2026/three-sided-y-zipper-design-0504 -
MIT CSAIL
@MIT_CSAIL
The top 10 programming languages on GitHub over the past 2+ years.
Notably, TypeScript has leaped to the top spot.
v/@github -
MIT CSAIL
@MIT_CSAIL
"Any sufficiently advanced technology is indistinguishable from magic." — Arthur C. Clarke
-
MIT CSAIL
@MIT_CSAIL
The Wright Brothers' patent for a "Flying Machine," AKA the world's first airplane. It was filed in 1903.
Images v/Library of Congress -
MIT CSAIL
@MIT_CSAIL
A free guide to prompt engineering from Google: shorturl.at/BaGVe
-
MIT CSAIL reposted
Pulkit Agrawal
@pulkitology
Eka means unity -- “one,” in Sanskrit and “first” in Finnish.
We’re building intelligence for the physical world in its native language: forces.
Until now, robotics faced a tradeoff — generality or speed. The real world requires both. Robotics also faced a data problem.
Our Vision–Force–Action (VFA) model — the first of its kind — breaks the generality-speed tradeoff and the data barrier.
It's a new foundation uniting performance, generality, and safety for putting capable robots in everyone's hands.
Today, I am excited to share our journey of pushing robots beyond human limits.
Today, dexterity becomes scalable.
Today, I welcome you to the Era of Eka.
Co-founded with @haarnoja, and so thrilled and grateful to be working with a dream team at @EkaRobotics.
Learn more: www.ekarobotics.com -
MIT CSAIL
@MIT_CSAIL
MIT PhD student Alex Zhang (@a1zhang) explains how AI models are "mismanaged geniuses" that could take on a much wider range of tasks.
Full video: www.youtube.com/watch?v=umbtFGPzdAM -
MIT CSAIL
@MIT_CSAIL
Highest paid workers are more likely to use AI for work, according to Financial Times breakdown: shorturl.at/uSC71
-
MIT CSAIL
@MIT_CSAIL
61 years ago this month, the Fast Fourier Transform was created, a powerful tool for image compression & data analysis.
Watch a classic MIT breakdown of FFT, perhaps the most-taught algorithm at the Institute: www.youtube.com/watch?v=iTMn0Kt18tg
v/@MITOCW -
MIT CSAIL
@MIT_CSAIL
Happy bday to the late Larry Tesler, inventor of "copy-paste."
Happy bday to the late Larry Tesler, inventor of "copy-paste."
Happy bday to the late Larry Tesler, inventor of "copy-paste."
cnet.co/4b5ZGrq (Pic v/Reddit) -
MIT CSAIL
@MIT_CSAIL
How do top reasoning models become overconfident?
MIT found that RL rewards correct answers w/o considering how sure the model is. By training them to estimate their confidence about each answer, the team boosted uncertainty estimates w/o hurting accuracy: -
MIT CSAIL
@MIT_CSAIL
Ask us your questions about Recursive Language Models (RLMs) & where LMs are underutilized or inefficient. We’ll pick some to feature in an upcoming explainer w/MIT PhD student Alex Zhang (@a1zhang), who recently developed RLMs.
For more on his work: alexzhang13.github.io -
MIT CSAIL
@MIT_CSAIL
This week, MIT CSAIL will join other top ML researchers at ICLR to tackle a shift in focus from more powerful AI to more reliable systems 🧵
Our papers at the conference show how to potentially make AI models stronger critical thinkers, more honest, & better at math ⬇️Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
“MathNet” 📊
MIT, KAUST & HUMAIN have built the world's largest collection of Olympiad-level math problems.
The dataset can help students prepare for competitions & revealed how AI models can improve at math problems, esp. ones w/visual reasoning: www.csail.mit.edu/news/mit-researchers-build-worlds-large…
MIT CSAIL
@MIT_CSAIL
This sampling of MIT CSAIL papers at ICLR shows a common need for efficient models that can reason about complex, real-world problems. More compute helps, but the ways these machines "think" also need refinement.
You can find more info about these projects on our website: -
MIT CSAIL
@MIT_CSAIL
Today, MIT & the IMO released MathNet, the world’s largest dataset of International Math Olympiad problems & solutions 🌍
MathNet is 5x larger than previous datasets & is sourced from over 40 countries across 4 decades: www.csail.mit.edu/news/mit-researchers-build-worlds-large… -
MIT CSAIL
@MIT_CSAIL
100 tips to get the most out of Claude, v/@BoucherNicolas.
-
MIT CSAIL
@MIT_CSAIL
Things we didn't have 25 years ago:
iPhone
Facebook
YouTube
Twitter/X
Instagram
Android
Bitcoin
Tesla
Gmail
WhatsApp
Snapchat
Zoom
Amazon Prime
Airbnb
Uber
Dropbox
LinkedIn
Reddit
ChatGPT
v/@stats_feed -
MIT CSAIL
@MIT_CSAIL
"Code never lies, comments sometimes do." — Ron Jeffries
-
MIT CSAIL
@MIT_CSAIL
A decade of imaging.
Compressed into three months.
Here’s how MIT CSAIL & Harvard taught an electron microscope to see like you do: www.youtube.com/watch?v=u1rKQddhxbc
MIT CSAIL
@MIT_CSAIL
The team's work is now in Nature Methods, by @YaronMeirovitch et al: www.nature.com/articles/s41592-025-02929-3
-
MIT CSAIL
@MIT_CSAIL
Read the lost thesis of Dennis Ritchie, creator of the C programming language & co-creator of Unix: archive.computerhistory.org/resources/access/text/2020/05…
Ritchie never got his PhD b/c he didn't want to pay Harvard the thesis binding fee.
(v/@IEEESpectrum) -
MIT CSAIL
@MIT_CSAIL
X back in 2006, v/@JamesLucasIT.
This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
If you could have dinner w/any computer scientist, past or present, who would it be?
-
MIT CSAIL
@MIT_CSAIL
#otd in 1980 a Japanese computer scientist published a paper proposing the “Neocognitron,” the neural net that directly inspired CNNs: www.rctn.org/bruno/public/papers/Fukushima1980.pdf
Kunihiko Fukushima’s paper explained back in 1986: www.youtube.com/watch?v=Qil4kmvm2Sw -
MIT CSAIL
@MIT_CSAIL
An interactive exhibit showing how a computer processor does math, v/@ProofofMaro.
-
MIT CSAIL
@MIT_CSAIL
Training a large AI model is expensive, but MIT CSAIL researchers have helped develop a new approach that cuts compute costs.
Using control theory, "CompreSSM" compresses models during training, shedding complexity. It makes models leaner & faster: news.mit.edu/2026/new-technique-makes-ai-models-leaner-fa… -
MIT CSAIL
@MIT_CSAIL
In honor of Git's 21st birthday, a poem written in Git: shorturl.at/fp80F
-
MIT CSAIL reposted
Zengyi Qin
@qinzytech
How do you train AI agents that can use computers like humans?
Introducing OSGym: Scalable OS Infrastructure for Computer Use Agents. It’s large-scale training made possible by extensive infra optimization:
• 1024 OS sandboxes running in parallel
• $0.23 / sandbox / day (90% cost down)
• 1420 trajectories per minute
• 37× faster disk provisioning
• 88% less physical disk usage
• RAM-bound orchestration (10× scalable than CPU-bound)
🧵 -
MIT CSAIL
@MIT_CSAIL
How do you train AI agents that can use computers like humans? 🧵
MIT CSAIL researchers introduce "OSGym": scalable OS infrastructure to improve the capabilities of computer use agents. It introduces large-scale training made possible by extensive infrastructure optimization:Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
So why did the researchers develop OSGym?
Training computer use agents is both a modeling problem & an infrastructure challenge:
• Each task needs a full OS (GUI, apps, browser, etc.)
• Scaling → massive CPU/RAM/disk costs
• Random crashes from real software environments
• Thousands of parallel runs required
It’s a key bottleneck, though it’s not often discussed.
MIT CSAIL
@MIT_CSAIL
The key insight behind OSGym: Don’t scale like traditional systems.
• Decentralize each OS replica (no global bottleneck)
• Shift from CPU-bound → RAM-bound scaling (10x cheaper)
• Use XFS reflink copy-on-write disks (37x faster disk provisioning)
The result: Scalable, fault-tolerant, & actually affordable agent training. -
MIT CSAIL
@MIT_CSAIL
What a difference 70 years makes: 3.75 MB in 1956 & 1 TB today.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
"Life is not easy for any of us, but what of that? We must have perseverance & above all confidence in ourselves. We must believe that we are gifted in something, & that this thing, at whatever cost, must be attained." — Marie Curie
-
MIT CSAIL
@MIT_CSAIL
A first-of-its-kind study from MIT measures how successful AI is in completing thousands of tasks done by workers in the US economy.
Across these real-world tasks, they found that AI capabilities are improving quickly, but performance is rising smoothly: www.csail.mit.edu/news/new-mit-research-overturns-prior-v… -
MIT CSAIL reposted
Ihtesham Ali
@ihteshamali
A MIT professor taught the same lecture every January for 40 years, and every single time it was standing room only.
I watched it at 2am and it completely rewired how I think about communication.
His name was Patrick Winston. The lecture is called "How to Speak."
His opening line hit like a truck: your success in life will be determined largely by your ability to speak, your ability to write, and the quality of your ideas in that order.
Not your GPA. Not your pedigree. Not your IQ. How you speak is what separates people who get heard from people who get ignored.
Here's the framework he drilled into MIT students for four decades.
He said never start with a joke. Start by telling people exactly what they're going to learn. Prime the pump before you pour anything in. He called it the "empowerment promise" give people a reason to stay in their seats within the first 60 seconds.
Then he broke down the 5S rule for making ideas stick: Symbol, Slogan, Surprise, Salient, and Story. Every idea worth remembering hits at least three of these.
The part that floored me was his "near miss" technique. Don't just show what's right show what almost looks right but isn't. That contrast is when the brain actually locks something in permanently.
His final rule before any big talk: end with a contribution, not a summary. Don't recap what you said. Tell people what you gave them that they didn't have before they walked in.
I've used this framework in pitches, interviews, and presentations ever since watching it, and the results are not subtle.
Patrick Winston passed away in 2019, but this lecture is still free on MIT OpenCourseWare. One hour, watched by millions, and it costs absolutely nothing.
The most important class MIT ever put on the internet isn't about code or math. It's about how to make people actually listen to you. -
MIT CSAIL
@MIT_CSAIL
Where can linearized dynamics be trusted w/contact?
MIT's "Contact Trust Region" answers this using manipulation theory, differentiable simulation & control. It enables manipulation w/less compute than RL, powering robots to be more energy-efficient: ctr.theaiinstitute.comThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
An introductory course about deep learning, courtesy of MIT: consent.youtube.com/m?continue=https%3A%2F%2Fwww.youtube.…
-
MIT CSAIL
@MIT_CSAIL
A prompt you can use in Claude to help build & launch apps, v/@milesdeutscher.
-
MIT CSAIL
@MIT_CSAIL
A cheat sheet to data structures, v/@PythonPr.
-
MIT CSAIL
@MIT_CSAIL
21 years ago MIT researchers got a computer-generated gibberish paper accepted to a predatory journal.
Generate your own here: pdos.csail.mit.edu/archive/scigen/ -
MIT CSAIL
@MIT_CSAIL
What if you could "reprogram" the way everyday things like bags & walls look?
MIT's "ChromoLCD" tool could bring that idea to life. It uses LCDs & LEDs to stamp high-res designs onto surfaces, possibly creating customizable fashion pieces & smart homes: news.mit.edu/2026/chromolcd-high-res-imagery-on-everyday-… -
MIT CSAIL
@MIT_CSAIL
The MIT paper that showed space can be more powerful than time in computers.
Full video: www.youtube.com/watch?v=kQo30UaEh8I -
MIT CSAIL
@MIT_CSAIL
The life of a computer programmer in the 1940s.
Images v/Computer History Museum -
MIT CSAIL
@MIT_CSAIL
A take from 1963 that aged pretty well.
-
MIT CSAIL
@MIT_CSAIL
Happy birthday Joseph Fourier, whose 1822 equation allows us to listen to mp3 audio files today: www.quantamagazine.org/what-is-the-fourier-transform-2025…
-
MIT CSAIL
@MIT_CSAIL
"A good programmer is someone who always looks both ways before crossing a one-way street." — Doug Linder
-
MIT CSAIL reposted
Nidhi Seethapathi
@nidhi_s91
How does perceived risk shape adaptation and learning?
Our new work led by @inseungkang reveals that locomotor adaptation proactively navigates a "fall risk landscape" , modulating learning parameters that dictate optimality to prioritize safety.
doi.org/10.64898/2026.03.02.709033 -
MIT CSAIL
@MIT_CSAIL
Is there a "secret sauce" driving the development of AI models?
FutureTech researchers at MIT CSAIL found that at the frontier, sheer scale drives LLM performance, but proprietary techniques & algorithmic advances matter more away from it: www.csail.mit.edu/news/3-questions-there-secret-sauce-ai-… -
MIT CSAIL
@MIT_CSAIL
Remember browsing the Internet on keypad phones?
v/@NatureMomentz -
MIT CSAIL
@MIT_CSAIL
Graphing out where AI complements human work (e.g. chief executives) & where it could pressure them to learn new skills (e.g. secretaries).
More details: www.reddit.com/r/dataisbeautiful/comments/1qkqygn/oc_whic…
Credits: xY2j-Ib2p9--NmEX-43- on Reddit & EBRD -
MIT CSAIL
@MIT_CSAIL
Happy Pi Day! Here's the first 100,000 digits of π, visualized.
v/AvoidableBoat67 on Reddit -
MIT CSAIL
@MIT_CSAIL
A free repository of curated ML system design case studies from 80+ leading companies: github.com/Engineer1999/A-Curated-List-of-ML-System-Desig…
v/@techNmak -
MIT CSAIL
@MIT_CSAIL
A prompt you can use in Claude to help you master prompt engineering, v/@milesdeutscher.
-
MIT CSAIL
@MIT_CSAIL
Reaching scientific goals: expectation vs. reality, v/Florian Aigner.
-
MIT CSAIL
@MIT_CSAIL
A guide to remembering everything you read: readwise-assets.s3.amazonaws.com/media/wisereads/articles…
v/@CuriousMindsHub -
MIT CSAIL
@MIT_CSAIL
"An expert is a person who has made all the mistakes that can be made in a very narrow field." — Niels Bohr
-
MIT CSAIL
@MIT_CSAIL
The software development life cycle.
-
MIT CSAIL
@MIT_CSAIL
MIT in the late 1800s (when it was located in Boston) vs. today (Cambridge).
-
MIT CSAIL
@MIT_CSAIL
Apple CEO Tim Cook advises MIT grads on how to measure their impact on humanity (2017).
Full Commencement Address: www.youtube.com/watch?v=ckjkz8zuMMs -
MIT CSAIL
@MIT_CSAIL
"There is a single light of science, & to brighten it anywhere is to brighten it everywhere." — Isaac Asimov
-
MIT CSAIL
@MIT_CSAIL
Free MIT course that covers tools & technologies that are assumed, but rarely taught, in most CS programs & software engineering jobs, like Bash, git, & AI tools: missing.csail.mit.edu/2026/course-shell/
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Grady Booch, who helped develop the Unified Modeling Language (UML). He has also advanced the fields of software engineering & object-oriented design.
-
MIT CSAIL
@MIT_CSAIL
GenAI models can create 3D designs for personal items that are clever, but don't sustain everyday use.
By augmenting genAI w/physics simulations, MIT’s "PhysiOpt" tool rapidly drafts objects that work as intended when fabricated — no extra training needed: -
MIT CSAIL
@MIT_CSAIL
This is MIT CSAIL: www.youtube.com/watch?v=pT2wbEyrbgg
This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
How you know you're getting older: the "dragon book" was published nearly forty years ago.
shorturl.at/DswlP -
MIT CSAIL
@MIT_CSAIL
Which research paper(s) changed how you view AI?
-
MIT CSAIL
@MIT_CSAIL
This month, “hello world” said “hello world!”
The term was coined in a seminal programming book published in 1978: “C Programming Language,” written by Brian Kernighan and Dennis Ritchie: go.aws/37TmOZq -
MIT CSAIL
@MIT_CSAIL
Often-overlooked abbreviations for tech terms:
QR Code - Quick Response Code 📱
PDF - Portable Document Format 📄
SIM - Subscriber Identity Module 💳
SMS - Short Message Service 💬
WWW - World Wide Web 🌐
GPS - Global Positioning System 🛰️
API - Application Programming Interface 🔗 -
MIT CSAIL
@MIT_CSAIL
The emergence of OpenClaw & MoltBook has heightened public interest & concerns around AI agents. Plus, the US gov't is launching the AI Agent Standards Initiative.
Amidst this, MIT CSAIL & collaborators are introducing the 2025 AI Agent Index, the only project of its kind to offer an ecosystem-wide view of AI agents. It documents the capabilities, safety features, & transparency of 30 top AI agents: aiagentindex.mit.eduPost not available here, view on X
MIT CSAIL
@MIT_CSAIL
Beyond these insights, researchers found that developers in different countries (mainly the US & China) are taking markedly different approaches to agent development.
MIT CSAIL
@MIT_CSAIL
The AI Agent Index collected 1350 fields of information on 30 state-of-the-art systems. Check out the website for more findings & data: aiagentindex.mit.edu
For questions or other inquiries, contact @StephenLCasper (scasper@mit.edu) & @LeonStaufer (lets2@cam.ac.uk). -
MIT CSAIL
@MIT_CSAIL
Happy 35th anniversary to Python!
Python 0.9.0, the first public version of the popular programming language, released #otd in 1991. -
MIT CSAIL
@MIT_CSAIL
A free MIT course about building scalable & high-performance software systems: ocw.mit.edu/courses/6-172-performance-engineering-of-soft… (v/@MITOCW)
Here, MIT prof. & CSAIL principal investigator Charles Leiserson discusses why performance matters in computing (Lecture 1). -
MIT CSAIL
@MIT_CSAIL
Which jobs will AI replace in the next 2 or 3 years?
On a new episode of the Silicon Valley Girl Podcast (@siliconvalleymm), MIT Prof. & CSAIL Director Daniela Rus discusses where AI is already enhancing our jobs & where humans are still needed: www.youtube.com/watch?v=JFilKfL7_9A -
MIT CSAIL
@MIT_CSAIL
When programming languages were released:
1957: Fortran
1958: Lisp
1959: COBOL
1964: BASIC
1970: Pascal
1972: C
1978: SQL
1985: C++
1991: Python
1993: R
1995: PHP, Java, & JavaScript
2001: C#
2009: Go
2012: Julia
2014: Swift
2015: Rust
2016: Kotlin
2017: WebAssembly
2023: Mojo -
MIT CSAIL
@MIT_CSAIL
Fun fact: The word “computer” originally meant a person who computes, not a machine. It was used back as far back as the 17th century.
Roughly 300 years later, the term was adapted to described electronic devices. -
MIT CSAIL reposted
Zhening Li (Zed)
@zli11010
Ever wished you could rewind your AI agent when it goes off track?
EnCompass lets you rewind any Python program containing LLM calls.
While normal program execution runs a program once from start to finish, 1/ -
MIT CSAIL
@MIT_CSAIL
The late Frank Gehry's architectural models of the Stata Center, which were displayed at MIT back in 2001. The building opened three years later.
Images v/MIT professor Peter Szolovits -
MIT CSAIL
@MIT_CSAIL
Candy heart messages written by GPT-2 (left) vs. GPT-5 (right). The latter is a bit more romantic.
Original image v/Janelle Shane -
MIT CSAIL
@MIT_CSAIL
MIT grad gives full roadmap to learning AI in just 17 minutes: www.youtube.com/watch?v=EWFFaKxsz_s
-
MIT CSAIL reposted
Russ Tedrake
@RussTedrake
I've been saying for years that the biggest challenge for simulation in robotics is not actually the physics engine (although you do have to get that right). The real challenge is capturing the *diversity* of the real world. There was no doubt that generative AI had the potential to change that, but it's still amazing to see it take shape.
Watching Nick's incredibly fast progress has convinced me that content generation might not actually be a bottleneck anymore. This is a beautiful combination of hardened tools for e.g. low-level mesh processing with the latest tools for generative asset creation, wrapped in a powerful agentic workflow. Please do give it a try and share your feedback.
Nicholas Pfaff
@NicholasEPfaff
Meet SceneSmith: An agentic system that generates entire simulation-ready environments from a single text prompt.
VLM agents collaborate to build scenes with dozens of objects per room, articulated furniture, and full physics properties.
We believe environment generation is no longer the bottleneck for scalable robot training and evaluation in simulation.
Website: scenesmith.github.io
👇🧵(1/8)This media couldn't be displayed. See it on the original post. -
MIT CSAIL reposted
Nicholas Pfaff
@NicholasEPfaff
Meet SceneSmith: An agentic system that generates entire simulation-ready environments from a single text prompt.
VLM agents collaborate to build scenes with dozens of objects per room, articulated furniture, and full physics properties.
We believe environment generation is no longer the bottleneck for scalable robot training and evaluation in simulation.
Website: scenesmith.github.io
👇🧵(1/8)This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Ray Kurzweil '70, an innovator in speech & text recognition. He's also a leading proponent of “the singularity.”
Photo v/The Academy of Achievement -
MIT CSAIL
@MIT_CSAIL
AI agents can be very effective when they use LLMs, but coding agents to work backwards to fix mistakes is time-consuming.
MIT method executes AI agent programs by backtracking & making multiple attempts, helping coders work w/these systems efficiently: news.mit.edu/2026/helping-ai-agents-search-to-get-best-re… -
MIT CSAIL reposted
Derya Unutmaz, MD
@DeryaTR_
I finally had a chance to read this paper. I am now convinced that Recursive Language Models (RLMs) are going to be the next big thing in AI advances! Attention is shifting toward very large context windows. Very impressive paper! Congrats to Alex who is a new PhD student at MIT.
alex zhang
@a1zhang
Much like the switch in 2025 from language models to reasoning models, we think 2026 will be all about the switch to Recursive Language Models (RLMs).
It turns out that models can be far more powerful if you allow them to treat *their own prompts* as an object in an external environment, which they understand and manipulate by writing code that invokes LLMs!
Our full paper on RLMs is now available—with much more expansive experiments compared to our initial blogpost from October 2025!
arxiv.org/pdf/2512.24601 -
MIT CSAIL
@MIT_CSAIL
Deep Learning job interview questions, fully solved & covering a wide range of key AI topics: arxiv.org/abs/2201.00650
credit: @papers_daily -
MIT CSAIL reposted
alex zhang
@a1zhang
Much like the switch in 2025 from language models to reasoning models, we think 2026 will be all about the switch to Recursive Language Models (RLMs).
It turns out that models can be far more powerful if you allow them to treat *their own prompts* as an object in an external environment, which they understand and manipulate by writing code that invokes LLMs!
Our full paper on RLMs is now available—with much more expansive experiments compared to our initial blogpost from October 2025!
arxiv.org/pdf/2512.24601 -
MIT CSAIL
@MIT_CSAIL
“Hardware eventually fails. Software eventually works.” — Michael Hartung
-
MIT CSAIL
@MIT_CSAIL
Happy 79th birthday to Gerry Sussman, the MIT prof. who co-wrote "the Wizard Book” (Structure and Interpretation of Computer Programs) w/Hal Abelson & Julie Sussman in 1984.
Read it for free here: mitp-content-server.mit.edu/books/content/sectbyfn/books_… -
MIT CSAIL
@MIT_CSAIL
What’s the oldest piece of tech you own?
-
MIT CSAIL
@MIT_CSAIL
A conceptual map of agentic AI, v/@LevelUpCoding_.
-
MIT CSAIL
@MIT_CSAIL
Today would have been the 83rd birthday of Professor Patrick Winston, who passed away in 2019 after directing MIT’s AI lab for 25 years.
Here are his complete video lectures for 6.034, which have been watched over 9 million times: www.youtube.com/watch?v=TjZBTDzGeGg&list=PLUl4u3cNGP63gFH… -
MIT CSAIL
@MIT_CSAIL
"Every great developer you know got there by solving problems they were unqualified to solve until they actually did it." — Patrick McKenzie
-
MIT CSAIL
@MIT_CSAIL
Key skills for aspiring data scientists to learn.
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Python creator Guido van Rossum. The open source language was named after comedy troupe Monty Python: spectrum.ieee.org/top-programming-languages-2025
IEEE Spectrum recently named it their top programming language: spectrum.ieee.org/top-programming-languages-2025 -
MIT CSAIL
@MIT_CSAIL
A free MIT introduction to computer science & Python programming: www.youtube.com/watch?v=xAcTmDO6NTI&list=PLUl4u3cNGP62A-y…
This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
Over 50 essential LLM interview questions & their answers: github.com/Devinterview-io/llms-interview-questions?tab=r…
v/devinterview -
MIT CSAIL reposted
Lior Alexander
@LiorOnAI
Claude can make blue1brown animations in minutes.
Education is about to explode. -
MIT CSAIL
@MIT_CSAIL
This repository on ML operations has free talks, books, papers and more: github.com/visenger/awesome-mlops
Developed by @visenger -
MIT CSAIL
@MIT_CSAIL
Thoughts on Claude Code?
-
MIT CSAIL
@MIT_CSAIL
Helpful study methods to help you ace your exams, v/@cooltechtipz.
-
MIT CSAIL
@MIT_CSAIL
A free beginner’s guide to machine learning: mlstory.org/pdf/patterns.pdf
v/@swapnakpanda -
MIT CSAIL
@MIT_CSAIL
What’s your favorite book about machine learning?
-
MIT CSAIL
@MIT_CSAIL
Things like precious opal appear iridescent thanks to structural color — microscopic structures that reflect light to produce radiant hues.
MIT system pairs software & hardware to design & program these color-changing properties onto everyday items: news.mit.edu/2026/mit-morphochrome-paints-objects-with-li… -
MIT CSAIL
@MIT_CSAIL
"Programming is not about typing, it's about thinking." — Rich Hickey
v/@CodeWisdom -
MIT CSAIL
@MIT_CSAIL
A flow chart for writing a research proposal, v/@acagamic.
-
MIT CSAIL
@MIT_CSAIL
8 effective prompts for debugging code, v/@systemdesignone.
-
MIT CSAIL
@MIT_CSAIL
GenAI can stylize 3D models before they're 3D printed into everyday items, but often weaken its mechanical integrity.
“MechStyle” simulates how genAI's edits impact the model's structure & ensures designs for things like glasses & hooks sustain daily use: news.mit.edu/2026/genai-tool-helps-3d-print-personal-item… -
MIT CSAIL reposted
Boyuan Chen
@BoyuanChen0
Introducing Large Video Planner (LVP-14B) — a robot foundation model that actually generalizes. LVP is built on video gen, not VLA. As my final work at @MIT, LVP has all its eval tasks proposed by third parties as a maximum stress test, but it excels!🤗
boyuan.space/large-video-planner -
MIT CSAIL
@MIT_CSAIL
A free throwback MIT course breaking down how machine learning techniques can be applied to healthcare: ocw.mit.edu/courses/6-s897-machine-learning-for-healthcar… (v/@MITOCW)
Here, MIT prof. & CSAIL principal investigator David Sontag discusses how AI can help sort thru medical data (Lecture 1).This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
The fastest growing & declining jobs by 2030, v/@wef.
-
MIT CSAIL
@MIT_CSAIL
The early days of Amazon — a company now valued at ~$2.4 trillion.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
"Beware of bugs in the above code; I have only proved it correct, not tried it." — Donald Knuth, Turing winner & author of "The Art of Computer Programming," who turns 88 today.
Image: Brian Flaherty for The New York Times -
MIT CSAIL
@MIT_CSAIL
MIT PhD student explains how AI is helping scientists w/conservation efforts, such as monitoring vulnerable ecosystems.
Full video: shorturl.at/Dejn8 -
MIT CSAIL
@MIT_CSAIL
Best practices for API design, v/Neo Kim.
-
MIT CSAIL
@MIT_CSAIL
A free guide to the fundamentals of machine learning, v/@swapnakpanda: cs.nyu.edu/~mohri/mlbook/
-
MIT CSAIL
@MIT_CSAIL
A signed letter from Steve Jobs back in 1983. It was put up for auction in 2021, eventually selling for nearly $500,000.
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Intel co-founder Gordon Moore (of "Moore’s Law"). In 1965, he predicted that the number of transistors on computer chips would double roughly every two years.
Image: Intel -
MIT CSAIL
@MIT_CSAIL
"Learning to break down problems into smaller pieces is one of the most important skills in computer science/life." — Addy Osmani (@addyosmani)
-
MIT CSAIL
@MIT_CSAIL
What are your goals as a computer scientist this year?
-
MIT CSAIL
@MIT_CSAIL
</2025>
-
MIT CSAIL
@MIT_CSAIL
A timeline of how quantum mechanics has evolved, v/@ScienceMagazine: www.science.org/doi/10.1126/science.ady6092
-
MIT CSAIL
@MIT_CSAIL
What are your AI predictions for 2026?
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Linus Torvalds, the father of Linux: www.zdnet.com/article/what-linus-torvalds-really-thinks-a…
Image via Amanda Lucier/The Washington Post/Getty -
MIT CSAIL
@MIT_CSAIL
Key mathematics equations & tables, v/@codek_tv.
-
MIT CSAIL
@MIT_CSAIL
A free, beginner-friendly guide to data structures in Python, v/@swapnakpanda: donsheehy.github.io/datastructures/fullbook.pdf
-
MIT CSAIL
@MIT_CSAIL
Merry Christmas to all our fellow mathematicians!
-
MIT CSAIL
@MIT_CSAIL
How robots prepare for the holidays, according to Meta AI, ChatGPT, & Midjourney.
-
MIT CSAIL
@MIT_CSAIL
The 25 most important mathematical definitions in data science, v/@DailyDoseOfDS_.
-
MIT CSAIL
@MIT_CSAIL
May I interest you in some empty object cookies?
v/@jeffistyping -
MIT CSAIL
@MIT_CSAIL
"The value of a prototype is in the education it gives you, not in the code itself." — Alan Cooper
-
MIT CSAIL
@MIT_CSAIL
Micro banners from 90s webpages & software, v/@WebDesignMuseum.
More classic designs: www.webdesignmuseum.org/exhibitions/web-banners-in-the-90s -
MIT CSAIL
@MIT_CSAIL
Born #otd in 1913: Sister Mary Kenneth Keller, who helped develop Basic. In 1965, she became the first woman to earn a PhD in computer science in the US: www.computer.org/csdl/magazine/an/2023/01/10076929/1LFQ2k…
-
MIT CSAIL
@MIT_CSAIL
A free MIT course breaking down mathematics for computer science & engineering: ocw.mit.edu/courses/6-042j-mathematics-for-computer-scien… (v/@MITOCW)
Here, MIT prof. Tom Leighton, who is also the CEO & co-founder of Akamai, discusses what a proof is (Lesson 1). -
MIT CSAIL
@MIT_CSAIL
LMs can struggle w/open-ended tasks that have constraints, such as advanced puzzles & math proofs.
MIT’s "DisCIPL" uses an LLM to steer smaller LMs to collaborate on these prompts. It achieves accuracy & efficiency comparable to much leading models: news.mit.edu/2025/enabling-small-language-models-solve-co… -
MIT CSAIL
@MIT_CSAIL
Andrej Karpathy, Eureka Labs founder & former Director of AI at Tesla, breaks down how LLMs like ChatGPT "download the Internet."
Watch the full beginner-friendly breakdown here: www.youtube.com/watch?v=7xTGNNLPyMI -
MIT CSAIL
@MIT_CSAIL
What was your favorite AI breakthrough this year?
-
MIT CSAIL
@MIT_CSAIL
The 5 websites LLMs cited the most in October (v/Semrush):
1. Reddit
2. LinkedIn
3. Wikipedia
4. Medium
5. YouTube
More data: www.semrush.com/blog/most-cited-domains-ai/ -
MIT CSAIL
@MIT_CSAIL
We mourn the passing of Frank Gehry (1929-2025), the architect who designed the Ray & Maria Stata Center.
He was a visionary known for playing w/a mix of curvy edges, steel, & occasional bright colors, including the ones you see across Stata. Gehry's other iconic designs include the Guggenheim Museum in Bilbao, the Disney Concert Hall in Los Angeles, & the Louis Vuitton Foundation in Paris, France.
The distinctive, colorful look of Stata reflects the creative fire Gehry brought to each building he envisioned. Undaunted by convention, he wanted his architectural projects to "defy categorization." Gehry's work certainly did — though "futuristic" would be a fitting description.
His work continues to foster the innovations taking place every day at CSAIL. Daniela Rus, MIT Prof. & CSAIL Director, on the ways he impacted our community:
"As we reflect on Frank Gehry’s extraordinary legacy, I’m grateful for the daily reminder of his vision & for the creativity his work continues to inspire. My condolences go out to his family, colleagues, & the many communities shaped by his remarkable contributions to architecture & design." -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late computing innovator Grace Hopper, who helped shape COBOL, made pioneering contributions to compilers, & popularized the word "bug."
Watch her discuss the importance of information during a lecture at MIT in 1985: www.youtube.com/watch?v=ZR0ujwlvbkQ -
MIT CSAIL
@MIT_CSAIL
This is what one "byte" of RAM looked like in 1946.
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Geoffrey Hinton, whose work in neural networks has led many to call him the "Godfather of AI."
These pioneering efforts won him the Nobel Prize last year: www.nytimes.com/2024/10/08/technology/nobel-prize-geoffre…
Image v/Aaron Vincent Elkaim & @nytimes -
MIT CSAIL
@MIT_CSAIL
"The function of good software is to make the complex appear to be simple." — Grady Booch
-
MIT CSAIL
@MIT_CSAIL
A new study from McKinsey reveals that many organizations are adopting AI, including AI agents. Still, many are in the early stages of scaling this technology: www.mckinsey.com/capabilities/quantumblack/our-insights/t…
-
MIT CSAIL
@MIT_CSAIL
Do "untrainable" neural networks just need a helping hand to do modern tasks?
MIT study reveals many "ineffective" networks may start from suboptimal starting points. CSAIL's short-term guidance method transferred structural knowledge to boost performance: -
MIT CSAIL
@MIT_CSAIL
MIT system uses rigorous mathematics to ensure robots flex, adapt, & interact w/people & objects in a safe & precise way.
It helps robots remain flexible & responsive w/o exceeding safe force limits: news.mit.edu/2025/new-control-system-teaches-soft-robots-… -
MIT CSAIL
@MIT_CSAIL
All major neural networks, in one chart.
v/The Asimov Institute -
MIT CSAIL
@MIT_CSAIL
"The first step of any project is to grossly underestimate its complexity and difficulty." — Nicoll Hunt
-
MIT CSAIL
@MIT_CSAIL
A breakdown of all Python essentials & how they can help you code data visualizations, v/@TDataScience: towardsdatascience.com/data-visualization-explained-part-…
-
MIT CSAIL
@MIT_CSAIL
Types of matrices & their operations, v/@PythonPr.
-
MIT CSAIL reposted
Sara Beery
@sarameghanbeery
How do we do more with biodiversity data we've already collected?
I gave a TED Talk on scientific discovery in ecological databases at a @TEDCountdown and @BezosEarthFund event for #nyclimateweek this year, and it's now live!
@TEDTalks @inaturalist #aiforconservation -
MIT CSAIL
@MIT_CSAIL
The AI boom, as explained in 16 charts: www.understandingai.org/p/16-charts-that-explain-the-ai-boom
-
MIT CSAIL
@MIT_CSAIL
Math inspires us all!
Happy #fibonacciday 11/23: www.livescience.com/37470-fibonacci-sequence.html -
MIT CSAIL
@MIT_CSAIL
In 1984, about 8% of homes in the US had computers. Today, roughly 97% do.
Stat v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
A map of ARPANET back in 1977. It was the first large-scale computer network, connecting institutions like MIT & Stanford. The system was a predecessor to the Internet.
v/The Computer History Museum -
MIT CSAIL
@MIT_CSAIL
The first password thief (at least, as we know it today) was born #otd in 1940. Allan Scherr was only allowed to use MIT’s time-sharing system 4 hrs/week, so he found a way to print out all the PWs and started logging in as other users.
it.slashdot.org/story/12/01/28/024220/how-allan-scherr-ha… -
MIT CSAIL
@MIT_CSAIL
The Fourier transform, visualized.
-
MIT CSAIL
@MIT_CSAIL
A free breakdown of all things math, including the questions you didn't know to ask: www.math.cmu.edu/~jmackey/151_128/bws_book.pdf
-
MIT CSAIL
@MIT_CSAIL
#otd in 1971 Intel unveiled the first single-chip CPU. 4 bits, 16 pins & 740KHz: spectrum.ieee.org/chip-hall-of-fame-intel-4004-microproce…
-
MIT CSAIL
@MIT_CSAIL
Happy bday to MIT prof. Shafi Goldwasser, Turing winner & pioneer in the field of cryptography: news.mit.edu/2013/goldwasser-and-micali-win-turing-award-…
-
MIT CSAIL
@MIT_CSAIL
A breakdown of the simplest neural network from the late MIT Prof. Patrick Winston during his 6.034 (Artificial Intelligence) course.
Clip v/@7etsuo & @MITOCW -
MIT CSAIL
@MIT_CSAIL
"Any sufficiently advanced technology is indistinguishable from magic." — Arthur C. Clarke
-
MIT CSAIL
@MIT_CSAIL
MIT robots at work!
Photos v/Alex Gagne -
MIT CSAIL
@MIT_CSAIL
15 courses for understanding neural networks, v/@tut_ml: shorturl.at/60Se8
-
MIT CSAIL
@MIT_CSAIL
"In programming, if someone tells you 'you’re overcomplicating this,' they’re either 10 steps behind you or 10 steps ahead of you." — Andrew Clark
credit: @CodeWisdom -
MIT CSAIL
@MIT_CSAIL
Happy birthday to MIT Prof. Barbara Liskov, Turing winner & programming pioneer.
Watch her break down data abstraction & object-oriented programming: www.youtube.com/watch?v=yCjEYSiVB04 -
MIT CSAIL
@MIT_CSAIL
While programming w/LLMs seems promising, code can be messy, opaque, & hard to change safely.
An MIT case study charts a more legible, modular path forward for software, bringing together features that'd otherwise be scattered across multiple services: news.mit.edu/2025/mit-researchers-propose-new-model-for-l… -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Michael Dertouzos, who directed MIT’s Laboratory for Computer Science for 25+ years.
Watch him discuss the future of computers on NBC’s Today Show w/late MIT prof. Joseph Weizenbaum back in 1983. -
MIT CSAIL reposted
Eva Yi Xie
@EvaYiXie
Connectome suggests brain’s synaptic weights follow heavy-tailed distributions, yet most analyses of RNNs assume Gaussian connectivity.
🧵⬇️ Our @AllenInstitute #neurips2025 paper shows heavy-tailed weights can strongly affect dynamics, trade off robustness + attractor dimension.
openreview.net/forum?id=J0SbYYY0po -
MIT CSAIL
@MIT_CSAIL
#otd in 1971 Bell Labs released the first Unix Programmer's Manual.
Download the free PDF here (third edition): shorturl.at/8NP0j
(v/@CoolSWEng) -
MIT CSAIL
@MIT_CSAIL
What’s one thing you’d like AI to be able to do, but it can’t yet?
-
MIT CSAIL
@MIT_CSAIL
A genAI model developed at MIT designs proteins & peptides of any modality to bind to different biomolecular targets.
The system, called "BoltzGen," unifies design & structure prediction. It's freely available for unrestricted academic & commercial use: boltz.com/boltzgen -
MIT CSAIL
@MIT_CSAIL
#otd in 1998 Yann LeCun submitted a paper on “gradient-based” deep learning for document recognition.
It took more than a decade before the world finally warmed to neural networks: www.semanticscholar.org/paper/Gradient-based-learning-app…
First image v/Erik Tanner & WIRED -
MIT CSAIL
@MIT_CSAIL
Why can’t programmers tell the difference between Halloween & Christmas?
Because oct 31 = dec 25. -
MIT CSAIL
@MIT_CSAIL
How robots would celebrate Halloween at MIT's Stata Center, according to ChatGPT, Midjourney, & Gemini.
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Marcian “Ted” Hoff, Jr., an engineer who co-invented the microprocessor.
How he did it: spectrum.ieee.org/ted-hoff -
MIT CSAIL reposted
MIT Jameel Clinic for AI & Health
@AIHealthMIT
Join us in-person at MIT on Oct. 30 @ 5pm to hear from @HannesStaerk on BoltzGen and the state of biomolecular design!
🗓️RSVP now: jclinic.mit.edu/events/boltzgen/ -
MIT CSAIL reposted
Hannes Stark
@HannesStaerk
Excited to release BoltzGen which brings SOTA folding performance to binder design! The best part of this project has been collaborating with many leading biologists who tested BoltzGen at an unprecedented scale, showing success on many novel targets and pushing its limits! 🧵..
-
MIT CSAIL reposted
Gabe Grand
@gabe_grand
Do AI agents ask good questions? We built “Collaborative Battleship” to find out—and discovered that weaker LMs + Bayesian inference can beat GPT-5 at 1% of the cost.
Paper, code & demos: gabegrand.github.io/battleship
Here's what we learned about building rational information-seeking agents... 🧵🔽 -
MIT CSAIL
@MIT_CSAIL
14 sorting algorithms in just 60 seconds.
v/@7etsuo -
MIT CSAIL
@MIT_CSAIL
To give context on age, Grace Hopper began computing at 38, completed the first compiler at 46, helped shape COBOL at 53, kept developing COBOL for the Navy in her 70s, retired from the Navy at 80 & then became a consultant for the Digital Equipment Corporation.
v/@cooperx86 -
MIT CSAIL
@MIT_CSAIL
Microsoft Word was released #otd in 1983. This is how it looked it back then.
Image v/Aldo Computers -
MIT CSAIL
@MIT_CSAIL
Ask us your questions about how AI & computer vision can help w/conservation efforts. We’re featuring a few in an upcoming explainer w/MIT PhD student Justin Kay.
For more on his work: justinkay.github.io -
MIT CSAIL reposted
Sharut Gupta
@sharut_gupta
[1/7] Paired multimodal learning shows that training with text can help vision models learn better image representations. But can unpaired data do the same?
Our new work shows that the answer is yes!
w/ @shobsund @ChenyuW64562111, Stefanie Jegelka and @phillip_isola -
MIT CSAIL reposted
Phillip Isola
@phillip_isola
If you are at ICCV, I'm giving a talk here at 3:30pm in Ballroom B on "Revisiting the symbol grounding problem in the age of multi-modal AI"
Will cover recent work on multimodal rep alignment in unpaired and unimodal models.
Shiry Ginosar
@shiryginosar
Join us TODAY for the 3rd Perception Test Challenge perception-test-challenge.github.io @ICCV2025!
Ballroom B, Full day
Amazing lineup of speakers: @farhadi, @AlisonGopnik, Phlipp Krahenbul, @phillip_isola -
MIT CSAIL
@MIT_CSAIL
It can be wasteful & expensive to keep up w/the latest fashion trends.
Luckily, MIT & Adobe’s “Refashion” program helps users create outlines for clothing w/components you can replace or restyle. It’s a more sustainable approach to fashion design: news.mit.edu/2025/refashion-software-designs-eco-friendly… -
MIT CSAIL
@MIT_CSAIL
BOOlean logic to brush up on before Halloween.
v/@38mo1 -
MIT CSAIL
@MIT_CSAIL
MIT Prof. & CSAIL principal investigator Dina Katabi has been elected to the National Academy of Medicine for building wireless devices that assist w/digital health using AI & radio signals, helping doctors detect diseases like Parkinson’s from digital biomarkers.
Having previously been elected to National Academy of Engineering & the National Academy of Sciences, she's now a member of all three National Academies: www.csail.mit.edu/news/dina-katabi-elected-national-acade… -
MIT CSAIL
@MIT_CSAIL
Some recommended reads, each w/MIT CSAIL authors.
-
MIT CSAIL
@MIT_CSAIL
Machine learning algorithms cheat sheet: medium.com/dataflair/beat-the-heat-with-machine-learning-…
v/Himani Bansal -
MIT CSAIL
@MIT_CSAIL
Key AI tools to use for things like video generation, data visualization, & coding assistance, v/René Remsik.
-
MIT CSAIL
@MIT_CSAIL
A free, beginner-friendly course from MIT on computing fundamentals. These classic lessons break down concepts such as programming language design, abstraction, & recursion: www.youtube.com/watch?v=-J_xL4IGhJA&list=PLE18841CABEA240…
v/@MITOCW -
MIT CSAIL
@MIT_CSAIL
Sometimes generative models struggle to follow even the most specific prompts.
MIT's new “GLASS Flows” approach boosts text-image alignment for large-scale models at inference time. It uses ODEs to simulate random changes over time — w/o any retraining: x.com/peholderrieth/status/1976338189834141870 -
MIT CSAIL
@MIT_CSAIL
MIT professor discusses what role AI should play when educators are teaching students how to program.
Part 2 of the full video: shorturl.at/5PtCB -
MIT CSAIL
@MIT_CSAIL
In honor of #adalovelaceday, learn about the woman who many view as the world's first programmer: cnet.co/2dZ1Ihp
-
MIT CSAIL
@MIT_CSAIL
MIT ranked No. 1 university in the US by Times Higher Education: www.timeshighereducation.com/student/best-universities/be…
-
MIT CSAIL
@MIT_CSAIL
MIT PhD student Justin Kay is developing computer vision algorithms that help monitor vulnerable ecosystems, supporting wildlife in the midst of threats such as habitat alterations & climate change.
More on these approaches, including "CODA," which helps conservationists choose which AI model to use on a particular dataset: www.csail.mit.edu/news/3-questions-how-ai-helping-us-moni… -
MIT CSAIL
@MIT_CSAIL
"Telling a programmer there’s already a library to do X is like telling a songwriter there’s already a song about love." — Pete Cordell
-
MIT CSAIL
@MIT_CSAIL
Linear regression, as explained by @rfeers.
-
MIT CSAIL
@MIT_CSAIL
To become home & factory assistants, robots will need to train on lots of demonstrations.
MIT & Toyota's GenAI tool creates virtual training grounds to help them inch toward that vision, arranging 3D items into physically realistic kitchens & restaurants: news.mit.edu/2025/using-generative-ai-diversify-virtual-t… -
MIT CSAIL
@MIT_CSAIL
MIT professor on how AI is reshaping what it means to "write code."
Part 1 of the full video: shorturl.at/DdmSF -
MIT CSAIL
@MIT_CSAIL
The Stata Center in 1999 ⏩ Today
The initial freeform sketch was drawn by Frank Gehry as an early mockup of the building (v/Gehry Partners). -
MIT CSAIL
@MIT_CSAIL
What are people asking ChatGPT?
v/@VisualCap -
MIT CSAIL
@MIT_CSAIL
Types of data structures, v/@PythonPr.
-
MIT CSAIL
@MIT_CSAIL
58 years ago, Larry Roberts presented his idea for an "ARPANet" for connecting multiple computers together across the United States.
Full paper: shorturl.at/6uMNq -
MIT CSAIL
@MIT_CSAIL
MIT & McMaster researchers have developed a compound that targets gut inflammation.
Called enterololin, it takes on gut bacteria tied to Crohn’s flare-ups while sparing the microbiome. The team mapped its action in months instead of years using genAI: www.csail.mit.edu/news/ai-maps-how-new-antibiotic-targets… -
MIT CSAIL
@MIT_CSAIL
The most commonly-used data cleaning commands in Python.
v/@Eyowhite3 -
MIT CSAIL
@MIT_CSAIL
The Turing Test was born #otd in 1950 w/Alan Turing’s seminal paper "Computing Machinery and Intelligence."
academic.oup.com/mind/article/LIX/236/433/986238 -
MIT CSAIL
@MIT_CSAIL
A CAPTCHA nightmare for programmers.
-
MIT CSAIL
@MIT_CSAIL
In German, the word "verschlimmbessern" means "to make something worse while trying to repairing it." In English, we call that "software development."
-
MIT CSAIL
@MIT_CSAIL
How do LLMs respond to questions about the election?
MIT researchers ran a 4-month study, querying 12 state-of-the-art models on over 12K prompts daily. They found that the models can reinforce certain political narratives, mislead, or even exhibit manipulative tendencies:Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
The team envisions comparing model outputs w/real-world polls, expanding studies to other democracies, examining newer reasoning or "thinking" models, & exploring downstream effects on users.
For now, the dataset helps us understand how LLMs are shaping modern political discourse.
MIT CSAIL
@MIT_CSAIL
Authors: @cen_sarah, @andrew_ilyas, Hedi Driss, Charlotte Park, @aspenkhopkins, @charapod, & @aleks_madry
Data: huggingface.co/datasets/sarahcen/llm-election-data-2024 -
MIT CSAIL
@MIT_CSAIL
A Python cheat sheet.
v/@QCompounding -
MIT CSAIL
@MIT_CSAIL
What if we could predict customer attrition with high accuracy w/o training on sensitive data?
MIT’s "FairRAG" is fast & privacy-preserving. It uses differentially private synthetic data & retrieval-augmented generation — & outperforms traditional ML & LLMs: www.csail.mit.edu/news/enhancing-privacy-and-decision-mak… -
MIT CSAIL
@MIT_CSAIL
A reliable excuse for programmers, v/@xkcd.
-
MIT CSAIL
@MIT_CSAIL
A free roadmap for AI engineering w/helpful resources: github.com/chiphuyen/aie-book
v/@HeyNina101 -
MIT CSAIL
@MIT_CSAIL
23 years ago today Tor was launched by MIT students for anonymous communications: www.torproject.org/about/history/
-
MIT CSAIL
@MIT_CSAIL
Jeff Dean, Google's Chief Scientist, tells a student what it's like to be a computer scientist.
v/@JeffDean -
MIT CSAIL
@MIT_CSAIL
"Design is not just what it looks like and feels like. Design is how it works." — Steve Jobs
-
MIT CSAIL
@MIT_CSAIL
Doctors monitoring prenatal health often turn to MRIs to get a closer look, but they can be hard to interpret.
An MIT system uses ML to model the shape & movements of fetuses in 3D more clearly, potentially helping doctors spot anomalies & make diagnoses: news.mit.edu/2025/machine-learning-tool-gives-doctors-mor… -
MIT CSAIL
@MIT_CSAIL
Google Maps on its launch day, back in 2005.
Initially, it only featured North America & the British Isles (v/@amazingmap). -
MIT CSAIL
@MIT_CSAIL
A visualization of how NBA players' shot selection has changed over the past 25 years. The mid-range jump shot has seen a noticeable dip in frequency, while three-point attempts have soared.
More: algorithmicathlete.com/blog/is-mid-range-dead
v/The Algorithmic Athlete -
MIT CSAIL
@MIT_CSAIL
What’s an invention you feel deserves more credit?
-
MIT CSAIL
@MIT_CSAIL
Beginner-friendly LLM course exploring foundations, architectures, training, deployment, and current trends: github.com/mlabonne/llm-course?tab=readme-ov-file
v/@Hesamation -
MIT CSAIL reposted
Ticha Sethapakdi
@tichaesque
Thank yous @MIT_CSAIL for the lovely article!! I had a lot of fun working on this paper and I’m so excited to present it at @ACMUIST this year :))
MIT CSAIL
@MIT_CSAIL
Barrier-grid animations are a clever way to create a moving picture that doesn’t actually move — w/o electronics.
MIT's "FabObscura" software helps you produce bolder designs for these displays, which can be attached to objects in homes, stores, & more: shorturl.at/M4Jl3 -
MIT CSAIL
@MIT_CSAIL
Barrier-grid animations are a clever way to create a moving picture that doesn’t actually move — w/o electronics.
MIT's "FabObscura" software helps you produce bolder designs for these displays, which can be attached to objects in homes, stores, & more: shorturl.at/M4Jl3 -
MIT CSAIL
@MIT_CSAIL
Today would have been the 84th birthday of Dennis Ritchie, inventor of C and co-creator of Unix: zd.net/2creeZi
(Article v/@jperlow, image v/Victoria Will/AP) -
MIT CSAIL
@MIT_CSAIL
How Netflix looked back in 1999.
It now has over 300M subscribers globally.
Image v/@OldInternetFeel -
MIT CSAIL
@MIT_CSAIL
"Learning to break down problems into smaller pieces is one of the most important skills in computer science/life." — Addy Osmani (@addyosmani)
-
MIT CSAIL
@MIT_CSAIL
A free beginner’s guide to machine learning: mlstory.org/pdf/patterns.pdf
v/@swapnakpanda -
MIT CSAIL
@MIT_CSAIL
Google Chrome was publicly released this month in 2008.
It now commands around 70% of the global browser market share, according to some estimates. -
MIT CSAIL
@MIT_CSAIL
Can we build objects that are mostly eco-friendly, but still strong where it counts?
MIT’s "SustainaPrint" software & hardware toolkit strengthens only the weakest parts of greener prints. It reduces plastic use while preserving structural integrity: news.mit.edu/2025/greener-way-3d-print-stronger-stuff-0904 -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late John McCarthy, father of AI & LISP: www.wired.com/2011/10/john-mccarthy-father-of-ai-and-lisp… (credit: @WIRED)
Image v/John McCarthy & @MITMuseum -
MIT CSAIL
@MIT_CSAIL
Ask us your questions about AI-assisted software engineering & programming languages. We’re featuring a few in an upcoming explainer w/MIT prof. & CSAIL Associate Director & COO Armando Solar-Lezama.
For more on his work: people.csail.mit.edu/asolar/ -
MIT CSAIL
@MIT_CSAIL
#otd in 1955, the term “artificial intelligence” was coined in a conference proposal: stanford.io/2WJJJGN
-
MIT CSAIL reposted
Manolis Kellis
@manoliskellis
MIT Course announcement: Machine Learning for Computational Biology #mlcb25
Fall'24 Lecture Videos: www.youtube.com/playlist?list=PLypiXJdtIca4gtioEPLIExlAKv…
Fall'24 Lecture Notes: docs.google.com/document/d/17eG2nS0kM3RyxDZsxpLiSKaNjFLCs…
(a) Genomes: Statistical genomics, gene regulation, genome language models, chromatin structure, 3D genome topology, epigenomics, regulatory networks.
(b) Proteins: Protein language models, structure and folding, protein design, cryo-EM, AlphaFold2, transformers, multimodal joint representation learning.
(c) Therapeutics: Chemical landscapes, small-molecule representation, docking, structure-function embeddings, agentic drug discovery, disease circuitry, and target identification.
(d) Patients: Electronic health records, medical genomics, genetic variation, comparative genomics, evolutionary evidence, patient latent representation, AI-driven systems biology.
Foundations and frontiers of computational biology, combining theory with practice. Generative AI, foundation models, machine learning, algorithm design, influential problems and techniques, analysis of large-scale biological datasets, applications to human disease and drug discovery.
First Lecture: Thu Sept 4 at 1pm in 32-144
With: Prof. Manolis Kellis @manoliskellis, Prof. Eric Alm @ejalm, TAs: Ananth Shyamal, Shitong Luo @luost26
Course website: compbio.mit.edu/mlcb
@MIT @MITEECS @MITdeptofBE @MITCSBPhD @MIT_CSAIL @Harvard @HarvardMed @BroadInstitute
Manolis Kellis
@manoliskellis
Today was my last lecture for @MIT #computationalbiology: #genomes, #networks, #evolution, #health. I recorded each and immediately posted online here: www.youtube.com/playlist?list=PLypiXJdtIca6GBQwDTo4bIEDV8…
Please do share, and let me know which topics need more explanations, clarifications, and corrections! -
MIT CSAIL
@MIT_CSAIL
"Program testing can be a very effective way to show the presence of bugs, but it is hopelessly inadequate for showing their absence." — Edsger Dijkstra
-
MIT CSAIL
@MIT_CSAIL
MIT prof. Regina Barzilay has been named to the TIME100 AI 2025 for developing machine-learning models that can predict diseases like breast cancer: shorturl.at/CCCqT
v/@TIME -
MIT CSAIL
@MIT_CSAIL
A new AI system from MIT could make flu vaccine strain selection more accurate.
Trained on decades of data, "VaxSeer" predicts dominant flu strains and IDs the most protective vaccine candidates. Its recommendations often outperformed/matched the WHO’s: www.csail.mit.edu/news/mit-researchers-develop-ai-tool-im… -
MIT CSAIL
@MIT_CSAIL
Will the future of robotics be written in code or in data?
A debate at ICRA among MIT profs. & their international colleagues revealed that we'll probably need both to unlock robots' real-world potential — but which takes the lead, & when, isn't clear yet: www.csail.mit.edu/news/mit-roboticists-debate-future-robo… -
MIT CSAIL
@MIT_CSAIL
Approx. Internet users globally (v/@JonErlichman):
1995: 16 million 💾
2005: 1 billion 💻
2015: 3.2 billion 📱
2025: 5.6 billion 🌐 -
MIT CSAIL
@MIT_CSAIL
Today in 1991, 21-year-old Linus Torvalds announced in a newsgroup that he was working on a free operating system that later came to be called Linux.
groups.google.com/forum/ -
MIT CSAIL
@MIT_CSAIL
30 years ago: Windows 95 hit stores to the tune of "Start Me Up." Microsoft reportedly paid ~$3 million (roughly $6.3M today) to license the song in their ad.
Video v/Whamtan on YouTube -
MIT CSAIL
@MIT_CSAIL
Now that it's been out a couple weeks, what are your thoughts on GPT-5?
-
MIT CSAIL
@MIT_CSAIL
One of the most influential programming textbooks was first published as a paperback 40 years ago today: MIT's "Structure and Interpretation of Computer Programs."
Read it for free here: rb.gy/5hvui -
MIT CSAIL reposted
Bonnie Berger Lab
@lab_berger
Excited to share our recent work: Sparse autoencoders uncover biologically interpretable features in protein language model representations now in PNAS. Thread below 🧵
-
MIT CSAIL
@MIT_CSAIL
Ranking top AI chatbots over the past year by user experience, visibility, & momentum, v/OneLittleWeb: onelittleweb.com/data-studies/best-ai-chatbots/
ChatGPT came in first, with Grok & Gemini rounding out the top three. -
MIT CSAIL
@MIT_CSAIL
A helpful list of best practices for API design (v/@alexxubyte).
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Margaret Hamilton, an MIT computer scientist who designed the flight software for the Apollo missions in the 1960s.
Image v/El Impresionante on Reddit -
MIT CSAIL
@MIT_CSAIL
ML models often fail to see that symmetric transformations (like rotations) leave data fundamentally unchanged.
MIT designed the first method for ML w/symmetry that's provably efficient in terms of data & compute, possibly aiding drug & materials discovery: news.mit.edu/2025/new-algorithms-enable-efficient-machine… -
MIT CSAIL
@MIT_CSAIL
Free MIT course breaks down foundation models & genAI: www.youtube.com/watch?v=y1fGlAECNFM&list=PLXV9Vh2jYcjbnv6…
Here, CSAIL researcher & lead lecturer Rickard Brüel Gabrielsson discusses AGI & a key innovation behind AI systems: learning from observation (Lecture 1). -
MIT CSAIL
@MIT_CSAIL
Ask us your questions about embodied intelligence or AI systems that interact w/the world. We’re featuring a few in an upcoming explainer w/MIT prof. Vincent Sitzmann (@vincesitzmann).
For more on his work: www.vincentsitzmann.com -
MIT CSAIL
@MIT_CSAIL
Happy birthday, LEGO! Founded in 1932, they now produce 36 billion bricks annually.
There are 80 LEGO bricks for every person in the world (not to mention a few to recreate MIT's Stata Center). -
MIT CSAIL
@MIT_CSAIL
The Harwell Dekatron is the world's oldest original working digital computer.
Built in 1951, it needs 10 seconds to multiply two numbers (v/@JonErlichman). -
MIT CSAIL
@MIT_CSAIL
A free MIT course breaking down fundamental math concepts in computer science: www.youtube.com/watch?v=sbpCTjmw85g&list=PLUl4u3cNGP61VNv…
Here, MIT prof. Erik Demaine breaks down state machines (Lecture 4).
v/@MITOCW -
MIT CSAIL
@MIT_CSAIL
M.C. Escher's artwork famously features depth-defying optical illusions.
New MIT tool creates multi-dimensional representations of these "impossible objects" that users can relight, smooth, & solve geometry problems on w/o bending or cutting shapes: news.mit.edu/2025/mit-meschers-tool-visualizes-edits-phys… -
MIT CSAIL
@MIT_CSAIL
#otd in 1991 Tim Berners-Lee first posted on an online newsgroup about a project he was working on called "The World Wide Web."
w3.org/People/Berners-Lee/1991/08/art-6487.txt -
MIT CSAIL
@MIT_CSAIL
Lung cancer is the leading cause of cancer-related deaths worldwide.
A deep learning model developed by MIT & MGH researchers called "Sybil" can accurately predict a patient’s risk of lung cancer up to 6 years in advance by analyzing a low-dose CT scan: www.youtube.com/watch?v=zZz80eC4mRgThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
Image generation systems normally use a tokenizer to compress & encode visual data, along w/a generator that combines & arranges these representations to create images.
MIT's new method creates, converts, & "inpaints" images w/o using a generator at all: news.mit.edu/2025/new-way-edit-or-generate-images-0721 -
MIT CSAIL
@MIT_CSAIL
"A computer will do what you tell it to do, but that may be much different from what you had in mind." — Joseph Weizenbaum
Image v/Joachim Schulz & Getty Images -
MIT CSAIL
@MIT_CSAIL
This month in 1981, IBM introduced its first PC. With a 4.77 MHz Intel 8088 microprocessor, the $1600 “Model 5150” soon became the best-selling computer in the world: www.wired.com/2011/08/0812ibm-5150-personal-computer-pc/
-
MIT CSAIL
@MIT_CSAIL
This diagram represents what's widely considered to be the world's first "non-trivial" computer program, developed in 1843 by Ada Lovelace for Charles Babbage's Analytical Engine.
(v/@fermatslibrary) -
MIT CSAIL
@MIT_CSAIL
How do language models track dynamic scenarios, like completing code or guessing your responses?
MIT research finds that LMs don't track "state changes" step by step — they use mathematical shortcuts that we can control to boost LMs' prediction skills: news.mit.edu/2025/unique-mathematical-shortcuts-language-… -
MIT CSAIL
@MIT_CSAIL
24 resources to help you learn about software architecture, v/@alexxubyte.
-
MIT CSAIL
@MIT_CSAIL
To give context on age, Grace Hopper began computing at 38, completed the first compiler at 46, helped shape COBOL at 53, kept developing COBOL for the Navy in her 70s, retired from the Navy at 80 & then became a consultant for the Digital Equipment Corporation.
v/@cooperx86 & photo v/Cynthia Johnson & Getty Images -
MIT CSAIL
@MIT_CSAIL
If you could build a specialized robot to automate one task, which one would you choose?
-
MIT CSAIL
@MIT_CSAIL
A report card grading the safety of leading AI model-makers, v/@FLI_org: futureoflife.org/wp-content/uploads/2025/07/AI-Safety-Ind…
Anthropic topped the list w/a C+, while DeepSeek scored lowest w/an F. -
MIT CSAIL
@MIT_CSAIL
A surprising mathematical pattern w/palindromes.
-
MIT CSAIL reposted
MIT OpenCourseWare
@MITOCW
We are so excited to share that 6.1200J Mathematics for Computer Science is now ready and waiting for you! Watch the videos on YouTube: www.youtube.com/playlist?list=PLUl4u3cNGP61VNvICqk2HXJTon…
Taught by Prof. Erik Demaine, Dr. Zachary Abel, and Dr. Byrnmor Chapman, this course covers elementary discrete mathematics for science and engineering, with a focus on mathematical tools and proof techniques useful in computer science. 💻
@MITEECS -
MIT CSAIL
@MIT_CSAIL
Can AI actually code for us? 🧵
MIT research reveals there’s a "long way to go" due to bottlenecks like assessment, codebase scale, & incorrect retrievals. The work reflects a vision to let humans focus on high-level design while routine work is automated: news.mit.edu/2025/can-ai-really-code-study-maps-roadblock…Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
Finally, we lay out a vision for ways we can move forward as a field. We discuss the necessity of large-scale community-driven efforts to collect open datasets & train open models for code.
We also discuss techniques such as designing RL environments, training LLMs to program w/humans collaboratively, & incorporating tools from research in software engineering.
MIT CSAIL
@MIT_CSAIL
Authors: Alex Gu, Naman Jain, Wen-Ding Li, Manish Shetty, Yijia Shao, Ziyang Li, Diyi Yang, Kevin Ellis, Koushik Sen, & Armando Solar-Lezama
Paper: arxiv.org/abs/2503.22625 -
MIT CSAIL
@MIT_CSAIL
A new handheld interface from MIT gives anyone the ability to train a robot for tasks in fields like manufacturing.
The versatile tool can teach a robot new skills using one of three approaches: natural teaching, kinesthetic training, & teleoperation: news.mit.edu/2025/new-tool-gives-anyone-ability-to-train-… -
MIT CSAIL
@MIT_CSAIL
56 years ago today, humans walked on the moon for the first time.
Original @TIME report of the monumental event, from 1969: time.com/5629566/1969-moon-landing-time/ -
MIT CSAIL
@MIT_CSAIL
"Programming is not about typing, it's about thinking." — Rich Hickey
v/@CodeWisdom -
MIT CSAIL
@MIT_CSAIL
A step-by-step guide to diffusion models: arxiv.org/pdf/2406.08929
v/@goyal__pramod
MIT CSAIL
@MIT_CSAIL
Authors: @PreetumNakkiran, @ArwenBradley, @oh_that_hat, & @advani_madhu
-
MIT CSAIL
@MIT_CSAIL
The Fibonacci sequence in Pascal's Triangle, v/@TheMathFlow.
-
MIT CSAIL
@MIT_CSAIL
MIT CSAIL's "PhysicsGen" system helps robots handle items efficiently by customizing & multiplying the data they train on.
It turns a few VR demonstrations into thousands of simulations, potentially helping build huge datasets for dexterous robots: news.mit.edu/2025/simulation-based-pipeline-tailors-train… -
MIT CSAIL
@MIT_CSAIL
Hello world! This is @nishanthkumar23 and I'll be taking over @MIT_CSAIL 's X/Twitter & Instagram for 24 hours! I'm a 4th year PhD @MITEECS working on AI/ML for Robotics and Computer Agents. Drop any and all questions about research, AI, MIT, or dogs (esp. robot dogs!) below 👇
-
MIT CSAIL
@MIT_CSAIL
A new approach from CSAIL & Google marks a shift toward teaching models to orchestrate their own parallel decoding strategy.
The team's "Parallel Structure Annotation" (PASTA) enables LLMs to generate text in parallel, accelerating their response times: www.csail.mit.edu/news/speeding-llm-inference-parallelism -
MIT CSAIL
@MIT_CSAIL
A guide to competitive programming: duoblogger.github.io/assets/pdf/memonvyftw/guide-t-cp.pdf
-
MIT CSAIL reposted
Russ Tedrake
@RussTedrake
TRI's latest Large Behavior Model (LBM) paper landed on arxiv last night! Check out our project website: toyotaresearchinstitute.github.io/lbm1/
One of our main goals for this paper was to put out a very careful and thorough study on the topic to help people understand the state of the technology, and to share a lot of details for how we're achieving it.
www.youtube.com/watch?v=BEXFnru5-NA&list=PL1V6-cXYDNBAGag… -
MIT CSAIL
@MIT_CSAIL
Which computer science innovator — past or present — would you like to have dinner with?
-
MIT CSAIL
@MIT_CSAIL
A Python cheat sheet.
v/@QCompounding -
MIT CSAIL
@MIT_CSAIL
Happy birthday, Nikola Tesla!
Image v/Culver Pictures -
MIT CSAIL
@MIT_CSAIL
AVs called "gliders" drift thru the ocean to collect data about aquatic environments.
MIT approach uses machine learning to explore new glider designs, testing & refining designs in simulation. It could help scientists monitor marine life: news.mit.edu/2025/ai-shapes-autonomous-underwater-gliders…This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
The 20 system design concepts every developer should know, according to @sahnlam.
-
MIT CSAIL
@MIT_CSAIL
Robots consider many actions, constraints, & mechanical capabilities when handling objects.
An approach from MIT CSAIL & NVIDIA speeds up this planning process, helping robots "think ahead" by considering thousands of solutions & refining the best ones: news.mit.edu/2025/new-system-enables-robots-to-solve-mani… -
MIT CSAIL
@MIT_CSAIL
"Controlling complexity is the essence of computer programming." — Brian Kernighan
-
MIT CSAIL
@MIT_CSAIL
A list explaining key ML concepts and papers: github.com/dair-ai/ML-Papers-Explained
v/@omarsar0 -
MIT CSAIL
@MIT_CSAIL
Where are the world's AI data centers? A recent report from @nytimes reveals only 32 countries have them: www.nytimes.com/interactive/2025/06/23/technology/ai-comp…
-
MIT CSAIL
@MIT_CSAIL
77 years ago this week Claude Shannon ushered in the field of information theory with his paper "A Mathematical Theory of Communication," which has been cited over 100,000 times: people.math.harvard.edu/~ctm/home/text/others/shannon/ent…
-
MIT CSAIL
@MIT_CSAIL
MIT CSAIL was founded #otd in 1963 as "Project MAC."
-
MIT CSAIL
@MIT_CSAIL
Could generative AI improve how robots perform?
MIT method uses genAI to refine specific parts of a robot blueprint & tests new 3D designs in simulation. It created a machine that out-jumped & landed more consistently than a human-designed robot: www.youtube.com/watch?v=dI1J4aPcbmc -
MIT CSAIL
@MIT_CSAIL
Apple released the first iPhone #otd in 2007.
18 years later, it has over 1.3 billion active users.
Image v/Sarah Tew & CNET -
MIT CSAIL
@MIT_CSAIL
A flow chart for writing a research proposal, v/@acagamic.
-
MIT CSAIL
@MIT_CSAIL
Can a robot learn what its own body looks like — just by watching itself move?
W/Neural Jacobian Fields (NJF), CSAIL researchers show that it can. Using just one external camera, NJF enables robots to infer the structure of their joints & limbs — no built-in sensors or prior models required.
This work opens new doors for adaptable, self-aware machines: www.csail.mit.edu/news/robots-know-themselves-mits-vision… -
MIT CSAIL
@MIT_CSAIL
LLMs sometimes exhibit position bias, where they overemphasize the beginning or end of a document or conversation while neglecting the middle.
A theoretical framework from MIT revealed that model architectures & training data contribute to this problem: news.mit.edu/2025/unpacking-large-language-model-bias-0617 -
MIT CSAIL
@MIT_CSAIL
A free MIT guide to key computer vision concepts: visionbook.mit.edu
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Alan Turing, the creator of modern computing.
Image v/Fine Art Images & Heritage Images/age fotostock -
MIT CSAIL
@MIT_CSAIL
QS ranks MIT the world's No. 1 university for the 14th year in a row, placing first in data science & AI, computer science, electrical & electronic engineering, & more for 2026: news.mit.edu/2025/qs-ranks-mit-worlds-no-1-university-0618
-
MIT CSAIL
@MIT_CSAIL
Time flies: GPT-4o (2025) vs. DALL·E 1 (2021).
-
MIT CSAIL
@MIT_CSAIL
Give us your hot takes about generative AI.
-
MIT CSAIL
@MIT_CSAIL
A roadmap of AI & machine learning, v/@hamptonism.
-
MIT CSAIL
@MIT_CSAIL
"Commenting your code is like cleaning your bathroom — you never want to do it, but it really does create a more pleasant experience for you and your guests." — Ryan Campbell (credit: @CodeWisdom)
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late John Tukey, who coined the terms “software” and “bit” (short for “binary digit”). He would have turned 1101110 today: www.nytimes.com/2000/07/28/us/john-tukey-85-statistician-…
v/@nytimes, photo by Alfred Eisenstadt -
MIT CSAIL
@MIT_CSAIL
50 key questions explaining core LLM concepts, v/Hao Hoang: drive.google.com/file/d/1wolNOcHzi7-sKhj5Hdh9awC9Z9dWuWMC…
-
MIT CSAIL
@MIT_CSAIL
True story: a kid saw this and said, "oh, you 3D-printed the 'Save' Icon."
Credit: @Bill_Gross -
MIT CSAIL
@MIT_CSAIL
How long it took 25% of the US to adopt new technology:
⚡ Electricity: 46 years
☎️ Telephone: 35 years
📻 Radio: 31 years
📺 Television: 26 years
🖥️ Personal computers: 16 years
🌐 World Wide Web: 7 years
📲 Smartphones: 4 years
🤖 GenAI: 2 years
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
MIT CSAIL researchers developed a computationally efficient method that could enable artists to design realistic simulations of elastic objects, like bouncy or squishy characters, for animated movies or video games: news.mit.edu/2025/animation-technique-simulates-squishy-o…
-
MIT CSAIL
@MIT_CSAIL
MIT CSAIL Director Daniela Rus has been named a 2025 Tech Power Player by @BostonGlobe.
Full list: globe.com/tech50 #globetech50 -
MIT CSAIL
@MIT_CSAIL
The earliest written instance of the word "byte" is from a paper published #otd in 1956 by IBM’s Werner Buchholtz.
Less commonly used is the word for a half-byte: a "nibble."
archive.computerhistory.org/resources/text/IBM/Stretch/pd… -
MIT CSAIL
@MIT_CSAIL
"Hardware eventually fails. Software eventually works." — Michael Hartung
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Web inventor & MIT Professor Emeritus Tim Berners-Lee! Things besides the WWW that didn’t exist before TBL:
- Apple
- Google
- Facebook
- YouTube
- Instagram
- Netflix
- Uber
- Spotify
- & most technologies we use today
Image v/Sotheby's -
MIT CSAIL
@MIT_CSAIL
A free guide to data structures & algorithms in Python: donsheehy.github.io/datastructures/fullbook.pdf
-
MIT CSAIL reposted
Gabriele Corso
@GabriCorso
Excited to unveil Boltz-2, our new model capable not only of predicting structures but also binding affinities! Boltz-2 is the first AI model to approach the performance of FEP simulations while being more than 1000x faster! All open-sourced under MIT license! A thread… 🤗🚀
This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
MIT CSAIL, @AIHealthMIT, & @RecursionPharma have released a protein-binding affinity model called "Boltz-2" that could expand the role of AI in drug discovery.
The system predicts how molecules interact — & how tightly they bind — in record time: www.csail.mit.edu/news/mit-releases-breakthrough-protein-…2 media couldn't be displayed. See them on the original post. -
MIT CSAIL reposted
Phillip Isola
@phillip_isola
Check out our new work, CycleReward, below!
* Product: a drop-in replacement for things like CLIP; f(img, text) --> score
* Idea: cycle-consistent mappings are good (encourages descriptive text, penalizes hallucinations), so we train a reward model to check for this
Hyojin Bahng
@hyojinbahng
Image-text alignment is hard — especially as multimodal data gets more detailed. Most methods rely on human labels or proprietary feedback (e.g., GPT-4V).
We introduce:
1. CycleReward: a new alignment metric focused on detailed captions, trained without human supervision.
2. CyclePrefDB: 866K preference pairs from cycle consistency.
📄 Paper arxiv.org/abs/2506.02095
🌐 Project Page cyclereward.github.io
💻 Code github.com/hjbahng/cyclereward
(1/🧵) -
MIT CSAIL
@MIT_CSAIL
MIT CSAIL, JHU, & USC researchers have developed "CapSpeech," a text-to-speech framework that generates voices w/controllable timbre & speaking style via text prompts.
Imagine designing a voice just by describing it. CapSpeech can customize age, timbre, accent, emotion, & more: wanghelin1997.github.io/CapSpeech-demo/This media couldn't be displayed. See it on the original post.Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
CapSpeech provides guidelines for data construction & model design to develop prompt-guided expressive text-to-speech systems across use cases like cinematic narration, game character voiceovers, conversational chatbots, emotion therapy, & more.
MIT CSAIL
@MIT_CSAIL
The CSAIL researchers welcome collaboration.
Please contact @mitqinzy and @yaqubhai for more info, including resources. -
MIT CSAIL
@MIT_CSAIL
Sometimes the best way to express an idea is by sketching it out.
A system from MIT CSAIL & Stanford captures this iterative process by teaching LLMs to create sequential sketches. It could work w/users to visually communicate concepts: news.mit.edu/2025/teaching-ai-models-to-sketch-more-like-…
MIT CSAIL
@MIT_CSAIL
This method, called "SketchAgent," uses a multimodal language model to turn natural language prompts into sketches in a few seconds.
For example, it can doodle a cat either on its own or through collaboration, drawing w/a human or incorporating text-based input to sketch each part separately.
MIT CSAIL
@MIT_CSAIL
The researchers showed that SketchAgent can create abstract drawings of diverse concepts, like a robot, butterfly, DNA helix, flowchart, & even the Sydney Opera House.
MIT CSAIL
@MIT_CSAIL
SketchAgent leverages an off-the-shelf LLM to draw stroke-by-stroke w/o training on any data — instead, the researchers developed a "sketching language" in which a sketch is translated into a numbered sequence of strokes on a grid.
SketchAgent also provides textual annotations of the meaning of each stroke in the sketch.
MIT CSAIL
@MIT_CSAIL
One day, the tool could be expanded into an interactive art game that helps teachers and researchers diagram complex concepts or give users a quick drawing lesson.
MIT CSAIL
@MIT_CSAIL
More details about SketchAgent: yael-vinker.github.io/sketch-agent/
Authors: @YVinker, @TamarRottShaham, Kristine Zheng, AlexZhao, Judith E. Fan, & Antonio Torralba -
MIT CSAIL
@MIT_CSAIL
9 distance measures in data science w/algorithms (v/@MaartenGr).
-
MIT CSAIL
@MIT_CSAIL
A beginner's guide to data structures & algorithms: livebook.manning.com/book/grokking-algorithms-second-edit…
v/@swapnakpanda -
MIT CSAIL
@MIT_CSAIL
Napster was released #otd in 1999.
v/@JonErlichman & @MTVNEWS -
MIT CSAIL
@MIT_CSAIL
Just a couple of robots duking it out at ICRA earlier this month.
Video v/@Hongyu_Lii of Brown University -
MIT CSAIL
@MIT_CSAIL
What's the most underrated programming language?
-
MIT CSAIL
@MIT_CSAIL
GenAI is almost 5x less accurate than humans when summarizing scientific research, according to new study: royalsocietypublishing.org/doi/epdf/10.1098/rsos.241776
-
MIT CSAIL
@MIT_CSAIL
89 years ago today, Alan Turing's "Universal Machine" proposal ushered in the computer age: www.scientificamerican.com/blog/guest-blog/how-alan-turin…
-
MIT CSAIL
@MIT_CSAIL
Types of databases, v/@Eyowhite3.
-
MIT CSAIL
@MIT_CSAIL
The first digital image was scanned #otd in 1957. The late Russell Kirsch’s grayscale image of his newborn son launched the field of raster graphics: www.nist.gov/news-events/news/2007/05/fiftieth-anniversar…
-
MIT CSAIL
@MIT_CSAIL
MIT study reveals that vision-language models, which are often used to analyze medical images to streamline medical diagnosis, can’t handle queries w/negation words like "no" and "not": news.mit.edu/2025/study-shows-vision-language-models-cant…
-
MIT CSAIL
@MIT_CSAIL
Why programmers like cooking, v/@rkoutnik.
-
MIT CSAIL
@MIT_CSAIL
Lost papers from Alan Turing — including his signed 1938 PhD dissertation and seminal 1936 work "On Computable Numbers" — were nearly shredded.
They're now set for auction, expected to sell for as much as £150,000 (roughly $201,380): www.theguardian.com/science/2025/may/21/alan-turing-paper… -
MIT CSAIL
@MIT_CSAIL
In May 1967, two computer scientists presented a paper about Simula 67, the first programming language to include abstract data types and classes: twobithistory.org/2019/01/31/simula.html
-
MIT CSAIL
@MIT_CSAIL
Crucial AI terms to know, v/@hasantoxr.
-
MIT CSAIL
@MIT_CSAIL
"The art of programming is the art of organizing complexity." — Edsger Dijkstra
-
MIT CSAIL
@MIT_CSAIL
A tip about using "neural networks," v/@xkcd.
-
MIT CSAIL
@MIT_CSAIL
What’s a common misconception about machine learning that you wish more people understood?
-
MIT CSAIL
@MIT_CSAIL
Basic Linux commands.
-
MIT CSAIL
@MIT_CSAIL
A new AI model from MIT CSAIL researchers crafts smooth, high-quality videos in seconds: news.mit.edu/2025/causevid-hybrid-ai-model-crafts-smooth-…
Their "CausVid" system combines the photorealism of diffusion models w/the speed of autoregressive approaches. Take a peek at its most compelling visuals 🧵Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
One day, CausVid could help w/things like:
• Helping viewers understand a live stream in a different language by rapidly generating a video that syncs w/an audio translation ▶️
• Rendering new content in a video game 🎮
• Quickly producing training simulations to teach robots new tasks 🤖
MIT CSAIL
@MIT_CSAIL
More details about the system: causvid.github.io
Authors: @TianweiY, @qiangz_ai, @xunhuang1995, @rzhang88, @elishechtman, Bill Freeman, & @fredodurand -
MIT CSAIL reposted
Massachusetts Institute of Technology (MIT)
@MIT
“Periodic table of machine learning” could fuel AI discovery: Researchers have created a unifying framework that can help scientists combine existing ideas to improve AI models or create new ones. news.mit.edu/2025/machine-learning-periodic-table-could-f…
-
MIT CSAIL reposted
Massachusetts Institute of Technology (MIT)
@MIT
New research enables robots to learn about an object’s weight, softness, or contents by picking it up and gently shaking it. The robot can determine properties like an object’s mass in a matter of seconds, without the need for cameras or other external sensors.
-
MIT CSAIL
@MIT_CSAIL
A cheat sheet for key Python concepts, v/@clcoding.
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Edsger Dijkstra, whose algorithm solved the "single-source shortest path" problem: www.youtube.com/watch?v=gdmfOwyQlcI
Image v/Leo van Velzen -
MIT CSAIL reposted
Tianwei Yin
@TianweiY
Video diffusion models generate high-quality videos but are too slow for interactive applications.
We @MIT_CSAIL @AdobeResearch introduce CausVid, a fast autoregressive video diffusion model that starts playing the moment you hit "Generate"!
A thread 🧵 -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Caltech professor Katie Bouman! As a PhD student at MIT she was part of the team that helped create the first-ever image of a black hole.
-
MIT CSAIL
@MIT_CSAIL
A free guide to prompt engineering from Google: www.kaggle.com/whitepaper-prompt-engineering
-
MIT CSAIL
@MIT_CSAIL
AI can struggle to analyze long data sequences like climate trends & financial metrics.
Modeled after the brain's neural oscillations, MIT's new AI method makes stable & efficient predictions. It could aid w/forecasting tasks in health care, AVs, & more: news.mit.edu/2025/novel-ai-model-inspired-neural-dynamics… -
MIT CSAIL
@MIT_CSAIL
Happy birthday to MIT professor Ron Rivest, co-inventor of #rsa public-key encryption.
Image v/Cody O’Loughlin & Quanta Magazine -
MIT CSAIL
@MIT_CSAIL
#otd in 1974 TCP/IP was born, when Vint Cerf & Bob Kahn's paper describing a protocol for sending packets across networks was published: www.cs.princeton.edu/courses/archive/fall06/cos561/papers…
-
MIT CSAIL
@MIT_CSAIL
May the 4th be with you (v/@MIT).
-
MIT CSAIL
@MIT_CSAIL
When programming languages were released:
1957: Fortran
1958: LISP
1959: COBOL
1964: BASIC
1970: Pascal
1972: C
1979: SQL
1985: C++
1991: Python
1993: R
1995: PHP, Java & JS
2001: C#
2009: Go
2012: Julia
2014: Swift -
MIT CSAIL
@MIT_CSAIL
#otd in 1962 MIT released CTSS, the first functional time-sharing system that allowed multiple people to use a computer terminal at the same time: www.youtube.com/watch?v=Q07PhW5sCEk
Video v/@GBH & @MITMuseumThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
Happy birthday to #basic, the programming language launched at @Dartmouth #otd in 1964 to encourage non-STEM students to use computers: time.com/69316/basic/ (v/@TIME)
-
MIT CSAIL
@MIT_CSAIL
What's your favorite AI chatbot?
-
MIT CSAIL
@MIT_CSAIL
"Every great developer you know got there by solving problems they were unqualified to solve until they actually did it." — Patrick McKenzie
-
MIT CSAIL
@MIT_CSAIL
Mid-day finds: Watch as a TV station made the jump to color programming live.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
A new technique from MIT can make AI-generated code adhere to whatever programming language or other format is being used, while remaining error-free: news.mit.edu/2025/making-ai-generated-code-more-accurate-…
-
MIT CSAIL
@MIT_CSAIL
3D modeling tools dictate how many objects look — but what about how they feel?
MIT's "TactStyle" program enables creators to use image prompts to stylize the textures of 3D models. It can customize the tactile properties of accessories, decor, & more: news.mit.edu/2025/3d-modeling-you-can-feel-0422 -
MIT CSAIL
@MIT_CSAIL
TO: Ray@tomlinson.com
SUBJECT: Happy 84th!
Today would have been the 84th birthday of Ray Tomlinson, who sent the first email more than 50 years ago while working at BBN Technologies. -
MIT CSAIL
@MIT_CSAIL
60 years ago this month, the Fast Fourier Transform was created, a powerful tool for image compression & data analysis.
Watch a classic MIT breakdown of FFT, perhaps the most-taught algorithm at the Institute: www.youtube.com/watch?v=iTMn0Kt18tg
v/@MITOCW -
MIT CSAIL
@MIT_CSAIL
The CTRL+C experience.
-
MIT CSAIL
@MIT_CSAIL
Marijuana was the world's first online transaction, between MIT & Stanford students on ARPANET: gizmodo.com/remember-how-the-first-thing-ever-sold-online…
Image v/Grok -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Moore's Law!
60 years ago today, Gordon Moore published his article on cramming more components onto integrated circuits: www.monolithic3d.com/uploads/6/0/5/5/6055488/gordon_moore…
Image v/@ComputerHistory -
MIT CSAIL
@MIT_CSAIL
What’s your favorite AI coding tool?
-
MIT CSAIL reposted
Gabe Grand
@gabe_grand
Tackling complex problems with LMs requires search/planning, but how should test-time compute be structured?
Introducing Self-Steering, a new meta-reasoning framework where LMs coordinate their own inference procedures by writing code!This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
A free MIT course that uses Python to solve different puzzles: ocw.mit.edu/courses/6-s095-programming-for-the-puzzled-ja…
-
MIT CSAIL reposted
Angie Boggust
@angie_boggust
#chi2025 paper on human–AI alignment! 🧵
Models can learn the right concepts but still be wrong in how they relate them.
✨Abstraction Alignment✨evaluates whether models learn human-aligned conceptual relationships.
It reveals misalignments in LLMs💬 and medical datasets🏥. -
MIT CSAIL
@MIT_CSAIL
MIT's Stata Center, Ghibli-style!
v/ChatGPT -
MIT CSAIL
@MIT_CSAIL
Bot in shining armor: Leonardo’s robot, designed by da Vinci around the year 1495, was a humanoid knight that could sit, wave its arms, and even use its mouth.
The design later inspired NASA’s Anthrobot on Mars: www.bbc.com/future/article/20160727-leon
Photo v/Erik Möller -
MIT CSAIL
@MIT_CSAIL
#otd in 1980 a Japanese computer scientist published a paper proposing the “Neocognitron,” the neural net that directly inspired CNNs: www.rctn.org/bruno/public/papers/Fukushima1980.pdf
Kunihiko Fukushima’s paper explained back in 1986: www.youtube.com/watch?v=Qil4kmvm2Sw -
MIT CSAIL
@MIT_CSAIL
Can LLMs help us discover the right molecules for new medicines?
An MIT-led approach combines an LLM w/graph-based models to create new molecules w/specific properties. Users can request molecules in plain language & get a description of how to make it: news.mit.edu/2025/could-llms-help-design-our-next-medicin… -
MIT CSAIL
@MIT_CSAIL
A summary of how AI agents work, v/@sahnlam.
-
MIT CSAIL
@MIT_CSAIL
Wouldn’t it be great to save a copy of your personal items & the memories they hold?
MIT’s "InteRecon" program reconstructs these objects & their interactive features in mixed reality, bringing cherished belongings like toys & iPods to digital spaces: news.mit.edu/2025/interecon-brings-personal-items-to-mixe…This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
The 3 laws of robotics, and why they’re in that order.
v/@xkcdComic -
MIT CSAIL
@MIT_CSAIL
What a difference 70 years makes: 3.75 MB in 1956 & 1 TB today.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
"A good programmer is someone who always looks both ways before crossing a one-way street." — Doug Linder
-
MIT CSAIL
@MIT_CSAIL
A concise summary of how HTTPS works, v/@sahnlam.
-
MIT CSAIL
@MIT_CSAIL
What’s your favorite book about machine learning?
-
MIT CSAIL
@MIT_CSAIL
#otd in 1976 Apple was founded.
-
MIT CSAIL
@MIT_CSAIL
This MIT CS class teaches you things that all the other classes don't teach you, like...
🖥️ Shell tools and scripting
🖥️ Vim
🖥️ Data wrangling
🖥️ Command-line environment
🖥️ Version control
Watch all 11 lectures for free here: www.youtube.com/watch?v=Z56Jmr9Z34Q&list=PLyzOVJj3bHQuloK… -
MIT CSAIL
@MIT_CSAIL
"In programming, if someone tells you "you’re overcomplicating this," they’re either 10 steps behind you or 10 steps ahead of you." — Andrew Clark (credit: @CodeWisdom)
-
MIT CSAIL
@MIT_CSAIL
The world’s first laptop launched #otd in 1983. RadioShack’s TRS-80 Model 100 weighed 3 pounds with a 20-hour battery, a word processor, and a spreadsheet program.
shorturl.at/noGM4 -
MIT CSAIL
@MIT_CSAIL
Register-transfer level verification captures vulnerabilities to side-channel attacks, reducing the risk of hidden bugs in the chip design before shipping them to production.
A new method from CSAIL uses shadow logic to improve RTL verification scalability while being applicable to different defense mechanisms: www.csail.mit.edu/news/giving-verification-more-logic-and… -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Keras, the open-source neural-network library, which was launched 10 years ago today.
It's now one of the world’s ten most-used software tools: github.com/keras-team/keras -
MIT CSAIL
@MIT_CSAIL
20 years ago today MIT researchers got a computer-generated gibberish paper accepted to a predatory journal: www.csail.mit.edu/news/how-fake-paper-generator-tricked-s…
Generate your own here: pdos.csail.mit.edu/archive/scigen/ -
MIT CSAIL
@MIT_CSAIL
Translating graphics into tactile charts for blind & low-vision readers requires compliance w/extensive design rules.
MIT’s "Tactile Vega-Lite" tool hardwires such guidelines w/i its program, potentially streamlining the process for designers & teachers: news.mit.edu/2025/making-graphs-more-accessible-blind-low… -
MIT CSAIL
@MIT_CSAIL
An MIT introduction to computer science & Python programming: www.youtube.com/watch?v=xAcTmDO6NTI&list=PLUl4u3cNGP62A-y…
This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
The encoder of a transformer, as explained by @rfeers.
-
MIT CSAIL
@MIT_CSAIL
Reasons why people who work w/computers seem to have a lot of free time: imgur.com/gallery/D2j11jY
Image v/Evil Jaymz -
MIT CSAIL
@MIT_CSAIL
Happy birthday Joseph Fourier, whose 1822 equation allows us to listen to mp3s today: gizmodo.com/digital-music-couldnt-exist-without-the-fouri…
-
MIT CSAIL
@MIT_CSAIL
Cable-driven mechanisms enable streamlined, precise motion, but are difficult to assemble by hand.
MIT’s "Xstrings" software tool & 3D printing method automates how these objects are put together, allowing users to customize designs for things like:
- A lizard robot
- A wall sculpture
- A claw that can seamlessly grab items
Full video: www.youtube.com/watch?v=EMMsmuC-xFU -
MIT CSAIL
@MIT_CSAIL
Atlas can now breakdance — look out, @Olympics!
Video v/Boston Dynamics -
MIT CSAIL
@MIT_CSAIL
Born #otd in 1927: the late Allen Newell, who in 1955 — before the term “AI” existed — co-wrote what many consider the world’s first AI program.
"Logic Theorist" proved 38 of 52 classic math theorems & established the field of heuristic programing: history-computer.com/Library/Logic Theorist memorandum.pdf -
MIT CSAIL
@MIT_CSAIL
A free beginner's guide to data structures & algorithms: pressbooks.palni.org/anopenguidetodatastructuresandalgori…
v/@swapnakpanda -
MIT CSAIL
@MIT_CSAIL
Happy Pi Day! Here's the first 100,000 digits of π, visualized.
v/AvoidableBoat67 on Reddit -
MIT CSAIL
@MIT_CSAIL
A new programming language from MIT could enable high-performance coding that can compete w/state-of-the-art libraries w/a few hundred lines of code, instead of tens or hundreds of thousands.
Called "Exo 2," the approach enables reusable scheduling libraries external to compilers: news.mit.edu/2025/high-performance-computing-with-much-le… -
MIT CSAIL
@MIT_CSAIL
QS ranks MIT #1 in computer science, data science and AI, electrical and electronic engineering, & more: news.mit.edu/2025/qs-world-university-rankings-rates-mit-…
-
MIT CSAIL
@MIT_CSAIL
"The most effective debugging tool is still careful thought, coupled with judiciously placed print statements." — Brian Kernighan, co-creator of Unix
-
MIT CSAIL
@MIT_CSAIL
Access to the book "Introduction to Vectors and Tensors":
- Vol. 1: oaktrust.library.tamu.edu/server/api/core/bitstreams/0603…
- Vol. 2: oaktrust.library.tamu.edu/server/api/core/bitstreams/9c9b…
v/swapnakpanda on X
MIT CSAIL
@MIT_CSAIL
MIT CSAIL
@MIT_CSAIL
Links here:
- Vol. 1: oaktrust.library.tamu.edu/handle/1969.1/2502
- Vol. 2: oaktrust.library.tamu.edu/handle/1969.1/3609
MIT CSAIL
@MIT_CSAIL
Access to the book "Introduction to Vectors and Tensors":
- Vol. 1: oaktrust.library.tamu.edu/server/api/core/bitstreams/0603…
- Vol. 2: oaktrust.library.tamu.edu/server/api/core/bitstreams/9c9b…
v/swapnakpanda on X -
MIT CSAIL
@MIT_CSAIL
An SQL query, as visualized by @bytebytego.
-
MIT CSAIL
@MIT_CSAIL
Computer engineering 60 years ago.
v/@JonErlichman & the San Diego Air and Space Museum -
MIT CSAIL
@MIT_CSAIL
Links here:
- Vol. 1: oaktrust.library.tamu.edu/handle/1969.1/2502
- Vol. 2: oaktrust.library.tamu.edu/handle/1969.1/3609
MIT CSAIL
@MIT_CSAIL
Access to the book "Introduction to Vectors and Tensors":
- Vol. 1: oaktrust.library.tamu.edu/server/api/core/bitstreams/0603…
- Vol. 2: oaktrust.library.tamu.edu/server/api/core/bitstreams/9c9b…
v/swapnakpanda on X -
MIT CSAIL
@MIT_CSAIL
The software development life cycle.
-
MIT CSAIL
@MIT_CSAIL
Born #otd in 1917: Betty Holberton, who...
-programmed the ENIAC (w/5 other women)
-helped standardize COBOL & FORTRAN w/Grace Hopper
-co-developed UNIVAC, & BINAC's C-10 instruction set
-co-wrote 1st generative programming system
www.nytimes.com/2001/12/17/business/frances-e-holberton-8… -
MIT CSAIL reposted
Massachusetts Institute of Technology (MIT)
@MIT
Quantum computing is advancing fast! Researchers are developing more efficient quantum circuits that could crack encryption, posing a major cybersecurity risk. mitsha.re/1vit50VbVHv
-
MIT CSAIL
@MIT_CSAIL
An introductory course about deep learning, courtesy of MIT: consent.youtube.com/m?continue=https%3A%2F%2Fwww.youtube.…
v/@MITDeepLearning -
MIT CSAIL
@MIT_CSAIL
Over the past year, MIT CSAIL researchers have made headway in improving how robots perceive their surroundings, plan their movements, and execute tasks.
Take a look at 5 recent highlights 🧵Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
5. Training general-purpose robots faster 🤖
The "Heterogeneous Pretrained Transformers" technique merges vast, diverse data to help train robots on many tasks w/o starting from scratch. This faster, cost-effective method outperformed traditional training by >20% in real-world tests.
MIT CSAIL
@MIT_CSAIL
These projects, among many others, illustrate how CSAIL is helping implement intelligent robots in everyday life.
Follow us for more updates on our lab’s latest research! -
MIT CSAIL
@MIT_CSAIL
What was your first handheld device (iPod, Walkman, etc)?
-
MIT CSAIL
@MIT_CSAIL
The late Frances Spence (right) was born #otd in 1922.
She was a trailblazer in computer programming as one of the first programmers of ENIAC, the first general-purpose electronic computer. -
MIT CSAIL
@MIT_CSAIL
MIT’s “Introduction to Algorithms,” published #otd in 1990, is the world’s most cited CS text, with 67K citations & over a million copies sold.
mitpress.mit.edu/9780262046305/introduction-to-algorithms/
@mitpress -
MIT CSAIL
@MIT_CSAIL
40 years ago vs. now: MIT's Building 20 (foreground) & its replacement, the Stata Center.
-
MIT CSAIL
@MIT_CSAIL
RIP to Skype.
The internet calling app will be officially phased out this May: www.cnn.com/2025/02/28/tech/skype-microsoft-shutdown/inde…
Image v/Reuters -
MIT CSAIL
@MIT_CSAIL
What are today's robots truly capable of?
AI scientist & "Love is Blind" star @camrhamilton came to MIT CSAIL to find out: www.youtube.com/watch?v=5AMWCXQmDBU -
MIT CSAIL
@MIT_CSAIL
Where do you think robotics will be in ten years?
-
MIT CSAIL
@MIT_CSAIL
How you know you're getting older: the "dragon book" was published nearly forty years ago.
www.reddit.com/r/programming/comments/67v0h/ask_reddit_ho… -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Steve Jobs, who would've turned 70 today.
Some of his best quotes: www.wsj.com/articles/BL-DGB-23002
Image v/Ted Thai, The LIFE Picture Collection, & Getty Images -
MIT CSAIL
@MIT_CSAIL
Valuations in 2009:
BlackBerry: $28 billion
Apple: $79 billion
Today:
BlackBerry: $3 billion
Apple: $3.6 trillion
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
This month, “hello world” said “hello world!”
The term was coined in a seminal programming book published in 1978: “C Programming Language,” written by Brian Kernighan and Dennis Ritchie: go.aws/37TmOZq -
MIT CSAIL
@MIT_CSAIL
MIT assistant professor Andreea Bobu discusses what inspired her to become a scientist focused on collaborative learning & autonomy.
Full video: www.youtube.com/watch?v=0JXszb1oxkk -
MIT CSAIL
@MIT_CSAIL
Visualizing scalar, vector, matrix, & higher-order tensors, v/@CalcInsights_.
-
MIT CSAIL
@MIT_CSAIL
"Anyone who has lost track of time when using a computer knows the propensity to dream, the urge to make dreams come true, & the tendency to miss lunch." — Tim Berners-Lee
-
MIT CSAIL
@MIT_CSAIL
"The most important numerical algorithm of our lifetime." — Gilbert Strang
(credit: @gabrielpeyre) -
MIT CSAIL
@MIT_CSAIL
Happy Valentine's Day from CSAIL, w/images from Midjourney.
-
MIT CSAIL
@MIT_CSAIL
Address space layout randomization (ASLR) scatters program instructions in a computer’s memory to prevent leaking precious data.
The problem: Hackers use microarchitectural side attacks to locate them. They can then use code-reuse attacks to steal passwords & alter systems — but an MIT CSAIL system might give ASLR the security boost it needs 🧵Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
The researchers found that Oreo could prevent microarchitectural side attacks on the Linux operating system w/o hampering the software it protects.
Their experiment involved simulating hardware attacks in gem5, a platform commonly used to study computer architecture.
MIT CSAIL
@MIT_CSAIL
Oreo is a lightweight security upgrade, & next, the team will look to address speculative execution attacks.
Read more: news.mit.edu/2025/to-keep-hardware-safe-cut-out-codes-clu…
Authors: Shixin Song (@zhizhuoxin), Joseph Zhang, & Mengjia Yan -
MIT CSAIL reposted
Aleksander Madry
@aleks_madry
Do current LLMs perform simple tasks (e.g., grade school math) reliably?
We know they don't (is 9.9 larger than 9.11?), but why?
Turns out that, for one reason, benchmarks are too noisy to pinpoint such lingering failures.
w/ @josh_vendrow @EdwardVendrow @sarameghanbeery
1/5 -
MIT CSAIL
@MIT_CSAIL
Read the lost thesis of Dennis Ritchie, creator of the C programming language & co-creator of Unix: archive.computerhistory.org/resources/access/text/2020/05…
Ritchie never got his PhD b/c he didn't want to pay Harvard the thesis binding fee.
(v/@IEEESpectrum) -
MIT CSAIL reposted
Boyuan Chen
@BoyuanChen0
Announcing Diffusion Forcing Transformer (DFoT), our new video diffusion algorithm that generates ultra-long videos of 800+ frames. DFoT enables History Guidance, a simple add-on to any existing video diffusion models for a quality boost. Website: boyuan.space/history-guidance (1/7)
-
MIT CSAIL
@MIT_CSAIL
New MIT faculty member & CSAIL principal investigator Kaiming He discusses AI’s role in lowering barriers between scientific fields & fostering collaboration across scientific disciplines.
“There is no way I could ever understand high-energy physics, chemistry, or the frontier of biology research, but now we are seeing something that can help us to break these walls, and that is the creation of a common language that has been found in AI”: news.mit.edu/2025/creating-common-language-kaiming-he-0207 -
MIT CSAIL
@MIT_CSAIL
How math was written before LaTeX, v/@Anthony_Bonato.
-
MIT CSAIL
@MIT_CSAIL
MIT professor on the tradeoffs of equipping AI w/human-like intelligence.
Part 2 of the full video: www.youtube.com/watch?v=GOu_SshPnXU -
MIT CSAIL
@MIT_CSAIL
"Students shouldn't go out into life w/o the ability to communicate.
Your success in life will be determined largely by...
-your ability to speak,
-your ability to write, &
-the quality of your ideas
**in that order**."
— MIT Prof. Patrick Winston, who would've turned 82 today -
MIT CSAIL
@MIT_CSAIL
Happy birthday to programming-language pioneer Ken Thompson.
💻 designer of Unix
💻 inventor of B (predecessor to C)
💻 co-inventor of Google's Go
www.youtube.com/watch?v=EY6q5dv_B-o -
MIT CSAIL
@MIT_CSAIL
What's the one generative Al tool you can't live without?
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Python creator Guido van Rossum. The open source language was named after comedy troupe Monty Python: www.hackster.io/news/the-ieee-computer-society-crowns-pyt…
AI image v/Grok -
MIT CSAIL
@MIT_CSAIL
MIT CSAIL PhD students weigh in on their favorite AI programs.
Full video: www.youtube.com/watch?v=ApbQ6GKNn5Y -
MIT CSAIL
@MIT_CSAIL
Thoughts on DeepSeek-R1 so far?
-
MIT CSAIL
@MIT_CSAIL
This repository on ML operations has free talks, books, papers and more: github.com/visenger/awesome-mlops
Developed by @visenger -
MIT CSAIL
@MIT_CSAIL
"An expert is a person who has made all the mistakes that can be made in a very narrow field." — Niels Bohr
-
MIT CSAIL
@MIT_CSAIL
An HP ad from 1968: "Is it possible to get a really good computer for less than $10,000?"
v/HP Computer Museum -
MIT CSAIL
@MIT_CSAIL
MIT professor on how to tell if an image is AI-generated.
Part 1 of the full video: www.youtube.com/watch?v=53kYaFx7nHs -
MIT CSAIL reposted
Massachusetts Institute of Technology (MIT)
@MIT
Developing and deploying powerful generative AI models comes with significant environmental consequences. “It is not just the electricity you consume when you plug the computer in. There are much broader consequences,” Elsa Olivetti says. mitsha.re/SSxH50UIvWz
-
MIT CSAIL
@MIT_CSAIL
MIT CSAIL’s new generative model can efficiently create molecular dynamics videos 🧵
Their "MDGen" system emulates molecular dynamics, connecting static structures. It's an early step toward designing molecules w/desired dynamics.Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
By hitting the "play button" on molecules, future versions of the tool could help chemists design new molecules & closely study how their drug prototypes for cancer & other diseases would interact w/the molecular structure it intends to impact.
MIT CSAIL
@MIT_CSAIL
Read more here: www.csail.mit.edu/news/toward-video-generative-models-mol…
Authors: Bowen Jing, Hannes Stärk (@HannesStaerk), Tommi Jaakkola, & Bonnie Berger (@lab_berger) -
MIT CSAIL
@MIT_CSAIL
The early days of Amazon.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
The BlackBerry turns 26 today.
Image: Al Seib/LA Times via Getty Images -
MIT CSAIL
@MIT_CSAIL
What was your first programming language and what do you remember most about learning it?
-
MIT CSAIL reposted
Dr. Peter Slattery
@PeterSlattery1
📢 The MIT AI Risk Repository (@MITAIRisk) was selected by the @ParisPeaceForum for showcasing at the AI Action Summit! Come and see our booth on February 10 👉 parispeaceforum.org/projects-ai-action-summit/. Visit airisk.mit.edu #aiactionsummit
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Wikipedia, which launched #otd in 2001.
Their homepage has changed a bit in the past 24 years. -
MIT CSAIL
@MIT_CSAIL
There's no place like home.
v/@PR0GRAMMERHUM0R -
MIT CSAIL
@MIT_CSAIL
5.52 billion internet users and counting today.
Photo v/@JonErlichman, stat v/Meltwater -
MIT CSAIL
@MIT_CSAIL
"The purpose of abstraction is not to be vague, but to create a new semantic level in which one can be absolutely precise." — Edsger Dijkstra
-
MIT CSAIL
@MIT_CSAIL
An ad from General Motors in 1969.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
A new printmaking technique produces images that can change colors when heated up ♨️
MIT researchers helped develop "Thermochromorph" to take advantage of full-color process printing. It uses thermochromic inks to make images transition into each other: www.csail.mit.edu/news/images-magically-transform-through…Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
Thermochromorph does have some potential limitations:
1. Image resolution is constrained by the smallest dot size that the team’s laser cutter can engrave, which could potentially be offset by screen printing.
2. For print quality, the pigments are not entirely invisible in their clear states, which means that the clarity of the transitions depends on how thickly the ink layers were applied during printmaking. Future iterations will explore different image-processing techniques to modify the overlay of halftone patterns for the hot & cold images to reduce these visual artifacts.
MIT CSAIL
@MIT_CSAIL
Authors: @tichaesque, Paris Myers (@medialab), Tianyu Yu, Juliana Covarrubias, Mackenzie Leake, & Stefanie Mueller
Paper: dl.acm.org/doi/10.1145/3680530.3695445
Full video: www.youtube.com/watch?v=z2DU1a8mANM -
MIT CSAIL
@MIT_CSAIL
62 years ago today: MIT PhD student Ivan Sutherland unveiled his “Sketchpad”, which pioneered the CAD technology used in today’s computers: ti.me/2rj5Q2f
-
MIT CSAIL
@MIT_CSAIL
If you could have dinner with any computer scientist, past or present, who would it be?
-
MIT CSAIL
@MIT_CSAIL
Boston Dynamics was officially spun out of MIT #otd in 1992.
A timeline of the company's evolution: www.youtube.com/watch?v=_EZQx87DyzM -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Intel co-founder Gordon Moore (of "Moore’s Law"). About 60 years ago, he predicted that the number of transistors on computer chips would double roughly every two years.
Image: Intel -
MIT CSAIL
@MIT_CSAIL
Deep Learning job interview questions, fully solved & covering a wide range of key AI topics: arxiv.org/abs/2201.00650
credit: @papers_daily -
MIT CSAIL
@MIT_CSAIL
What are your goals as a computer scientist this year?
-
MIT CSAIL
@MIT_CSAIL
</2024>
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Bjarne Stroustrup, inventor of C++: bigthink.com/videos/big-think-interview-with-bjarne-strou…
-
MIT CSAIL reposted
Zengyi Qin
@qinzytech
DeepSeek: 200 people + $5.5M cost beats Llama3.1-405B
JetMoE: 4 people + $0.08M cost beats Llama2-7B and Llama-13B
@deepseek_ai @AIatMeta @Alibaba_Qwen @xai @ZihangDai @tingchenai @Guodzh @TheGregYang @elonmusk give us JetMoE team $6M and see what we can do -
MIT CSAIL
@MIT_CSAIL
A Python cheat sheet.
v/@QCompounding -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Linus Torvalds, the father of #linux: www.zdnet.com/article/linus-torvalds-talks-ai-rust-adopti…
Image via Steve Double/Camera Press/Redux -
MIT CSAIL
@MIT_CSAIL
The 25 most important mathematical definitions in data science, v/@DailyDoseOfDS_.
-
MIT CSAIL
@MIT_CSAIL
"Beware of bugs in the above code; I have only proved it correct, not tried it." — Donald Knuth, Turing winner & author of "The Art of Computer Programming"
-
MIT CSAIL
@MIT_CSAIL
Merry Christmas to all our fellow mathematicians! (v/@DiegoKuonen)
-
MIT CSAIL
@MIT_CSAIL
May I interest you in some empty object cookies?
v/@jeffistyping -
MIT CSAIL
@MIT_CSAIL
How robots prepare for the holidays, according to Midjourney.
-
MIT CSAIL
@MIT_CSAIL
Second nature to developers, but a source of unfamiliarity to many others.
v/@xkcd -
MIT CSAIL
@MIT_CSAIL
Which LLMs hallucinate the most? The least?
A leaderboard for how often these models produce hallucinations when summarizing a document: github.com/vectara/hallucination-leaderboard -
MIT CSAIL reposted
Sharut Gupta
@sharut_gupta
Excited to introduce Contextual Self-Supervised Learning (ContextSSL) for vision, an approach to learning a general representation that can adapt to be invariant or equivariant to task-specific symmetries given a few examples as the context.
Paper: arxiv.org/pdf/2405.18193 -
MIT CSAIL
@MIT_CSAIL
MIT professor Hari Balakrishnan on how he's used Al predictions to make roads safer from accidents: www.youtube.com/watch?v=jR9Sk4cLQT4
-
MIT CSAIL reposted
Zhou Xian
@zhou_xian_
Everything you love about generative models — now powered by real physics!
Announcing the Genesis project — after a 24-month large-scale research collaboration involving over 20 research labs — a generative physics engine able to generate 4D dynamical worlds powered by a physics simulation platform designed for general-purpose robotics and physical AI applications.
Genesis's physics engine is developed in pure Python, while being 10-80x faster than existing GPU-accelerated stacks like Isaac Gym and MJX. It delivers a simulation speed ~430,000 faster than in real-time, and takes only 26 seconds to train a robotic locomotion policy transferrable to the real world on a single RTX4090 (see tutorial: genesis-world.readthedocs.io/en/latest/user_guide/getting…).
The Genesis physics engine and simulation platform is fully open source at github.com/Genesis-Embodied-AI/Genesis. We'll gradually roll out access to our generative framework in the near future.
Genesis implements a unified simulation framework all from scratch, integrating a wide spectrum of state-of-the-art physics solvers, allowing simulation of the whole physical world in a virtual realm with the highest realism.
We aim to build a universal data engine that leverages an upper-level generative framework to autonomously create physical worlds, together with various modes of data, including environments, camera motions, robotic task proposals, reward functions, robot policies, character motions, fully interactive 3D scenes, open-world articulated assets, and more, aiming towards fully automated data generation for robotics, physical AI and other applications.
Open Source Code: github.com/Genesis-Embodied-AI/Genesis
Project webpage: genesis-embodied-ai.github.io
Documentation: genesis-world.readthedocs.io
1/nThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
Could multimodal vision language models (VLMs) help biodiversity researchers retrieve images for their studies? 🤔
MIT CSAIL, @ucl, @inaturalist, @EdinburghUni, & @UMassAmherst researchers designed a performance test to find out. Each VLM’s task: Locate & reorganize the most relevant results w/i the team’s "INQUIRE" dataset, composed of five million wildlife pictures & 250 search prompts from ecologists & other biodiversity experts: www.csail.mit.edu/news/ecologists-find-computer-vision-mo…Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
What makes the INQUIRE dataset so challenging?
The answer: its domain specificity. INQUIRE includes search queries based on discussions w/ecologists, biologists, oceanographers, & other experts about the types of images they’d look for, including animals’ unique physical conditions & behaviors.
MIT CSAIL
@MIT_CSAIL
The researchers are working w/iNaturalist to develop a query system to better help scientists & other curious minds find the images they actually want to see. Their working demo allows users to filter searches by species, enabling quicker discovery of relevant results like, say, the diverse eye colors of cats
Try it for yourself here: inquire-demo.csail.mit.edu
MIT CSAIL
@MIT_CSAIL
Authors: @EdwardVendrow, @panom1, Alex Shepard, @GBrostow, @ProfKateJones, @oisinmacaodha, @sarameghanbeery, & @Grant_Van_Horn
Paper: arxiv.org/abs/2411.02537
Related thread:
Eddie Vendrow
@EdwardVendrow
🎯 How can we empower scientific discovery in millions of nature photos?
Introducing INQUIRE: A benchmark testing if AI vision-language models can help scientists find biodiversity patterns- from disease symptoms to rare behaviors- hidden in vast image collections.
Thread👇🧵 -
MIT CSAIL
@MIT_CSAIL
This is what one "byte" of RAM looked like in 1946.
-
MIT CSAIL
@MIT_CSAIL
A Git command cheat sheet, v/@Prathkum.
-
MIT CSAIL
@MIT_CSAIL
What’s an invention you feel deserves more credit?
-
MIT CSAIL
@MIT_CSAIL
The full version of OpenAI’s o1 reasoning model has been out for a week. Thoughts so far?
-
MIT CSAIL
@MIT_CSAIL
Can we teach a robot its limits to do chores safely & correctly? 🧵
To help robots execute open-ended, multi-step tasks, MIT CSAIL researchers used vision models to see what’s near the machine & model its constraints. An LLM sketches up a plan that’s checked in a simulator to ensure it’s safe & realistic, potentially aiding household robots: www.csail.mit.edu/news/teaching-robot-its-limits-complete…
MIT CSAIL
@MIT_CSAIL
Their trial-&-error method is called "Planning for Robots via Code for Continuous Constraint Satisfaction" (PRoC3S).
It tests long-horizon plans to ensure they satisfy all constraints & enables a robot to perform diverse tasks, like writing individual letters, drawing a star, & sorting & placing blocks in different positions.
MIT CSAIL
@MIT_CSAIL
The researchers’ method uses an LLM pre-trained on text from across the Internet. Before asking PRoC3S to do a task, the team provided their language model w/a sample task (like drawing a square) that’s related to the target one (drawing a star).
The sample task includes a description of the activity, a long-horizon plan, & relevant details about the robot’s environment.
MIT CSAIL
@MIT_CSAIL
In simulations, PRoC3S successfully drew stars & letters 8/10 times each. It also could stack digital blocks in pyramids & lines, & place items w/accuracy, like fruits on a plate.
Across each of these digital demos, the CSAIL method completed the requested task more consistently than comparable approaches like "Code as Policies."
MIT CSAIL
@MIT_CSAIL
Next, the CSAIL engineers brought their approach to the real world. Their method developed & executed plans on a robotic arm, teaching it to put blocks in straight lines.
PRoC3S also enabled the machine to place blue & red blocks into matching bowls & move all objects near the center of a table.
MIT CSAIL
@MIT_CSAIL
In the future, PRoC3S could help robots complete more intricate chores in dynamic environments like houses, where they may be prompted to do a general chore composed of many steps (like "make me breakfast").
MIT CSAIL
@MIT_CSAIL
For future work, the researchers aim to improve results using a more advanced physics simulator, & to expand to more elaborate longer-horizon tasks via more scalable data-search techniques.
They also plan to apply PRoC3S to mobile robots (such as quadrupeds) for tasks that include walking & scanning surroundings.
MIT CSAIL
@MIT_CSAIL
Authors: Aidan Curtis (@AidanCurtis3), Nishanth Kumar (@nishanthkumar23), Jing Cao, Tomás Lozano-Pérez, & Leslie Pack Kaelbling (@MIT_LISLab)
Website: aidan-curtis.github.io/proc3s.github.io/
Full video: www.youtube.com/watch?v=hBq0EeFSsPo -
MIT CSAIL
@MIT_CSAIL
How can we really know if a chatbot is giving a reliable answer? 🧵
MIT CSAIL’s "ContextCite" tool can ID the parts of external context used to generate any particular statement from a language model, improving trust by helping users easily verify the statement:Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
The current model requires multiple inference passes, but the team is working to streamline this process to make detailed citations available on demand.
Another challenge is the inherent complexity of language. Some sentences in a given context are deeply interconnected & removing one might distort the meaning of others.
MIT CSAIL
@MIT_CSAIL
Authors: Ben Cohen-Wang (@bcohenwang), Harshay Shah (@harshays_), Kristian Georgiev (@kris_georgiev1), & Aleksander Mądry (@aleks_madry)
Demo & Paper: gradientscience.org/contextcite/
Related thread:
Aleksander Madry
@aleks_madry
How is an LLM actually using the info given to it in its context? Is it misinterpreting anything or making things up?
Introducing ContextCite: a simple method for attributing LLM responses back to the context: gradientscience.org/contextcite
w/ @bcohenwang, @harshays_, @kris_georgiev1 -
MIT CSAIL
@MIT_CSAIL
Born #otd in 1815: Ada Lovelace, the woman often credited with creating "the world's first program": www.scientificamerican.com/article/ada-lovelaces-180-year…
-
MIT CSAIL
@MIT_CSAIL
A breakdown of the simplest neural network from the late MIT Prof. Patrick Winston during his 6.034 (Artificial Intelligence) course.
Clip v/@7etsuo -
MIT CSAIL
@MIT_CSAIL
"Learning to break down problems into smaller pieces is one of the most important skills in computer science/life." — Addy Osmani (@addyosmani)
-
MIT CSAIL
@MIT_CSAIL
Decades of computing innovation: looking back on over 20 years of MIT CSAIL & 60 years of Project MAC.
Full video: www.youtube.com/watch?v=-woLSTt7dl8 -
MIT CSAIL
@MIT_CSAIL
Happy birthday to Geoffrey Hinton, whose work in neural networks has led many to call him the "Godfather of AI."
These pioneering efforts recently won him the Nobel Prize: www.nytimes.com/2024/10/08/technology/nobel-prize-geoffre…
Image v/Aaron Vincent Elkaim & @nytimes -
MIT CSAIL
@MIT_CSAIL
Perhaps you’ve imitated the sound of a cow or a bell before — but did you know you’re also "sonically sketching" what you heard? 🧵
Inspired by the cognitive science of how we communicate, MIT CSAIL researchers have developed an AI system that can produce these human-like vocal imitations w/no training & w/o ever having "heard" a human vocal impression before: www.csail.mit.edu/news/teaching-ai-communicate-sounds-hum…This media couldn't be displayed. See it on the original post.
MIT CSAIL
@MIT_CSAIL
The researchers engineered their system to produce & interpret sounds much like we do.
They first built a model of the human vocal tract that simulates how vibrations from the voice box are shaped by the throat, tongue, & lips. Then, they used a cognitively-inspired AI algorithm to control this vocal tract model & make it produce imitations in context-specific ways.
MIT CSAIL
@MIT_CSAIL
The model can effectively take many sounds from the world & generate human-like imitations of them — including noises like leaves rustling, a snake’s hiss, & an approaching ambulance siren.
Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
The team still has work to do w/the current iteration of their model: it struggles w/some consonants, like "z," which led to inaccurate impressions of sounds like bees buzzing.
They also can’t yet replicate how humans imitate speech, music, or sounds that are imitated differently across different languages, like a heartbeat.
MIT CSAIL
@MIT_CSAIL
Authors: Matt Caren, Kartik Chandra (@_k_a_c_h_), Joshua B. Tenenbaum, Jonathan Ragan-Kelley (@jrk), & Karima Ma
Paper: arxiv.org/abs/2409.13507
Full video: www.youtube.com/watch?v=O1EXc8kN3aI
Listen to more sounds from their model: matthewcaren.github.io/vocal-imitation/ -
MIT CSAIL
@MIT_CSAIL
Ask us your questions about AI reasoning, animal communication systems, & intelligent agents. We’ll pick some to feature in an upcoming explainer w/MIT CSAIL PhD student Pratyusha Sharma (@pratyusha_PS).
For more on her work: pratyushasharma.github.io -
MIT CSAIL
@MIT_CSAIL
MIT CSAIL Director & EECS Professor Daniela Rus was recently awarded the 2025 IEEE Edison Medal: www.csail.mit.edu/news/mit-csail-director-and-eecs-profes…
Joining electrical engineers including Alexander Graham Bell & Nikola Tesla, Rus received the prize for her leadership & pioneering work in robotics. -
MIT CSAIL
@MIT_CSAIL
Could AI help adapt our urban forests to climate change? 🧵
MIT CSAIL, @GoogleAI, & Purdue University scientists’ "Tree-D Fusion" system merges AI & tree-growth models w/Google's Auto Arborist data to create accurate, 3D urban trees. The project produced the first-ever large-scale database of 600,000 environmentally aware, simulation-ready tree models across North America: news.mit.edu/2024/advancing-urban-tree-monitoring-ai-powe…Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
Certain issues remain unsolved. Perhaps the most vexing is the "entangled tree problem," when neighboring trees grow into each other, their intertwined branches create a puzzle that no current AI system can fully unravel.
Now, the researchers are exploring applications beyond street view imagery, extending to platforms like iNaturalist & wildlife camera traps.
MIT CSAIL
@MIT_CSAIL
Authors: Jae Joong Lee (@_jaejoong_lee_), Bosheng Li, Sara Beery (@sarameghanbeery), Jonathan Huang, Songlin Fei, Raymond A. Yeh (@RaymondYeh), & Bedrich Benes
Paper: arxiv.org/abs/2407.10330 -
MIT CSAIL
@MIT_CSAIL
Netscape co-founder Marc Andreessen after the company went public in 1995.
v/@CNET -
MIT CSAIL
@MIT_CSAIL
65 years ago today, MIT spin-off DEC began shipping a minicomputer called the PDP-1.
Little did they know, it would be used to develop the first multiplayer video game Spacewar & usher in US hacker culture: www.computerhistory.org/pdp-1/introduction/
v/@ComputerHistory -
MIT CSAIL
@MIT_CSAIL
#otd in 1853 George Boole’s algebra & logic book introduced the operations “and”, “or” and “not” & set the groundwork for computers as we know them today. Free ebook here: www.gutenberg.org/files/15114/15114-pdf.pdf (v/@gutenberg_org)
-
MIT CSAIL
@MIT_CSAIL
34 years ago this month Tim Berners-Lee and Robert Cailliau emailed colleagues outlining a proposal for a hypertext project that they dubbed “WorldWideWeb.”
w3.org/Proposal.html -
MIT CSAIL
@MIT_CSAIL
An MIT CSAIL Thanksgiving, according to Midjourney.
-
MIT CSAIL reposted
Alex Albert
@alexalbert__
Introducing the Model Context Protocol (MCP)
An open standard we've been working on at Anthropic that solves a core challenge with LLM apps - connecting them to your data.
No more building custom integrations for every data source. MCP provides one protocol to connect them all: -
MIT CSAIL
@MIT_CSAIL
33 free online AI courses from @Mashable: mashable.com/article/free-ai-chatgpt-courses-november-2024
-
MIT CSAIL
@MIT_CSAIL
MIT professor Erik Demaine & artist-in-residence Martin Demaine discuss how they create unique origami sculptures from folded paper on the @PBS series "Craft in America."
The episode premieres on Dec. 27th at 9 & 10 PM depending on your area: www.youtube.com/user/craftinamerica?cbrd=1&ucbcb=1This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
Math inspires us all!
Happy #fibonacciday 11/23: www.livescience.com/37470-fibonacci-sequence.html -
MIT CSAIL
@MIT_CSAIL
LLMs rely heavily on transformers, which make use of an "attention mechanism" for modeling interactions among inputs. But this attention mechanism is inefficient and thus difficult to scale to longer contexts. 🧵
Because of this, researchers have been experimenting w/"linear transformers," which remove the nonlinearity in the attention mechanism & allow you to essentially restructure the computations more like a recurrent neural network (RNN).
MIT CSAIL
@MIT_CSAIL
Linear transformers have their own downsides, and there have been ongoing efforts in the community on improving their performance while maintaining training efficiency.
Recently, CSAIL researchers have proposed a novel parallel algorithm for efficiently training an expressive linear-transformer variant called "DeltaNet," which uses the classic delta rule as the recurrent update equation. This algorithm lets you scale DeltaNet to modern language modeling settings, where it is found to outperform other RNN-like models such as linear transformers and state-space models.
MIT CSAIL
@MIT_CSAIL
Read the full story on CSAIL News: www.csail.mit.edu/news/transforming-transformers
-
MIT CSAIL
@MIT_CSAIL
MIT professors on the expansive potential of generative AI, including language, code, & images: www.youtube.com/watch?v=ttxfgdmiuoI
-
MIT CSAIL
@MIT_CSAIL
MIT assistant professor Sara Beery on how her research group applies AI to monitor how species are changing around the world, including using tools like acoustics, sonar, & remote sensing data.
Full video: www.youtube.com/watch?v=ch6K00NORb8 -
MIT CSAIL
@MIT_CSAIL
MIT researchers recently developed "Boltz-1," which can predict the 3D structure of proteins, RNA, DNA, & small molecules at an accuracy comparable to AlphaFold 3.
A thread on the first commercially available, open-source model to achieve this feat ⬇️
Gabriele Corso
@GabriCorso
Thrilled to announce Boltz-1, the first open-source and commercially available model to achieve AlphaFold3-level accuracy on biomolecular structure prediction! An exciting collaboration with @jeremyWohlwend, @pas_saro and an amazing team at MIT and Genesis Therapeutics. A thread!
-
MIT CSAIL
@MIT_CSAIL
Take a peek inside an LLM, including all its layers & matrix operations: bbycroft.net/llm
v/@santiviquez & @BrendanBycroft -
MIT CSAIL
@MIT_CSAIL
What's a tech milestone you wish you could relive?
-
MIT CSAIL
@MIT_CSAIL
17 years later, the iPhone makes over $1 billion in revenue every two days.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
#otd in 1971 Intel unveiled the first single-chip CPU. 4 bits, 16 pins & 740KHz: spectrum.ieee.org/chip-hall-of-fame-intel-4004-microproce…
-
MIT CSAIL
@MIT_CSAIL
For roboticists, one challenge towers above the others: there isn’t enough data.
To accelerate the deployment of intelligent robots in the real world, MIT CSAIL’s "LucidSim" uses genAI & physics engines to create diverse & realistic virtual training grounds for robots. W/o any real-world data, their robot achieved expert-level performance in difficult tasks: www.csail.mit.edu/news/can-robots-learn-machine-dreamsThis media couldn't be displayed. See it on the original post.Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
LucidSim is paving the way toward a new generation of intelligent machines that elevate & amplify human effort. The researchers envision these machines could safely and autonomously learn how to navigate our complex world w/o ever setting foot in it.
The team’s immediate plan is to scale up to a human-sized robot and teach it to manipulate objects on the move, using purely synthesized images.
MIT CSAIL
@MIT_CSAIL
Authors: Alan Yu (@alany1_), Ge Yang (@EpisodeYang), Ran Choi, Yajvan Ravan, John Leonard (@jleonardmit), & Phillip Isola (@phillip_isola)
Paper: arxiv.org/abs/2411.00083
Full video: www.youtube.com/watch?v=fgnJQrvTj70
Website: lucidsim.github.io -
MIT CSAIL
@MIT_CSAIL
The Nobel Prize-winning tool AlphaFold 3 is now open-source: idp.nature.com/transit?redirect_uri=https%3A%2F%2Fwww.nat…
The AI model can accurately predict the structure & interactions of proteins, DNA, RNA, and more.
Image v/@GoogleDeepMind -
MIT CSAIL
@MIT_CSAIL
On July 4th, 1934, Marie Curie died of aplastic anaemia after years of radiation exposure.
Remarkably, even her lab notebook from 1899-1902 is still radioactive.
v/@NobelPrize -
MIT CSAIL
@MIT_CSAIL
Can we design a generative AI algorithm that works for any data type?
Together with a team at FAIR, CSAIL researchers propose a "unified theory of GenAI" called "Generator Matching" (GM). GM brings together a range of GenAI models, allowing researchers to build models for any data type (including multimodal) and w/completely new designs: arxiv.org/abs/2410.20587Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
Why does GM work? The trick is to make everything linear: Train generators, which are linear operators, to fulfill the Kolmogorov equation, a linear equation.
Then, marginalizing across data points is simple leading to training methods that work on large datasets.
MIT CSAIL
@MIT_CSAIL
Authors: Peter Holderrieth (@peholderrieth), Marton Havasi, Jason Yim, Neta Shaul, Itai Gat, Tommi Jaakkola, Brian Karrer, @RickyTQChen, & @lipmanya
Related thread:
Peter Holderrieth
@peholderrieth
New paper out!
We introduce “Generator Matching” (GM), a method to build GenAI models for any data type (incl. multimodal) with any Markov process. GM unifies a range of state-of-the-art models and enables new designs of generative models.
arxiv.org/abs/2410.20587
(1/5) -
MIT CSAIL
@MIT_CSAIL
Happy bday to MIT Prof. Barbara Liskov, Turing winner & programming pioneer: news.mit.edu/2009/turing-liskov-0310
-
MIT CSAIL
@MIT_CSAIL
In the field of reprogrammable surfaces (or items whose appearances we can digitally alter), users can redesign things like a wall or mug w/important information like health stats.
The portable light system & design tool "PortaChrome" uses UV & RGB LEDs to activate photochromic dye, reprogramming everyday objects like shirts. Its accompanying software can help users turn items into multicolor displays of fashion designs & health data: news.mit.edu/2024/portable-light-system-can-digitize-ever…Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
PortaChrome can quickly convert our surroundings into dynamic displays, but the researchers believe it could benefit from further speed boosts.
They'd like to use smaller LEDs, which would likely result in a surface that could be reprogrammed in seconds with a higher-resolution design due to increased light intensity.
MIT CSAIL
@MIT_CSAIL
Authors: Yunyi Zhu (@yunyizhu_), Cedric Honnet (@CedHon), Yixiao Kang , Junyi Zhu, Angelina J. Zheng, Kyle Heinz, Grace Tang, Luca Musk, Michael Wessely, & Stefanie Mueller
Paper: dl.acm.org/doi/10.1145/3654777.3676458
Additional Credits: University of California at Berkeley (@UCBerkeley) & Aarhus University (@AarhusUni_int)
Full video: www.youtube.com/watch?v=eQcwTSWOzR8 -
MIT CSAIL
@MIT_CSAIL
#otd in 1971 Bell Labs released the first Unix Programmer's Manual.
Download the free PDF here: dspinellis.github.io/unix-v3man/v3man.pdf
(v/@CoolSWEng) -
MIT CSAIL
@MIT_CSAIL
#otd in 1998 @ylecun submitted a paper on “gradient-based” deep learning for document recognition.
It took more than a decade before the world finally warmed to neural networks: www.semanticscholar.org/paper/Gradient-based-learning-app… -
MIT CSAIL
@MIT_CSAIL
Why can’t programmers tell the difference between Halloween & Christmas? Because oct 31 = dec 25.
#happyhalloween -
MIT CSAIL
@MIT_CSAIL
In scientific research, measurements that are too small to be quantified are usually labeled as "below the limit of quantification" & discarded.
MIT CSAIL researchers recently developed the "hypometric genetics" approach, which uses these typically disregarded measurements to improve genetic discovery up to 2.8x. Their method could impact personalized genomic medicine & drug development: www.csail.mit.edu/news/method-goes-below-limit-enhance-ge…
MIT CSAIL
@MIT_CSAIL
Authors: @manoliskellis & @yk_tani
Related thread: x.com/yk_tani/status/1848750077679997268
Additional credits: @broadinstitute, @MITEECS, @uk_biobank, @AJHGNews, & @GeneticsSociety
Yosuke Tanigawa
@yk_tani
I am excited to see our paper @AJHGNews. We introduce ‘hypometric genetics,’ an approach to investigate the genetic basis of binarized traits representing the presence of below-the-limit-of-quantification quality control indicators.
Paper📜: authors.elsevier.com/a/1jzoqgeXA9Lf
Thread: 🧵 -
MIT CSAIL
@MIT_CSAIL
#otd in 1991 Tim Berners-Lee first described the HTML mark-up language in a message to the WWW-talk mailing list: lists.w3.org/Archives/Public/www-talk/1991Oct/0003.html
-
MIT CSAIL
@MIT_CSAIL
55 years ago today, UCLA professor Leonard Kleinrock & his team sent the first online message: www.usatoday.com/story/tech/2019/10/29/first-internet-mes…
Funny enough, it was a typo: "LO" instead of the intended "LOGIN."
v/UCLA & Reuters -
MIT CSAIL
@MIT_CSAIL
14 sorting algorithms in just 60 seconds.
v/@7etsuo -
MIT CSAIL
@MIT_CSAIL
Linus Torvalds: fan of AI's potential, but not its hype.
He says AI is currently 90% marketing and 10% reality, and in five years, we'll see how useful it really is.
v/@tsarnick -
MIT CSAIL
@MIT_CSAIL
Another fall day flying by at CSAIL.
v/Mike Grimmett -
MIT CSAIL
@MIT_CSAIL
41 years ago today, Microsoft Word was released.
v/@JonErlichman -
MIT CSAIL
@MIT_CSAIL
Google's early employees began their work in a garage & made millions by the time they left.
A thread on the company's first 10 non-founding hires & where they are now 🧵
John Kim
@0xjohnkim
These folks gambled their careers on Larry & Sergei's search engine startup.
They wrote Google's first lines of code. Sold its first ads. Grew its first teams.
They started in a garage, and walked away with millions.
Where are Google's first 10 non-founding hires now? 👇 -
MIT CSAIL
@MIT_CSAIL
Much like Tetris, sustainable fashion design is a puzzle that involves configuring unique shapes in a way that fills up the available space 👚
To ensure that fashion designers leave no material to waste, MIT CSAIL, @AdobeResearch, & @UW’s "WasteBanned" application helps fashion designers visualize how to use every piece of fabric for their apparel. You can design, edit, & simulate how your eco-friendly clothing will look, then use that outline to bring your concepts to the real world: www.csail.mit.edu/news/zero-waste-design-tool-more-sustai…
MIT CSAIL
@MIT_CSAIL
It may seem easy to cut up garments & sew them together.
In reality, if you cut one piece, you’re also modifying others. This intertwined process, called linked editing, requires creativity to avoid creating waste. WasteBanned helps you visualize changes before you begin to cut to minimize environmental impact.
MIT CSAIL
@MIT_CSAIL
This CAD tool lets users enter the dimensions of their fabric & digitally cut, sew, move, measure, & label panels.
Users can then cut the panel into multiple edges, make a slit in the fabric, or split an edge by making a break point.
MIT CSAIL
@MIT_CSAIL
For example, you could outline a skirt by cutting three pieces & labeling them as the front, back right, & back left panels before digitally sewing them together.
You could also create a sweatshirt in 13 panels, marking each piece & then stitching accordingly.
MIT CSAIL
@MIT_CSAIL
The WasteBanned app also allows you to clear, undo, or redo any edits you’ve made, a notable difference from the irreversible changes we make in the real world.
Users can also add constraints, such as zero waste limits that ensure you use all your material, & user-added restrictions that limit the types of cuts you can make.
MIT CSAIL
@MIT_CSAIL
In a user study, participants observed that WasteBanned was easy & enjoyable to use.
One user (who studied fashion as an undergrad) even brought their digital designs to the real world, sewing together a bell sleeve tunic made of pink floral fabric from their outline.
MIT CSAIL
@MIT_CSAIL
While this software tool is a nice step toward "materials-first" design, the researchers would like to help users explore a set of equivalent-length seam curves that would produce well-fitting & stylish garments.
Moreover, the team is considering building a custom constraint solver, enabling curved cuts & seams w/different lengths.
MIT CSAIL
@MIT_CSAIL
Authors: Ruowang Zhang, Stefanie Mueller, Gilbert Louis Bernstein, Adriana Schulz (@AdrianaSchulz7), & Mackenzie Leake
Paper: dl.acm.org/doi/10.1145/3654777.3676395 -
MIT CSAIL
@MIT_CSAIL
Jeff Dean, Google's Chief Scientist, tells a student what it's like to be a computer scientist.
v/@JeffDean -
MIT CSAIL
@MIT_CSAIL
Web development timeline (v/@fireship_dev):
1990: HTML invented
1994: CSS invented to fix HTML
1995: JS invented to fix HTML/CSS
2006: jQuery invented to fix JS
2010: AngularJS invented to fix jQuery
2013: React invented to fix AngularJS
2014: Vue invented to fix React & Angular -
MIT CSAIL
@MIT_CSAIL
Take a trip inside MIT's Stata Center, home to many CSAIL researchers.
This media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
True story: a kid saw this and said, "oh, you 3D-printed the 'Save' Icon." twitter.com/Bill_Gross/status/920406104911233024
Credit: @Bill_Gross -
MIT CSAIL
@MIT_CSAIL
Sequence models have skyrocketed in popularity for their ability to analyze data & predict what to do next.
MIT’s "Diffusion Forcing" method combines the strengths of next-token prediction (like w/ChatGPT) & video diffusion (like w/Sora), training neural networks to handle corrupted data while predicting the next steps. This flexible, reliable sequence model helps produce higher-quality artificial videos and guides more precise decision-making for robots & AI agents: news.mit.edu/2024/combining-next-token-prediction-video-d…
MIT CSAIL
@MIT_CSAIL
By sorting through noisy data & reliably predicting the next steps in a task, Diffusion Forcing can aid a robot in ignoring visual distractions to complete manipulation tasks.
It can also generate stable & consistent video sequences and even guide an AI agent through digital mazes. In the future, it could enable household and factory robots to generalize to new tasks and improve AI-generated entertainment.
MIT CSAIL
@MIT_CSAIL
To combine the strengths of both models, CSAIL researchers created a sequence model training technique called "Diffusion Forcing."
The name comes from "Teacher Forcing," the conventional training scheme that breaks down full sequence generation into the smaller, easier steps of next-token generation.
MIT CSAIL
@MIT_CSAIL
Diffusion Forcing found common ground between diffusion models & teacher forcing: they both use training schemes that involve predicting masked (noisy) tokens from unmasked ones.
In the case of diffusion models, they gradually add noise to data, which can be viewed as fractional masking.
MIT CSAIL
@MIT_CSAIL
Diffusion Forcing trains neural networks to cleanse a collection of tokens, removing different amounts of noise w/i each one while simultaneously predicting the next few tokens.
MIT CSAIL
@MIT_CSAIL
When implemented into a robotic arm, for example, it helped swap two toy fruits across three circular mats, a minimal example of a family of long-horizon tasks that require memories.
Despite starting from random positions and seeing distractions like a shopping bag blocking the markers, it placed the objects into its target spots.
MIT CSAIL
@MIT_CSAIL
To generate videos, they trained Diffusion Forcing on Minecraft gameplay & colorful digital environments created w/i Google's DeepMind Lab Simulator.
When given a single frame of footage, the method produced more stable, higher-resolution videos than comparable baselines, including a Sora-like full-sequence diffusion model and ChatGPT-like next-token models.
MIT CSAIL
@MIT_CSAIL
Across each demo, Diffusion Forcing acted as a full sequence model, a next-token prediction model, or both.
According to the researchers, this versatile approach could potentially serve as a powerful backbone for a "world model" one day.
MIT CSAIL
@MIT_CSAIL
Authors: Boyuan Chen (@BoyuanChen0), Diego Martí Monsó, Yilun Du (@du_yilun), Max Simchowitz (@max_simchowitz), Russ Tedrake (@RussTedrake), Vincent Sitzmann (@vincesitzmann)
Paper: www.boyuan.space/diffusion-forcing/
Full video: www.youtube.com/watch?v=m4rhzR-h-JU -
MIT CSAIL
@MIT_CSAIL
MIT Prof. on how AI is impacting chip design & the future of Moore’s Law.
Full video: www.youtube.com/watch?v=w-3cFirlVqI -
MIT CSAIL
@MIT_CSAIL
Some tips & tricks for reducing how much memory you use when loading larger LLMs & other models in PyTorch: github.com/rasbt/LLMs-from-scratch/blob/main/ch05/08_memo…
v/@rasbt -
MIT CSAIL
@MIT_CSAIL
"The best learners are the people who push through the discomfort of being objectively bad at something." — Tommy Collison
-
MIT CSAIL
@MIT_CSAIL
Perhaps the best/only way to explain neural networks to a five-year-old: www.reddit.com/r/explainlikeimfive/comments/7buzbs/eli5_w… #machinelearning #eli5
(v/kouhoutek on @Reddit) -
MIT CSAIL
@MIT_CSAIL
MIT Media Lab's λ-2D project, which uses ideas from lambda calculus to explore coding via drawing: www.media.mit.edu/projects/2d-an-exploration-of-drawing-a…
v/@_DragonHead_ -
MIT CSAIL
@MIT_CSAIL
Published #otd in 1986: a seminal 4-page paper on backpropagation from Geoff Hinton that helped reignite the field of neural networks.
towardsdatascience.com/learning-backpropagation-from-geof… -
MIT CSAIL
@MIT_CSAIL
MIT professor on whether LLMs really understand complex financial concepts.
Part 2 of the full video: www.youtube.com/watch?v=7pOom_qlAH4 -
MIT CSAIL
@MIT_CSAIL
In honor of #adalovelaceday, learn about the woman who many view as the world's first programmer: cnet.co/2dZ1Ihp
-
MIT CSAIL
@MIT_CSAIL
Geoffrey Hinton, whose work in neural networks has led many to call him the “Godfather of AI”, co-wins the Nobel Physics Prize for AI research: www.nytimes.com/2024/10/08/science/nobel-prize-physics.html
-
MIT CSAIL reposted
Marc Andreessen 🇺🇸
@pmarca
Dear LazyX: What are the best X accounts or web sites that track all of the important AI papers and talks? 👊🤜
-
MIT CSAIL
@MIT_CSAIL
#otd in 1997, some advice that aged like batteries:
"I’d shut [Apple] down and give the money back to shareholders" — Dell CEO Michael Dell: cnet.co/3ZL55lf (v/@CNET) #tdih -
MIT CSAIL
@MIT_CSAIL
#otd in 1991 the Linux kernel was publicly released by Helsinki grad student Linus Torvalds: spectrum.ieee.org/linux-at-25-why-it-flourished-while-oth…
-
MIT CSAIL
@MIT_CSAIL
What’s a programming book that changed how you code?
We'll start: Michael Sipser's "Introduction to the Theory of Computation." -
MIT CSAIL
@MIT_CSAIL
#otd in 1967 Larry Roberts presented his idea for an “ARPANet” for connecting multiple computers together across the United States.
Full paper: www.researchgate.net/publication/234826251_Multiple_Compu… -
MIT CSAIL
@MIT_CSAIL
What does seeing a face in the moon have to do w/AI identifying tigers?
MIT CSAIL has introduced a human-labeled dataset of these illusory faces in everyday objects (a phenomenon called "pareidolia"), revealing how humans & AI detect faces differently. The group also discovered a "Goldilocks zone" where pareidolia is most likely to occur: news.mit.edu/2024/ai-pareidolia-can-machines-spot-faces-i…
A thread on their key insights 🧵Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
The study could potentially help face detection systems by reducing false positives, which might impact self-driving cars, human-computer interaction, & robotics.
The dataset & models could also help areas like product design.
MIT CSAIL
@MIT_CSAIL
Authors include Mark Hamilton (@mhamilton723), Vasha DuTell (@VashaDutell), and William T. Freeman.
Website: mhamilton.net/facesinthings
Full video: www.youtube.com/watch?v=Cigr5xnQ51o -
MIT CSAIL reposted
Liquid AI
@liquidai
Today we introduce Liquid Foundation Models (LFMs) to the world with the first series of our Language LFMs: A 1B, 3B, and a 40B model. (/n)
-
MIT CSAIL
@MIT_CSAIL
The Turing Test was born #otd in 1950 with Alan Turing’s seminal paper "Computing Machinery and Intelligence."
academic.oup.com/mind/article/LIX/236/433/986238 -
MIT CSAIL
@MIT_CSAIL
How can we distribute data points more uniformly for multi-dimensional problems in robotics, finance, & computational science? 🧵
MIT CSAIL’s AI-powered method uses graph neural networks to do this, self-optimizing to avoid clumps & voids. This "Message-Passing Monte Carlo (MPMC)" approach significantly improves accuracy in simulations & computations: www.csail.mit.edu/news/how-ai-improving-simulations-smart…
PNAS Paper: www.pnas.org/doi/10.1073/pnas.2409913121?url_ver=Z39.88-2…Post not available here, view on X
MIT CSAIL
@MIT_CSAIL
In robotics, path & motion planning often rely on sampling-based algorithms, which guide robots through decision-making processes.
MPMC’s improved uniformity could lead to more efficient robotic navigation & real-time adaptations for things like AVs or drone technology.
MIT CSAIL
@MIT_CSAIL
Going forward, the team plans to make MPMC points even more accessible to everyone, addressing the current limitation of training a new GNN for every fixed number of points and dimensions.
Authors: T. Konstantin Rusch (@tk_rusch), Nathan Kirk, Michael M. Bronstein (@mmbronstein), Christiane Lemieux, and Daniela Rus -
MIT CSAIL reposted
Konstantin Rusch
@tk_rusch
Our Message-Passing Monte Carlo paper has now been published @PNASNews
GNNs meet low-discrepancy sampling to improve accuracy of simulations in science and engineering 🤖🧬
Wonderful collaboration w/ @mmbronstein, N. Kirk, C. Lemieux,
D. Rus @MIT_CSAIL
www.pnas.org/doi/10.1073/pnas.2409913121 -
MIT CSAIL
@MIT_CSAIL
In German, the word "Verschlimmbessern" means “to make something worse while repairing it.” In English, we call that "software development."
-
MIT CSAIL
@MIT_CSAIL
"The robots are coming. And that’s a good thing." — MIT prof. & CSAIL Director Daniela Rus
Rus & science writer Gregory Mone's "The Mind's Mirror: Risk & Reward in the Age of AI" shows how AI works, how to mitigate dangerous outcomes, & possible futures: mitpressbookstore.mit.edu/book/9781324079323 -
MIT CSAIL
@MIT_CSAIL
The seven stages of a project:
1. Denial
2. Anger
3. Bargaining
4. Depression
5. Acceptance
6. Saturday
7. Sunday
v/@engineers_feed -
MIT CSAIL
@MIT_CSAIL
RT if you were part of that tiny 1993 bar.
(v/@markrendle) -
MIT CSAIL
@MIT_CSAIL
All major neural networks, in one chart: towardsdatascience.com/the-mostly-complete-chart-of-neura…
v/The Asimov Institute -
MIT CSAIL
@MIT_CSAIL
An IBM slide from 1979.
-
MIT CSAIL
@MIT_CSAIL
70 years ago this month, researchers ran the first FORTRAN program, considered the beginning of general-purpose programming: www.informationweek.com/it-leadership/fortran-7-reasons-w…
Image v/@NASA -
MIT CSAIL
@MIT_CSAIL
22 years ago today Tor was launched by MIT students for anonymous communications: www.torproject.org/about/history/ #cshistory
-
MIT CSAIL
@MIT_CSAIL
What’s one sci-fi movie that impacted your life?
-
MIT CSAIL
@MIT_CSAIL
MIT professor on how LLMs could eventually provide trustworthy financial advice.
Full video: www.youtube.com/watch?v=ui7kRlJMqjM -
MIT CSAIL
@MIT_CSAIL
54(ish) questions w/an MIT CSAIL AI & health researcher.
Full video: www.youtube.com/watch?v=ocjBezrYYqg -
MIT CSAIL
@MIT_CSAIL
The energy consumption of different programming languages (v/@burkov): haslab.github.io/SAFER/scp21.pdf
One finding: Python consumes 76x more energy while being 72x slower than C. -
MIT CSAIL
@MIT_CSAIL
Top websites for accessing free books (v/@nrqa__):
1. ZLIB
2. Project Gutenberg
3. Open Library
4. Google Books
5. BookBub
6. Free eBooks
7. Internet Archive
8. Smashwords
9. LibriVox
10. ManyBooks
11. PDF Drive -
MIT CSAIL
@MIT_CSAIL
76 years ago this week Claude Shannon ushered in the field of information theory with his paper "A Mathematical Theory of Communication", which has been cited over 100,000 times: www.siam.org/publications/siam-news/articles/celebrating-… (v/@TheSIAMNews)
-
MIT CSAIL
@MIT_CSAIL
"The purpose of abstraction is not to be vague, but to create a new semantic level in which one can be absolutely precise." — Edsger Dijkstra
-
MIT CSAIL
@MIT_CSAIL
To understand which diseases or physical traits you’re predisposed to, you'll find the answers sprinkled across your DNA. One powerful method for decoding this genetic forecast: polygenic scores.
MIT CSAIL & University of Tokyo researchers' “GenoBoost” polygenic scoring algorithm captures both additive and non-additive dominance effects for the first time, making more nuanced genetic predictions for individuals: www.csail.mit.edu/news/unlocking-new-predictive-powers-pe… -
MIT CSAIL
@MIT_CSAIL
To give context on age, Grace Hopper began computing at 38, completed the first compiler at 46, helped shape COBOL at 53, kept developing COBOL for the Navy in her 70s, retired from the Navy at 80 & then became a consultant for the Digital Equipment Corporation.
v/@cooperx86 -
MIT CSAIL
@MIT_CSAIL
Cardiovascular disease is the leading cause of death worldwide, primarily damaging the heart’s atriums and ventricles.
Despite being the entryway to the heart, the right atrium has been overlooked — until MIT CSAIL & collaborators developed the most detailed map of its cell types and their functions across different disease states: www.sciencedirect.com/science/article/pii/S26663791240024…
A thread 🫀 -
MIT CSAIL
@MIT_CSAIL
Happy birthday to #basic, the programming language launched at @Dartmouth #otd in 1964 to encourage non-STEM students to use computers: time.com/69316/basic/ (v/@TIME)
-
MIT CSAIL
@MIT_CSAIL
Today is a big day in open-source history: 31 years ago @CERN released the source code for the World Wide Web for anyone to use.
v/@BTGroup -
MIT CSAIL
@MIT_CSAIL
LLM-generated plans in abstract spaces lack physical grounding.
Human demonstrations in high-dimensional continuous spaces lack labels for motion constraints.
MIT CSAIL-led research grounds language plans in demonstrations as “manipulation modes” to build robust policies: news.mit.edu/2024/engineering-household-robots-have-littl…
A thread 🧵 -
MIT CSAIL
@MIT_CSAIL
In 1949, it took the early ENIAC computer 70 hours to figure 2,037 digits of Pi. #piday
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to the World Wide Web!
MIT professor Tim Berners-Lee’s invention was initially described by his CERN supervisor as “vague but exciting”: webfoundation.org/2019/03/web-birthday-30/
(v/@timberners_lee & @webfoundation, image courtesy of Sotheby’s) -
MIT CSAIL
@MIT_CSAIL
Ada Lovelace’s approach to needlepoint contributed to her pioneering computer programming concepts.
Her experience w/embroidery, which involves following repetitive patterns, paralleled the formulaic steps of mathematical calculations: theconversation.com/ada-lovelaces-skills-with-language-mu… -
MIT CSAIL
@MIT_CSAIL
How you know you're old: the "dragon book" was published nearly forty years ago.
www.reddit.com/r/programming/comments/67v0h/ask_reddit_ho… -
MIT CSAIL
@MIT_CSAIL
In 1946, six women programmed the first digital computer: www.digitaltrends.com/computing/remembering-eniac-and-the…
-
MIT CSAIL
@MIT_CSAIL
55 years ago Douglas Engelbart's "Mother of All Demos" introduced such concepts as the mouse, video conferencing, word processing & hypertext: www.smithsonianmag.com/innovation/douglas-engelbart-inven… (via @SmithsonianMag)
-
MIT CSAIL
@MIT_CSAIL
An inclusive training strategy makes genetic prediction models more accurate for everyone, promoting health equity.
CSAIL scientists developed a new method that improves genetic predictions by directly including admixed and ancestry-diverse individuals: news.mit.edu/2023/making-genetic-prediction-models-more-i… -
MIT CSAIL
@MIT_CSAIL
True story: a kid saw this and said, "oh, you 3D-printed the 'Save' Icon." twitter.com/Bill_Gross/status/920406104911233024
Credit: @Bill_Gross -
MIT CSAIL
@MIT_CSAIL
Read the lost thesis of Dennis Ritchie, creator of C programming language & co-creator of Unix: archive.computerhistory.org/resources/access/text/2020/05…
Ritchie never got his PhD b/c he didn't want to pay Harvard the thesis binding fee.
(v/@IEEESpectrum) -
MIT CSAIL
@MIT_CSAIL
Rest in Peace, Doug Lenat.
-
MIT CSAIL
@MIT_CSAIL
The first issue of the defunct computer magazine Byte was published this month in 1975.
v/Small Systems Journal -
MIT CSAIL
@MIT_CSAIL
On August 31, 1955, the term “artificial intelligence” was coined in a conference proposal: stanford.io/2WJJJGN
-
MIT CSAIL
@MIT_CSAIL
Change is hard, but necessary when building safe & equitable ML models.
MIT researchers analyze subpopulation shifts, showing how ML models have performed poorly w/underrepresented subgroups in health care & offering new strategies to mitigate the issue: news.mit.edu/2023/how-machine-learning-models-can-amplify… -
MIT CSAIL
@MIT_CSAIL
Happy 87th bday Margaret Hamilton, the MIT scientist whose code put a man on the moon: s.si.edu/22CL8H8
-
MIT CSAIL
@MIT_CSAIL
What’s the best free software you’re using?
-
MIT CSAIL
@MIT_CSAIL
RIP to MIT professor Patrick Winston, who passed away #otd in 2019. He was director of MIT’s AI lab for 25 years. Here are his complete video lectures for 6.034 (Artificial Intelligence): www.youtube.com/watch?v=TjZBTDzGeGg&list=PLUl4u3cNGP63gFH…
-
MIT CSAIL
@MIT_CSAIL
Alzheimer’s affects 1 in 9 Americans over 65, yet its impact on brain vascular cells has been unclear.
MIT scientists characterized gene expression patterns for 22,500 brain vascular cells, revealing key insights about the disease: www.csail.mit.edu/news/atlas-human-brain-blood-vessels-hi…
@manoliskellis -
MIT CSAIL
@MIT_CSAIL
Today is a big day in open-source history: 30 years ago @CERN released the source code for the World Wide Web for anyone to use.
home.bt.com/news/on-this-day/april-30-1993-internet-revol… (v/@BTGroup) #todayinhistory -
MIT CSAIL
@MIT_CSAIL
#otd in 2019 scientists released the first-ever photo of a black hole. Here’s the moment that Prof. Katie Bouman and colleagues first processed the image: www.nytimes.com/2019/04/11/science/katie-bouman-black-hol…
-
MIT CSAIL
@MIT_CSAIL
If you're looking for:
- JavaScript, PHP, Python
- React, Vue, Svelte
- Node, Django, GraphQL
- Mongo, Firestore, RDS
- Kubernetes, Docker
- AWS, Google Cloud, Azure
You're not looking for a full-stack dev, you're looking for an entire IT department.
credit: @VittoStack -
MIT CSAIL
@MIT_CSAIL
Free MIT videos & online materials from more than 2,400 courses, including intro classes in computer science, AI and algorithms.
Browse our open CS courses here: ocw.mit.edu/search/?d=Electrical Engineering and Computer…
(v/@MITOCW) -
MIT CSAIL
@MIT_CSAIL
Today would have been the 80th birthday of Professor Patrick Winston, who passed away in 2019 after directing MIT’s AI lab for 25 years.
Here are his complete video lectures for 6.034, which have been watched nearly 5 million times: www.youtube.com/watch?v=TjZBTDzGeGg&list=PLUl4u3cNGP63gFH… -
MIT CSAIL
@MIT_CSAIL
800 free computer science classes you can take online right now: github.com/Developer-Y/cs-video-courses
-
MIT CSAIL
@MIT_CSAIL
Free MIT videos & online materials from more than 2,400 courses, including intro classes in computer science, AI and algorithms.
Browse our open CS courses here: ocw.mit.edu/search/?d=Electrical Engineering and Computer…
(v/@MITOCW) -
MIT CSAIL
@MIT_CSAIL
An IBM slide from 1979.
-
MIT CSAIL
@MIT_CSAIL
All major neural networks, in one chart: towardsdatascience.com/the-mostly-complete-chart-of-neura…
v/The Asimov Institute -
MIT CSAIL
@MIT_CSAIL
Alan Turing first defined what’s now the Turing Machine 85 years ago #otd, shaping how machines compute info. today: www.forbes.com/sites/gilpress/2017/11/05/the-turing-test-…
Image v/Godrey Argent Studio & The Royal Society -
MIT CSAIL
@MIT_CSAIL
Published #otd in 1986: a seminal 4-page paper on backpropagation from Geoff Hinton that helped reignite the field of neural networks: towardsdatascience.com/learning-backpropagation-from-geof…
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Mary Allen Wilkes, who in 1961 designed the software for LINC, which many consider to be the world’s first minicomputer. blogs.wellesley.edu/cs/2019/05/10/mary-allen-wilkes-59-wo…
(v/@Wellesley, image v/@nytimes) -
MIT CSAIL
@MIT_CSAIL
Over 65 years ago today researchers ran the first FORTRAN program, considered the beginning of general-purpose programming: www.edn.com/1st-fortran-program-runs-september-20-1954/
(credit @billamend) -
MIT CSAIL
@MIT_CSAIL
Wikipedia articles influence legal decisions of Irish judges, scientists find: www.csail.mit.edu/news/how-wikipedia-influences-judicial-…
@MIT @MaynoothUni -
MIT CSAIL
@MIT_CSAIL
#otd in 1969, Apollo 11 took off - and then landed a man on the moon - thanks to software from an MIT team led by 32-year-old Margaret Hamilton.
Full video: www.youtube.com/watch?v=kTn56jJW4zY (v/@MAKERSwomen)
More: www.smithsonianmag.com/smithsonian-institution/margaret-h… (v/@SmithsonianMag) #apollo11 @NASA -
MIT CSAIL
@MIT_CSAIL
This robotic mouth chants algorithmically generated prayers: theprayer.diemutstrebe.com
-
MIT CSAIL
@MIT_CSAIL
"The Computer Girls," was a 1967 Cosmopolitan piece about a strange new field dominated by women. It was called programming: www.siliconrepublic.com/people/women-in-technology-the-co…
-
MIT CSAIL
@MIT_CSAIL
Supercomputing savant @JDongarra won the Turing Award for making the machines usable & helping democratize supercomputing in the cloud.
He believes MIT CSAIL's #julialang is the way of the future.
zd.net/3jblcCo -
MIT CSAIL
@MIT_CSAIL
Born #otd in 1927: Allen Newell, who in 1955 - before the term “AI” existed - co-wrote what many consider the world’s first AI program.
"Logic Theorist" proved 38 of 52 classic math theorems & established the field of heuristic programing.
Details: history-computer.com/Library/Logic Theorist memorandum.pdf -
MIT CSAIL
@MIT_CSAIL
When programming languages were created:
1957: Fortran
1958: LISP
1964: BASIC
1972: C
1978: SQL
1980: C++
1991: Python
1993: R
1995: PHP, Java & JS
2001: C#
2009: Go
2014: Swift -
MIT CSAIL
@MIT_CSAIL
The difference between Github and Gitlab via @YuHelenYu
-
MIT CSAIL
@MIT_CSAIL
800 free computer science classes you can take online right now: github.com/Developer-Y/cs-video-courses
-
MIT CSAIL
@MIT_CSAIL
An IBM slide from 1979.
-
MIT CSAIL
@MIT_CSAIL
Tech birthdays in 2022:
Lyft: 10 years
Snapchat: 11 years
Uber: 13 years
Twitter: 16 years
Facebook: 18 years
Tesla: 19 years
Google: 24 years
Mozilla: 24 years
Netflix: 26 years
Amazon: 28 years
Apple: 46 years
Microsoft: 47 years
Intel: 54 years
HP: 83 years
IBM: 111 years -
MIT CSAIL
@MIT_CSAIL
RT if you were part of that tiny 1993 bar.
(via @markrendle) -
MIT CSAIL
@MIT_CSAIL
9 Distance Measures in data science with algorithms.
via: @gp_pulipaka -
MIT CSAIL
@MIT_CSAIL
Google was launched on this day, in 1998.
-
MIT CSAIL
@MIT_CSAIL
A SQL cheatsheet for coding needs: github.com/ABZ-Aaron/cheat-sheets
-
MIT CSAIL
@MIT_CSAIL
#otd in 1985 one of the most influential programming textbooks was published: MIT’s “Structure and Interpretation of Computer Programs”.
Read it here for free: web.mit.edu/alexmv/6.037/sicp.pdf -
MIT CSAIL
@MIT_CSAIL
MIT has been named the world’s top university by the QS World University Rankings.
This is the 10th year in a row MIT has received this distinction: news.mit.edu/2021/qs-ranks-world-no-1-21-22-0608
Credit @MIT -
MIT CSAIL
@MIT_CSAIL
MIT CSAIL is excited to announce the Mentored Opportunity in Research Postdoctoral Fellowship (Meteor), to broaden participation in computing and artificial intelligence: www.csail.mit.edu/meteor
-
MIT CSAIL
@MIT_CSAIL
Brain implant gives robotic arm a sense of touch: arstechnica.com/science/2021/05/researchers-add-sense-of-…
-
MIT CSAIL
@MIT_CSAIL
Today's a big day in open-source history: #otd in 1993 @CERN released the source code for the World Wide Web for anyone to use. home.bt.com/news/on-this-day/april-30-1993-internet-revol…
(v/@BTGroup) #tdih #opensource -
MIT CSAIL
@MIT_CSAIL
#otd in 1980 a Japanese computer scientist published a paper proposing the “Neocognitron,” the neural net that directly inspired CNNs.
Kunihiko Fukushima’s paper: www.rctn.org/bruno/public/papers/Fukushima1980.pdf -
MIT CSAIL
@MIT_CSAIL
160 years of MIT, today.
-
MIT CSAIL
@MIT_CSAIL
BREAKING: MIT "smart clothes" use special tactile fibers to sense a person’s movement & determine what pose they're in.
Potential applications:
🏀 coaching
♿ rehabilitation
👴🏽 elder care
Paper: www.nature.com/articles/s41928-021-00558-0
More: www.csail.mit.edu/news/smart-clothes-can-measure-your-mov… (v/@NatureElectron) -
MIT CSAIL
@MIT_CSAIL
Happy birthday Joseph Fourier, whose 1822 equation allows us to listen to mp3s today: gizmodo.com/digital-music-couldnt-exist-without-the-fouri…
-
MIT CSAIL
@MIT_CSAIL
Describe coding in exactly six words.
We’ll RT all the best ones.
From the folks @StackOverflow:
Build. Deploy. Test. Debug. Debug. Debug.
#codingin6words #thursdaythoughts -
MIT CSAIL
@MIT_CSAIL
MIT’s “Introduction to Algorithms,” published #otd in 1990, is the world’s most cited CS text, with 9K citations & half a million copies sold.
mitpress.mit.edu/blog/1990-introduction-algorithms
(@mitpress) -
MIT CSAIL
@MIT_CSAIL
#otd in 2004 Bill Gates told an audience of MIT students "if you invent a breakthrough in artificial intelligence, so machines can learn, that is worth 10 Microsofts.''
www.nytimes.com/2004/03/01/business/microsoft-amid-dwindl… -
MIT CSAIL
@MIT_CSAIL
Today's the 75th anniversary of ENIAC, widely considered to be the first digital computer, coded by six female programmers. www.digitaltrends.com/computing/remembering-eniac-and-the…
@DigitalTrends
#eniacday
(photo: Corbis/Getty Images) -
MIT CSAIL
@MIT_CSAIL
Wow, we're sooooo impressed by this AI that can detect sarcasm. venturebeat.com/2020/11/18/ai-researchers-made-a-sarcasm-…
Paper: www.aclweb.org/anthology/2020.findings-emnlp.124/
(@NVIDIAAI v/@VentureBeat)
#ml #ai #machinelearning -
MIT CSAIL
@MIT_CSAIL
AI calculates breast cancer chances regardless of race: www.csail.mit.edu/news/robust-ai-tools-predict-future-cancer
-
MIT CSAIL
@MIT_CSAIL
WATCH: MIT's robot cheetah front-flips over its brother.
Brother flips out, in its own way.
(v/@MITEngineering)
Full video: youtu.be/Jd_QNW1kFfc
#robots #roboticsThis media couldn't be displayed. See it on the original post. -
MIT CSAIL
@MIT_CSAIL
A demo from 1993 of 32-year-old Yann LeCun showing off the world's first convolutional network for text recognition. #tbt #ml #neuralnetworks #cnns #machinelearning
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to Linus Torvalds, the father of #linux: tcrn.ch/2hXs1H1
-
MIT CSAIL
@MIT_CSAIL
The 10 most popular programming languages of 2020:
1. C
2. Java
3. Python
4. C++
5. C#
6. VisualBasic
7. JavaScript
8. PHP
9. R
10. SQL
#tiobe index
zd.net/3mHA5MM
(v/@LiamT @ZDNet)
#javascript #js #coding #php #python #sql #code #java #r #coding #programming -
MIT CSAIL
@MIT_CSAIL
Today @Microsoft & @codeorg are launching a new K-12 curriculum on AI & ethics.
Check out the team's 7-part video series:
code.org/en-US/artificial-intelligence
(featuring Microsoft CEO Satya Nadella & experts from Amazon, MIT, OpenAI, Tesla & more)
#aiinschools #aiethics -
MIT CSAIL
@MIT_CSAIL
Stay at 127.0.0.1
Wear a 255.0.0.0
#happythanksgiving -
MIT CSAIL
@MIT_CSAIL
Happy birthday to "Calvin and Hobbes", the comic with the most sensible take on "Moore's Law."
-
MIT CSAIL
@MIT_CSAIL
BREAKING: MIT cybersecurity experts warn Big Tech that blockchain voting isn't even close to ready for future elections.
Paper led by Ron Rivest, the "R" in "RSA": people.csail.mit.edu/rivest/pubs/PSNR20.pdf
More: www.csail.mit.edu/news/mit-experts-no-dont-use-blockchain…
#blockchain #electionsecurity -
MIT CSAIL
@MIT_CSAIL
30 years ago today Tim Berners-Lee and Robert Cailliau emailed colleagues outlining a proposal for a hypertext project that they dubbed "WorldWideWeb".
w3.org/Proposal.html
(v/@w3c) -
MIT CSAIL
@MIT_CSAIL
Computer scientists develop a better way to find approximate solutions to the 44-year-old "traveling salesperson" problem.
Paper: arxiv.org/pdf/2007.01409.pdf (v/@UW)
More: www.quantamagazine.org/computer-scientists-break-travelin… (v/@EricaKlarreich @QuantaMagazine)
#mondaymotivation -
MIT CSAIL
@MIT_CSAIL
After 2 weeks of multiple health screens and asking everyone to quarantine, I surprised my closest inner circle with a trip to a private island where we could pretend things were normal just for a brief moment in time.
-
MIT CSAIL
@MIT_CSAIL
This autonomous car can drive itself using only 19 control neurons.
Video: www.youtube.com/watch?v=8KBOf7NJh4Y
More: www.csail.mit.edu/news/new-deep-learning-models-require-f…
(work w/@ISTAustria @tuvienna)
#selfdrivingcars #autonomy #ml #dl #machinelearning -
MIT CSAIL
@MIT_CSAIL
"The Computer Girls," was a 1967 Cosmopolitan piece about a strange new field dominated by women. It was called programming: www.siliconrepublic.com/people/women-in-technology-the-co…
#tbt #womenintech #girlswhocode -
MIT CSAIL
@MIT_CSAIL
Stay at 127.0.0.1
Wear a 255.0.0.0 -
MIT CSAIL
@MIT_CSAIL
MIT offers free video lectures and online materials from more than 2,400 courses, including intro classes in computer science, AI and algorithms.
Browse our open CS courses here: ocw.mit.edu/search/?d=Electrical Engineering and Computer…
(v/@MITOCW)
#wednesdaywisdom -
MIT CSAIL
@MIT_CSAIL
"We [humans] are not the inventors of the first digital computer - we're its descendants."
MIT computational biologist Manolis Kellis on digital inheritance and the human genome: youtu.be/brslF-Cy3HU
@lexfridman #genomics -
MIT CSAIL
@MIT_CSAIL
In the fall of 1973, MIT students helped invented Maze, the first networked game & the first first-person shooter.
The government banned it from ARPANET after learning that half of all network activity was from MIT & Stanford students playing it.
www.polygon.com/features/2015/5/21/8627231/the-first-firs…
#tbt -
MIT CSAIL
@MIT_CSAIL
BREAKING: the Trump administration just dropped its plan to deport international students who take online-only classes, in a direct response to a lawsuit filed by MIT & Harvard.
www.politico.com/news/2020/07/14/trump-administration-dro…
(v/@PerezJr @politico) -
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late John Tukey, who coined the terms “software” and “bit” (short for “binary digit”).
He would have turned 1101000 today: www.nytimes.com/2000/07/28/us/john-tukey-85-statistician-…
(@nytimes, photo by Alfred Eisenstadt) -
MIT CSAIL
@MIT_CSAIL
BREAKING: In the name of open access, MIT has ended journal negotiations w/Elsevier. news.mit.edu/2020/guided-by-open-access-principles-mit-en…
MIT's Framework for Publisher Contracts has been endorsed by 100+ institutions, incl. Duke, Harvard, UCLA & Yale: libraries.mit.edu/scholarly/publishing/framework/
(v/@mitlibraries) #openaccess -
MIT CSAIL
@MIT_CSAIL
Is it time to #cancelmooreslaw?
Not so fast: in a new Science paper, MIT team says that researchers most focus on:
1️⃣ #software performance engineering 👩💻
2️⃣ new #algorithms
3️⃣ streamlined hardware💻
Paper: science.sciencemag.org/content/368/6495/eaam9744
Exec. summary: drive.google.com/file/d/1-gMiSxqnCxgflWJjsV5l4ABkoLAh-fbL… -
MIT CSAIL
@MIT_CSAIL
Just out in Science: MIT researchers say that, with the end of Moore's Law, we have to forget about silicon and start looking elsewhere.
Paper: science.sciencemag.org/content/368/6495/eaam9744
Executive summary: drive.google.com/file/d/1-gMiSxqnCxgflWJjsV5l4ABkoLAh-fbL…
@ScienceMagazine @NewsfromScience -
MIT CSAIL
@MIT_CSAIL
We are excited to announce that, effective today, we have decided to change the name of our lab to the MIT Computer Science and Æ Laboratory, or MIT CSÆL.🧝🏽♀️
All related media inquiries can be directed to Y Æ̃ SR-71 at Y_Æ̃SR-71@csail.mit.edu. -
MIT CSAIL
@MIT_CSAIL
MIT study: relaxing quarantine measures would create “exponential explosion” in COVID-19 cases & consequences that would be "catastrophic."
news.mit.edu/2020/new-model-quantifies-impact-quarantine-… (v/@MITMechE)
#covid19 #covid19usa -
MIT CSAIL
@MIT_CSAIL
This robotic mouth chants algorithmically generated prayers: theprayer.diemutstrebe.com
-
MIT CSAIL
@MIT_CSAIL
COVID-19 may quickly deplete the world's supply of ventilators, respirators and other medical supplies.
Join a global effort to design and develop open-source solutions to this problem: www.facebook.com/groups/670932227050506/about/
More info: hackaday.com/2020/03/12/ultimate-medical-hackathon-how-fa…
(v/@hackaday) -
MIT CSAIL
@MIT_CSAIL
In German the word "Verschlimmbessern" means "to make something worse while repairing it."
In English we call that "software development."
#fridayfeels -
MIT CSAIL
@MIT_CSAIL
“Any fool can write code that a computer can understand. Good programmers write code that humans can understand.” - Martin Fowler
(comic v/@PHDcomics)
#wednesdaywisdom #coding -
MIT CSAIL
@MIT_CSAIL
The five stages of g̶r̶i̶e̶f̶ debugging.
(comic credit: Sourcecode)
#fridayfeels -
MIT CSAIL
@MIT_CSAIL
#otd in 1958 Ivan Sutherland pushed the world’s first paper on virtual reality, “A head-mounted three-dimensional display.” He called the device “The Sword of Damocles.” www.cacs.usc.edu/education/cs653/Sutherland-HeadmountedDi…
-
MIT CSAIL
@MIT_CSAIL
Watch 9 of MIT's mini-cheetahs do coordinated dance moves & back-flips.
Full video: youtu.be/G6fMV1UPzkg
(v/@MITMeche) -
MIT CSAIL
@MIT_CSAIL
WATCH: MIT's self-driving car (on right) can see other cars around corners by sensing tiny changes in shadows: www.youtube.com/watch?v=NqFEd-9J1lE
More info: news.mit.edu/2019/helping-autonomous-vehicles-see-around-…
(joint work w/@ToyotaResearch) -
MIT CSAIL
@MIT_CSAIL
The top 20 free online CS courses of all time, via @classcentral.
Full list: www.classcentral.com/report/top-moocs-2019-edition/ #learntocode #mondaymotivation -
MIT CSAIL
@MIT_CSAIL
Happy birthday, #nikolatesla!
In a 1926 interview, he said that we would one day be able to carry a wireless phone "in [our] vest pocket." www.tfcbooks.com/tesla/1926-01-30.htm
#wednesdaywisdom -
MIT CSAIL
@MIT_CSAIL
Describe programming in only six words.
We’ll RT all the best ones.
Ours:
Turning ideas and caffeine into code.
#programmingin6words #wednesdaywisdom -
MIT CSAIL
@MIT_CSAIL
Happy Passover! In 1977 MIT professor Ron Rivest came home after a Manischewitz-filled sedar and couldn’t sleep - so he stayed up & wrote an entire paper that became the basis for RSA encryption: www.math.uchicago.edu/~may/VIGRE/VIGRE2007/REUPapers/FINA… v/(@Calderbank160O @UChicago) #tdih
-
MIT CSAIL
@MIT_CSAIL
Left: MIT computer scientist Katie Bouman w/stacks of hard drives of black hole image data.
Right: MIT computer scientist Margaret Hamilton w/the code she wrote that helped put a man on the moon.
(image credit @floragraham)
#ehtblackhole #blackholeday #blackhole -
MIT CSAIL
@MIT_CSAIL
BREAKING: the first-ever image of a black hole, developed with the help of MIT computer scientists. www.washingtonpost.com/science/2019/04/10/see-black-hole-…
#ehtblackhole #eventhorizontelescope -
MIT CSAIL
@MIT_CSAIL
3 years ago MIT grad student Katie Bouman led the creation of a new algorithm to produce the first-ever image of a black hole.
Today, that image was released.
More info: www.theguardian.com/science/2019/apr/10/black-hole-pictur…
2016 story: news.mit.edu/2016/method-image-black-holes-0606
#ehtblackhole #eventhorizontelescope -
MIT CSAIL
@MIT_CSAIL
"The Computer Girls," was a 1967 Cosmopolitan piece about a strange new field dominated by women. It was called programming: www.siliconrepublic.com/people/women-in-technology-the-co…
-
MIT CSAIL
@MIT_CSAIL
Humans take ~14K brain-hours to learn to run. Robots can learn in less than half as many CPU hours, but it sometimes looks like this: vids.onsizzle.com/humanoid-27-dofs-21-actuators-me-trying… (credit: @eron_gj)
-
MIT CSAIL
@MIT_CSAIL
Bias in AI: translating English to Turkish, a gender neutral language, then that same Turkish phrase back to English (credit @_DianeKim) #tuesdaythoughts
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to computing legend #gracehopper, inventor of #cobol, compilers, & the word "bug." www.cbsnews.com/news/grace-hopper-she-taught-computers-to… (article: @cbs, photo: Bettmann Archive)
-
MIT CSAIL
@MIT_CSAIL
The secret to making ML systems less discriminatory: collecting better data.
“Even the best mathematical models are no match for a biased dataset,” says MIT's Irene Chen.
www.csail.mit.edu/news/making-ml-systems-less-biased (image v/@TechTarget) -
MIT CSAIL
@MIT_CSAIL
45 years ago today TCP/IP was born, when Vint Cerf & Bob Kahn submitted a paper describing a protocol for sending packets across networks: www.cs.princeton.edu/courses/archive/fall06/cos561/papers…
-
MIT CSAIL
@MIT_CSAIL
The Turing Test was born #otd in 1950, in Alan Turing’s seminal paper “Computing Machinery and Intelligence”: www.wired.com/article/turing-test (credit: @wired)
-
MIT CSAIL
@MIT_CSAIL
VIDEO: How 3D-printed “motion sculptures” could help athletes improve their technique: news.mit.edu/2018/creating-3-d-printed-motion-sculptures-… #mosculp #uist2018 #uist
-
MIT CSAIL
@MIT_CSAIL
Happy birthday #nikolatesla! Fun fact: his statue in Silicon Valley radiates free WiFi: www.reddit.com/r/todayilearned/comments/339md3/til_theres… (photo: MIT)
-
MIT CSAIL
@MIT_CSAIL
Syrian-born MIT professor Dina Katabi's wireless system can "see" through walls to monitor the health of the elderly: www.washingtonpost.com/news/innovations/wp/2015/10/28/mit… #celebrateimmigrants
-
MIT CSAIL
@MIT_CSAIL
Wireless smart-health system uses AI to sense movement through walls: news.mit.edu/2018/artificial-intelligence-senses-people-t…
-
MIT CSAIL
@MIT_CSAIL
25 years ago today the first major web browser was released: Mosaic 1.0. www.wired.com/2010/04/0422mosaic-web-browser/ #otd #todayintech
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to LISP, first unveiled in a 1960 paper written @MIT: github.com/papers-we-love/papers-we-love/blob/main/comp_s… (credit: @xkcdComic)
-
MIT CSAIL
@MIT_CSAIL
20 years ago this week Sergey Brin and Larry Page published the CS paper that birthed Google: stanford.io/2IBq2o3 (credit: Stanford University)
-
MIT CSAIL
@MIT_CSAIL
#otd 29 years ago Tim Berners-Lee first proposed the idea for the World Wide Web @CERN.
His boss' response: "Vague but exciting..."
betterworld.mit.edu/world-wide-web/ #happybirthdaywww #www @timberners_lee -
MIT CSAIL
@MIT_CSAIL
Bias in AI: translating English to Turkish, a gender neutral language, then that same Turkish phrase back to English (credit @_DianeKim)
-
MIT CSAIL
@MIT_CSAIL
#otd in 1946, six women programmed the first digital computer: www.digitaltrends.com/computing/remembering-eniac-and-the… #eniacday
-
MIT CSAIL
@MIT_CSAIL
Why it's important to teach math. (credit @SMBCComics)
-
MIT CSAIL
@MIT_CSAIL
“Rudolph the robot reindeer / had a very shiny chrome” #happyholidays from Boston Dynamics’ robotic reindeer www.youtube.com/watch?v=RDZu04v7_hc (h/t @GordonHunt)
-
MIT CSAIL
@MIT_CSAIL
How to accidentally break a Skype bot. www.reddit.com/r/ProgrammerHumor/comments/4w3cnz/so_i_acc… (h/t @Reddit)
-
MIT CSAIL
@MIT_CSAIL
The world's first program was coded in 1842 by #adalovelace, born over 200 years ago today! www.mentalfloss.com/technology/computers/ada-lovelace-fir…
-
MIT CSAIL
@MIT_CSAIL
Happy 70th birthday to Geoffrey Hinton, whose work in neural networks has led many to call him the “godfather of AI”: surveymrs.com
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to MIT’s Turing Award–winning cryptographer Silvio Micali: bitcoinmagazine.com/business/interview-cryptographer-silv…
-
MIT CSAIL
@MIT_CSAIL
Bias in AI: translating English ➡️Turkish, a gender neutral language, then that same Turkish phrase back to English (credit @_DianeKim)
-
MIT CSAIL
@MIT_CSAIL
BREAKING: new MIT system enables glasses-free 3D TV. news.mit.edu/2017/watch-3-d-movies-at-home-sans-glasses-0712
-
MIT CSAIL
@MIT_CSAIL
#til: In 36 years, the avg. price per gigabyte of hard drive storage went from $400K to $0.019. @Reddit www.statisticbrain.com/average-cost-of-hard-drive-storage/
-
MIT CSAIL
@MIT_CSAIL
Happy 70th birthday to Prof. Ron Rivest, who co-invented #rsa public-key encryption: news.mit.edu/2011/rivest-unlocks-cryptographys-past-looks…
-
MIT CSAIL
@MIT_CSAIL
Happy birthday to the late Ray Tomlinson, the creator of email & @MIT alum: www.wired.com/2012/07/ray-tomlinson-email/
-
MIT CSAIL
@MIT_CSAIL
Which programming languages are used most late at night? @StackOverflow crunches the numbers: stackoverflow.blog/2017/04/19/programming-languages-used-…
-
MIT CSAIL
@MIT_CSAIL
This week MIT launched a new AI initiative w/Microsoft & other tech companies: news.mit.edu/2017/csail-launches-artificial-intelligence-…
-
MIT CSAIL
@MIT_CSAIL
Things that wouldn't exist without Tim Berners-Lee: HTML, HTTP, URLs, and memes. ind.pn/2oy9fM2 #turing50
-
MIT CSAIL
@MIT_CSAIL
To enable robots w/more sensors, MIT team turns to beetles for inspiration: news.mit.edu/2017/goldbug-beetle-printable-sensor-laden-s…
-
MIT CSAIL
@MIT_CSAIL
Happy birthday Joseph Fourier, whose 1822 equation allows us to listen to mp3s today: gizmodo.com/digital-music-couldnt-exist-without-the-fouri…
-
MIT CSAIL
@MIT_CSAIL
Why you should learn math, in one comic strip: smbc-comics.com/index.php?id=4150 credit @SMBCComics
-
MIT CSAIL
@MIT_CSAIL
Computing jobs are now the #1 source of ALL new wages in the US, says @codeorg: codeorg.medium.com #csforall
-
MIT CSAIL
@MIT_CSAIL
Born 100 yrs ago today: AI pioneer Herb Simon, father of “heuristic programming.” amturing.acm.org/award_winners/simon_1031467.cfm
-
MIT CSAIL
@MIT_CSAIL
Remember when Disney bought MIT for $7 billion 18 years ago? hacks.mit.edu/Hacks/by_year/1998/disney_buys_mit/ #aprilfools
-
MIT CSAIL
@MIT_CSAIL
The #voyager spacecraft’s computer has a total memory of only 69kb www.wired.com/2013/09/vintage-voyager-probes/
-
MIT CSAIL
@MIT_CSAIL
.@Toyota's $1bn AI team to include fmr CSAIL director @rodneyabrooks: for.tn/1Z9dxXc
-
MIT CSAIL
@MIT_CSAIL
DNA data takes hours to search. Bonnie Berger has found a way to speed it up: www.csail.mit.edu/news/searching-big-data-faster

















































































































































































































































































































































































































































































































































































































































































































































