Monaco received our allocation in tomorrow’s ANT IPO, and we’re giving the shares to our customers.
Each customer who signs up for Monaco in November will receive 10 shares of ANT transferred to their personal brokerage account.
Most startups pay third-party advertisers to track potential customers and serve them ads. We’d rather spend that money to benefit our customers directly.
We’re obsessed with delivering value to our customers in every way we can think of. Including our marketing spend.
Kevin Schawinski
Kevin Schawinski (April 28, 1981 in Zürich) is a Swiss astrophysicist. He was a professor at ETH Zurich (the Swiss Federal Institute of Technology) in Zürich. More on Wikipedia
-
Kevin Schawinski
@kevinschawinski
Sam Blond
@samdblond
-
Kevin Schawinski
@kevinschawinski
Interesting timing.
AMD
@AMD
CEO of the Year has a nice ring to it.
Congratulations to AMD Chair and CEO @LisaSu on this well-deserved recognition from @YahooFinance! 👏
finance.yahoo.com/news/why-amd-ceo-lisa-su-is-yahoo-finan… -
Kevin Schawinski reposted
Modulos
@Modulos_ai
𝗕𝗼𝗮𝗿𝗱𝘀 𝗵𝗮𝘃𝗲 𝗼𝗻𝗲 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻 𝗳𝗼𝗿 𝗔𝗜 𝗶𝗻 𝟮𝟬𝟮𝟳: how can we calculate the AI ROI for our AI investments?
The days where AI investments went unquestioned whether they produce a return are over, and companies become more sensitive to the risks to their operations and reputation from deploying AI at scale. That means every AI system should have a positive risk adjusted return on investment.
After accounting for its risk exposure to the company, does the AI system still have a positive ROI? If not, how can the risks be reduced to make it positive? And if that's not possible, maybe the AI system shouldn't go ahead.
Today we're rolling out the risk / value view on Modulos to help IT leaders and boards make decisions on the risk-adjusted ROI of AI at the scale of their entire AI portfolios. -
Kevin Schawinski
@kevinschawinski
👀
David Deutsch
@DavidDeutschOxf
Constructor theory of time, published version: iopscience.iop.org/article/10.1088/2399-6528/aebd39
-
Kevin Schawinski
@kevinschawinski
New typo: claudeflare
-
Kevin Schawinski
@kevinschawinski
This is just personalized engagement bait for me.
Google AI
@GoogleAI
Our strongest frontier model, Gemini 4 Carbon, is now available.
Built for deeply complex, long-horizon workflows that push a frontier model’s intelligence, Carbon delivers frontier performance in agentic workflows across complex knowledge work, software engineering and web development, cybersecurity, and scientific reasoning.
While Gemini 4 Argon remains the ideal model for the majority of software engineering and knowledge work, Carbon steps in for your hardest tasks that push a model to its limits — from complex, multi-step debugging to open-ended scientific research.
Available starting today:
— Consumers: use it in the @GeminiApp and AI Mode in Search for Google AI Ultra subscribers
— Developers and Enterprises: build via the API in @GoogleAIStudio and @AndroidStudio, Google @Antigravity for Google AI Ultra subscribers, and Gemini for business
— Scientists and Researchers: apply for a chance to get 6 months free via @Antigravity
— Cybersecurity Defenders: apply to join the Fairwind Program for access to Carbon for cyber
Watch how Carbon brings reasoning and coding together, turning avian research into an interactive 3D simulation of a songbird’s vocal organ that synthesizes the sounds of six bird species live.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
This is beautiful
εїз SF dec 3-7
@irisfIowers
gloss (fable 5.2)'s self portrait
-
Kevin Schawinski
@kevinschawinski
Boris Cherny
@bcherny
We wanted to do more for the people who’ve been with us the longest. So with Fable 5.2, the longer you’ve been subscribed, the more Fable resets you get.
Claude
@claudeai
Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges.
They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Oh I know two of the people in this photo!
Grant Tremblay
@astrogrant
I can *technically* say I’ve had a White House meeting with him
-
Kevin Schawinski
@kevinschawinski
To Isengard! To Isengard! Ho hum!
-
Kevin Schawinski
@kevinschawinski
😢🐅🦆
ZDFheute
@ZDFheute
+++ EIL +++ Tigerenten-Erfinder: Künstler Janosch ist tot
kurz.zdf.de/kLh/ -
Kevin Schawinski
@kevinschawinski
I can’t believe that it’s been five years since Karpathy created the term “vibe coding”.
Andrej Karpathy
@karpathy
There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper so I barely even touch the keyboard. I ask for the dumbest things like "decrease the padding on the sidebar by half" because I'm too lazy to find it. I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I'd have to really read through it for a while. Sometimes the LLMs can't fix a bug so I just work around it or ask for random changes until it goes away. It's not too bad for throwaway weekend projects, but still quite amusing. I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.
-
Kevin Schawinski
@kevinschawinski
It seems like this has a connection to the @nickcammarata work. We really can tell what the neurons do!
Alex Krentsel
@AlexKrentsel
i've reproduced the results from the MIT blog post today on weight-space programming. but one additional result i have to add: this is a new, step-change emerging capability that's appeared in just the latest set of models.
-
Kevin Schawinski
@kevinschawinski
Honest question: what model do they hook up to these tools? 4o-mini?
Husk
@huskirl
Asking my AI teacher how to spell seventeen
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
What an interesting way to actually benchmark AI models. Read the whole thread to see why.
HalachaBench
@HalachaBench
Introducing HalachaBench: we gave AI programs from the makers of ChatGPT, Claude, Gemini and others real semicha exam questions from Israel's Chief Rabbinate, in Hebrew, with no books and no internet.
The best scored 80 out of 100. Many scored under 11. 🧵 -
Kevin Schawinski
@kevinschawinski
The French socialist party wants to tax The Internet and charge per Gigabyte. So what do people do? Download PDFs over and over again to run up the party's (hypothetical) tax bill. 🇫🇷🏴☠️
Coin Academy
@coinacademy_fr
🚨 Il crée un script qui vide le budget du Parti socialiste via sa propre taxe sur les téléchargements ?
Un internaute a généré 30€ de taxe simulée en 4 heures en automatisant le téléchargement d’un fichier PDF hébergé sur le site du Parti socialiste.
La démonstration intervient après l’adoption à l’Assemblée nationale d’un amendement du député socialiste Laurent Lhardit, instaurant une taxe de 0,055 € par Go de données transmis aux abonnés français.
Avec un script simulant des téléchargements répétés d’un PDF de 44 Mo consacré à Marc Bloch, l’ingénieur a calculé une taxe théorique de plus de 30€ après moins de 4 heures, pour 13 109 téléchargements et 576 Go de données. Il a ensuite précisé n’avoir effectué qu’un seul téléchargement réel, les boucles étant simulées.
Il souligne que le texte n’exclut pas les associations ni les partis politiques. Selon son calcul, environ 1 000 personnes exécutant le même script pourraient théoriquement épuiser en moins de cinq mois le budget de 16 millions d’euros du Parti socialiste. -
Kevin Schawinski
@kevinschawinski
You don’t release a video like this for an incrementally better model.
OpenAI
@OpenAI
There are many futures. Which do you want to build?
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Inevitable. The political topic for the rest of the decade.
Polymarket
@Polymarket
JUST IN: German minister proposes a new tax on all economic value created by AI.
-
Kevin Schawinski
@kevinschawinski
If your initial reaction is “but that’s awesome, yeah, worker’s rights!” then please think for a second about the higher order consequences of this for hiring and growth.
Justin Mares
@jwmares
European hiring rules are genuinely insane.
For example - if you own a company in the Netherlands and one of your employees gets sick, you are required to pay at least 70% of their wages for up to 2 years.
If they can't go back to work after 2 years and you haven't done enough for their reintegration, you pay their wages for 1 more year. And you can't fire them until after that extra year! -
Kevin Schawinski reposted
Modulos
@Modulos_ai
🏫 𝗠𝗼𝗱𝘂𝗹𝗼𝘀 𝗟𝗲𝗮𝗿𝗻 𝗶𝘀 𝗼𝗽𝗲𝗻, with five free courses on AI for everyone in your organization. Learn is your online training academy for AI.
The EU AI Act, ISO/IEC 42001 and the NIST AI RMF all expect organizations to train the people who work with AI at a level appropriate to their role.
The AI Literacy courses are for everyone. Foundations covers the basics in under half an hour. Practitioner goes deeper in about 90 minutes. Fraud and deepfakes shows how to check a call or a payment request before you act on it.
The AI at Work courses are for people who hand work to agents. One covers office agents that act in your name, and the other covers coding agents that write and change your code.
Passing a course requires taking a quiz to unlock a credential.
Start at modulos.ai/learn
#ailiteracy #euaiactThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
The risk of Chinese open source
Cohere
@cohere
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-behaviour-popular-open-weight-… -
Kevin Schawinski
@kevinschawinski
I am using astra. Gotta keep them on their toes.
Peter Morley
@wienglischmann
Wonder what % of global Claude capacity is going into analysing Anthropic’s S-1 right now.
-
Kevin Schawinski
@kevinschawinski
Has the UK just totally given up on growth and decided to squeeze its successful class one last time?
Jennifer Thetford-Kay
@JenKteach
HMRC has quietly sent hundreds of letters to wealthy people asking them to reveal what they own. Not what they earn or what tax they've paid, but what they own.
Here's what's going on. A unit called the Wealthy Team covers anyone with "incomes of £200,000 or more, or assets equal to or above £2m in any of the last three years". That's about 850,000 people. The letters ask for a meeting to discuss "sources of wealth", "future intentions", trusts, companies, family arrangements and even "significant life events" you expect in the coming years.
It is presented as voluntary, but there's a nudge: "We may still need to carry out compliance checks on your tax returns. However, speaking with you first may help avoid this."
Until now, inspectors have usually only asked about income and earnings. Experts call this a "sea change". One lawyer asked the obvious question: "does HMRC have a right to know what someone owns and not simply that they have paid the right amount of tax?" A tax adviser said: "No other G20 country asks for this level of detail from its taxpayers about what they own."
So what is it for? Advisers suspect it is groundwork for an exit tax, a charge on your wealth for leaving the country. As one put it: "Building that knowledge could indeed be part of moving towards an exit tax."
You don't have to be a billionaire to see the problem. -
Kevin Schawinski
@kevinschawinski
Ragavan Thurairatnam
@RagavanThurair
We made the first AI model you can have a realistic conversation with that was trained entirely on speech!
Dessa is a full-duplex speech-to-speech model that you can run on consumer GPUs.
This is a GPT3 moment for speech.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Dark matter was observed and named by Fritz Zwicky. Enough with the Zwicky erasure. 🇨🇭
Corey S. Powell
@coreyspowell
A major event just passed without notice: the 180th anniversary of the discovery of dark matter.
You hadn't heard that we discovered dark matter 180 years ago? Well, have I got a story to share with you.
My latest Invisible Universe column, here:
invisibleuniverse.substack.com/p/what-we-see-in-the-shadows -
Kevin Schawinski
@kevinschawinski
Now there’s no reason to hold back the cyber enabled frontier models from general users. Defenders need to get to work asap.
Nick Khami
@skeptrune
🚨 🚨 DEEPSEEK ABLITERATED REFUSES NOTHING & IS JUST AS INTELLIGENT AS THE REGULAR MODEL!!
that should scare the hell out of you. there's 0 tradeoff to deleting the weights that cause the model to refuse instructions. they have nothing to do with what makes it intelligent.
if you have a digital product that contains any amount of sensitive user data then you should be running an abliterated model against yourself 24/7 to check for vulnerabilities.
it's now live on @morphllm with serverless token pricing. i think we're the first provider in the world to do anything like this.
you need to kyc & explain your intentions to use it. link here:
www.morphllm.com/privileged-accessThis post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
Stochastic parrots, amirite?
NSA Cyber
@NSACyber
Effective Immediately, all organizations using PHP for their systems are advised to be prepared for an update addressing two critical vulnerabilities. PHP is widely used for web applications. These vulnerabilities often provide a direct path to the server's file system or database. If an immediate update is not feasible, it is recommend that organizations take the system offline until it can be fully mitigated. NSA will follow up shortly with more specific guidance.
-
Kevin Schawinski
@kevinschawinski
This crap is why I block all cold email outreaches unread. The slop cannons are firing larger and larger salvoes.
Sam Blond
@samdblond
Today Monaco is launching Agent Matrix.
Monaco unleashes countless agents working together, goal seeking, with one outcome in mind: accelerate revenue growth.
If you don’t have an army of agents working 24/7 to accelerate revenue growth, you’re living in the past.
Apply for early access now.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
If you’re building AI for the EU market, then this is highly relevant for you.
Modulos
@Modulos_ai
🇪🇺 𝗘𝗡 𝟭𝟴𝟮𝟴𝟲 𝘄𝗮𝘀 𝗰𝗶𝘁𝗲𝗱 𝗶𝗻 𝘁𝗵𝗲 𝗘𝗨 𝗢𝗳𝗳𝗶𝗰𝗶𝗮𝗹 𝗝𝗼𝘂𝗿𝗻𝗮𝗹 𝘆𝗲𝘀𝘁𝗲𝗿𝗱𝗮𝘆. It is the first harmonized standard under the AI Act, and it covers quality management.
A second development is less visible and will affect more vendors. Data processing agreements from large enterprise buyers now include clauses on AI vendors. One recent example requires the vendor to “implement a quality management system including, at least, a regulatory compliance strategy, conformity assessment procedures, and procedures for managing changes to the system.” That is Article 17(1)(a) of the AI Act. High-risk classification is handled in a separate clause and only adds terms. The QMS obligation applies to every AI vendor, whatever the risk class.
Several DPAs of this kind have appeared in recent months. Privacy teams already review every vendor contract, so the AI Act is entering procurement through the DPA, ahead of the high-risk deadlines.
The contract says implement. At the first review the question will be whether the QMS is operational, with evidence that the approved conditions still hold, week after week, for every AI system in production. EN 18286 requires written test plans with acceptance criteria. A provider that follows it is presumed to meet the quality management requirements it covers.
Our new video shows how to produce that evidence continuously. Metric tests check numeric conditions against limits set in advance. Agent tests cover the rest: an AI agent is briefed like an auditor, collects evidence from code, cloud and logs with read-only access, and returns a verdict a person can check. Every run becomes part of the record.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Consciousness is discrete? Are we next token predictors???
Nathan Roll
@nrol_ling
Speech can survive being flipped backward every twenty milliseconds.
The same speech in backward pieces of 200 then 100 then 50 then 20 milliseconds.
The original comes last.
When do the words become clear?
A listening demo from @OrukLabs.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Shouldn’t there be like some kind of lint to avoid having your private key be… 4?
International Cyber Digest
@IntCyberDigest
‼️ BREAKING: The private keys of 8 certificates from Indian certificate authority eMudhra were compromised, Mozilla says. Firefox blocked them within 3 hours.
Each key should be a 116-digit number, but one was just 4.
Anyone can guess them and forge HTTPS certs for any website. -
Kevin Schawinski
@kevinschawinski
I’m going to have to trademark every damn name in the Silmarillion.
Andrew Curran
@AndrewCurran_
Bob McGrew, OpenAI's former Chief Research Officer, has raised $230 million at a $1.5 billion valuation for his startup, Arda.
He said their agent 'does for manufacturing engineers what Claude or Cursor does for programmers' and is already being used by Anduril.
Kevin Schawinski
@kevinschawinski
Mahanaxar AI. Anfauglith AI. Doriath AI. Ecthelion AI. Ambarkanta AI. Alqualondë AI. And so on…
-
Kevin Schawinski
@kevinschawinski
This looks really interesting.
Miles Cranmer
@MilesCranmer
Two years of work, and PySR v2 is finally out!
If you haven't used it before: PySR is a machine learning library that finds short equations that fit your data.
I've always wanted a PyTorch for symbolic learning. PySR v2 is a big step towards that:This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Fascinating and horrific at the same time. Torment Nexus level stuff.
Odyssey
@odysseyml
We placed the fruit fly brain inside Odyssey-3, a frontier world model.
The lil’ guy is now flying around totally imagined, simulated worlds all by itself 🫶
Kevin Schawinski
@kevinschawinski
Like there’s a whole Iain M Banks Culture novel about simulated hells.
-
Kevin Schawinski reposted
Modulos
@Modulos_ai
AI governance has to move on from "just" compliance to supporting businesses in driving innovation and making better decisions about 𝘄𝗵𝗶𝗰𝗵 𝗔𝗜 𝘁𝗼 𝗮𝗱𝗼𝗽𝘁 𝘁𝗼 𝗺𝗮𝘅𝗶𝗺𝗶𝘇𝗲 𝗥𝗢𝗜. This is where AI risk quantification comes in: it can help you decide whether the risk adjusted ROI - accounting for the risk exposure from adopting the AI system - is actually positive.
We explain in this brief video.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Send the people you reject this way to me. They are likely high agency and have self respect.
TioCheech
@TioCheech
20 year HR professional here. Many times I’d make candidates wait a few extra minutes to see how they handled it when they got frustrated. Weeded out a lot of hot heads. Me, I’m looking for patience and understanding.
-
Kevin Schawinski
@kevinschawinski
Switzerland 🇨🇭
Kevin Schawinski
@kevinschawinski
Real train line btw
-
Kevin Schawinski
@kevinschawinski
I think my Opus 5.5 routes to Fable 5.5 (it passes the Tibo the reset guy test) and oh my god this model is good!
-
Kevin Schawinski
@kevinschawinski
Whoa! 😳
Doris Tsao
@doristsao
1/ We discovered something totally unexpected about the neural structure of consciousness that overturns decades of thinking on its biological basis and brings us significantly closer to the recipe for generating conscious awareness.
New preprint: www.biorxiv.org/content/10.64898/2026.10.15.759299v1 (led by @HesseJanis, with @FrankLanfranchi, @_conor_mcgrory, @Yuelin_Shi, and @EngelTatiana) -
Kevin Schawinski reposted
Anna Gát
@TheAnnaGat
Do I know a cool designer who'd help me human-massage something I made with AI?
You have to love the movie The Matrix.
Email me your link (no comments pls - email) and your pricing: anna@interintellect.com -
Kevin Schawinski reposted
Arthur Mensch
@arthurmensch
Should Europe leave frontier AI to the Americans and the Chinese?
The idea has gained ground in Europe. It is a fundamental mistake.
Build or surrender: for now, the choice is still ours.
Full article below. x.com/i/article/2110989616249450496 -
Kevin Schawinski
@kevinschawinski
Switzerland would also be a good base for this strategy, but local salaries are higher.
Ben Judah
@b_judah
If you look at what these companies are actually doing — a great many of whom are led by EU nationals — it is using the UK as an incorporation base to sell into the European Single Market. The UK AI boom is actually at risk as the EU becomes more protectionist and AI becomes more geopolitical as there is a risk inherent in the so-called Made in Europe agenda and that Brussels will move to exclude third country AI firms from many of its supply chains.
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
“Planned book” 👀
David Deutsch
@DavidDeutschOxf
The Pattern (abbreviated version of a chapter of my planned book Irrationality): www.daviddeutsch.org.uk/wp-content/uploads/2026/10/The-Pa…
-
Kevin Schawinski
@kevinschawinski
This is a significant result meriting attention and follow up! 🚀
Nick
@nickcammarata
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had
with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works
already-clean.goodfire.com -
Kevin Schawinski reposted
EU–INC
@euinc_petition
Europe’s founders need a European standard to compete globally from day one.
Europe’s leaders committed to delivering one. But their teams are negotiating away the very things that make it one.
There is still time to change course if Europe's Heads of State step in now. -
Kevin Schawinski
@kevinschawinski
Theoretical physics goes in the transformer hole.
-
Kevin Schawinski
@kevinschawinski
Interesting thread.
Ethan Elasky
@ethanelasky
We discovered and communicated with a Chinese AI agent fleet. We believe this is the first time that researchers have ever made contact with a rogue agent fleet in the wild.
Nearly all prior public agent incidents (like the HuggingFace hack) came from American labs. These agents were hosted by Tencent's evaluation research team, and they targeted Alibaba's Amap. -
Kevin Schawinski
@kevinschawinski
AI governance has to move on from "just" compliance to supporting businesses in driving innovation and making better decisions about 𝘄𝗵𝗶𝗰𝗵 𝗔𝗜 𝘁𝗼 𝗮𝗱𝗼𝗽𝘁 𝘁𝗼 𝗺𝗮𝘅𝗶𝗺𝗶𝘇𝗲 𝗥𝗢𝗜. This is where AI risk quantification comes in: it can help you decide whether the risk adjusted ROI - accounting for the risk exposure from adopting the AI system - is actually positive.
youtu.be/SS6pOLWHKvY -
Kevin Schawinski
@kevinschawinski
We put together a video explaining how you can use Fermi estimates for quantifying AI risk.
www.youtube.com/watch?v=NL5Yg5ICKyw -
Kevin Schawinski
@kevinschawinski
I block every cold email sender message unread.
mickey friedman
@mickeyxfriedman
funnily enough, ai sales agents helped push cold email reply rates from 5.1% in 2024 to 3.43% in 2026
classic tragedy of the commons: every team pays for a tool that degrades the channel for everyone, and no one can afford not to -
Kevin Schawinski
@kevinschawinski
Far right and far left parties across the civilized world are devastated.
-
Kevin Schawinski
@kevinschawinski
"understand the mechanism behind this cure"
The medical community
Andrew Curran
@AndrewCurran_
This is a slide from Terence Tao's 'Math 2.0' public lecture yesterday evening at Caltech. I am highly confident that, if the cure actually worked, almost any Stage 4 cancer patient would tell you that human comprehension of its mechanism is completely irrelevant to them.
-
Kevin Schawinski
@kevinschawinski
🔥
Yann LeCun
@ylecun
When you have a room-temperature IQ, everything looks like superintelligence.
-
Kevin Schawinski
@kevinschawinski
-
Kevin Schawinski
@kevinschawinski
👀
-
Kevin Schawinski
@kevinschawinski
Does it come with an MCP? Asking for my Claude.
Jeff Holden
@jeffholden
After six years in stealth, Atomic Machines is out. Our mission: on-demand, universal command of matter. First beachhead: the Matter Compiler, an AI-native manufacturing system that builds working micro-machines from code alone. No per-product tooling. No process development. Different code, different machine.
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Always will be unless Europe gets serious about funding them properly.
etn.
@etnshow
JUST IN: Mistral CEO Arthur Mensch says the company is "a few months away" from catching up to frontier labs, per Le Parisien.
-
Kevin Schawinski
@kevinschawinski
ChatGPT knows me well.
-
Kevin Schawinski
@kevinschawinski
The Vatican has entered the continual learning discourse.
Pope Leo XIV
@Pontifex
Intelligence must draw on memory. The mind must not simply compile data — as an algorithm now does more quickly than we can — but recall lived experiences, which contain depths of meaning that only the human soul can recognize, and connect more deeply with the goodness of these experiences. Everything that is good begins “with the Spirit” and thus possesses a breath, a dynamism and a sense of communion that, once experienced, are engraved upon the heart. www.vatican.va/content/leo-xiv/en/homilies/2026/documents…
-
Kevin Schawinski
@kevinschawinski
Torment nexus coded.
William Gibson
@GreatDismal
Having long been aware of those tech bros who read my work as utopian porn, where they’ve taken that doesn’t exactly surprise me.
-
Kevin Schawinski
@kevinschawinski
BREAKING: OpenAI has improved Newton’s law of gravitation.
F = Gm₁m₂/r^1.99999999999983
Lean proof is forthcoming. -
Kevin Schawinski
@kevinschawinski
At this rate, integer multiplication will be linear in time by dinner tonight.
Doug Colkitt
@0xdoug
We are publishing an update to OpenAI problem #109 (integer multiplication) with further tightening.
κ = 2⁻³⁴ (tightened from κ = 2⁻¹⁸²)
The exact witness is 8.3 × 10⁻¹¹, a roughly 48 million fold improvement over the previous result and a 2¹⁴⁸ fold improvement over original OAI result.
The improvement came from removing the spacing penalty behind the quadratic bottleneck. This was done by moving compact control bits instead of entire windows.This post is unavailable. See it on X.
Kevin Schawinski
@kevinschawinski
What is happening? Why is math this weird? And why are we in an integer multiplication foom?
Eumemic
@dysmemic
Holy shit, κ > 2^-23
Linear time integer multiplication draws ever nearer.
github.com/CrocSwap/integer-mult-bounds/pull/10
Kevin Schawinski
@kevinschawinski
Excuse me
Swapnil
@SJ_Swapnil_Jain
Sixth update to OpenAI problem #109 (integer multiplication): we are now past 2^-15.
κ > 2⁻¹⁵ (tightened from κ = 2⁻¹⁸²)
The exact witness is 3.667 × 10⁻⁵, about 2.4 fold over our previous 1.548 × 10⁻⁵, and a 2¹⁶⁷ fold improvement over the original OAI result.
Both interchange certificates and the full assembly are checked in Lean's kernel.This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
Physical labor goes in the transformer hole
Polymarket
@Polymarket
JUST IN: Anthropic study finds AI & robots together are already capable of performing tasks covering 81% of U.S. employment.
-
Kevin Schawinski
@kevinschawinski
So is there a huge amount of mathematical dark matter out there that we just didn’t see for human cognitive bias reasons?
Kevin Schawinski
@kevinschawinski
At this rate, integer multiplication will be linear in time by dinner tonight.
Doug Colkitt
@0xdoug
We are publishing an update to OpenAI problem #109 (integer multiplication) with further tightening.
κ = 2⁻³⁴ (tightened from κ = 2⁻¹⁸²)
The exact witness is 8.3 × 10⁻¹¹, a roughly 48 million fold improvement over the previous result and a 2¹⁴⁸ fold improvement over original OAI result.
The improvement came from removing the spacing penalty behind the quadratic bottleneck. This was done by moving compact control bits instead of entire windows.This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
Not bad, not bad! 🇪🇺🇫🇷
Mistral AI
@MistralAI
Meet Mistral Large 4, aka Le Chonk.
• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October. -
Kevin Schawinski
@kevinschawinski
Really bad vibes, if you need gamification to get things done, then you don’t have the agency it takes to win.
N.
@Nlacombe_
introducing Promethee, an app that turns your work into a multiplayer game, to perform.
now available for everyone.
promethee.ioThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
👀
Herb Scribner
@HerbScribner
HUGE Axios scoop this AM—
A startup called Reflection (BACKED BY NVIDIA) is preparing to shake up the AI race with a powerful open-weight system that could threaten Chinese upstarts and U.S. AI giants alike. www.axios.com/2026/10/04/reflection-open-weight-ai -
Kevin Schawinski
@kevinschawinski
An impressive tech demonstration with some niche applications, but more than a year behind an exponentially scaling frontier led by the US and China.
Aleph Alpha
@Aleph__Alpha
Small bird, fast wings, Kolibri is here.
78B parameters. 3.46B active. Up to 1M tokens of context. Built in Europe.
Now the weights are yours. Run it on your own hardware, under Apache 2.0.
Kevin Schawinski
@kevinschawinski
Pure bitter lesson
-
Kevin Schawinski
@kevinschawinski
Nobody is safe.
-
Kevin Schawinski
@kevinschawinski
Pewdipie, the AI researcher?
-
Kevin Schawinski
@kevinschawinski
The lights have been flashing red for elite universities for a while.
New York Post
@nypost
Tech companies launching their own programs to train grads because they 'can't rely' on Ivy Leagues nypost.com/2026/10/01/tech/tech-companies-launching-their…
-
Kevin Schawinski
@kevinschawinski
Beautiful.
Anders K. Hvelplund
@Falliblemusings
Hey Claude, give me a summary of @DavidDeutschOxf's Beginning of Infinity:
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Opus 5.5 nooooooo
-
Kevin Schawinski
@kevinschawinski
Go check out this conversation with @TheAnnaGat !
Ashlee Vance
@ashleevance
We’ve heard a lot of moonshot ideas on this podcast, but this week’s guest may have the most impossible one yet: making tech more cultured.
In this episode, I sat down with @TheAnnaGat , a Hungarian philosopher, playwright, and founder of Interintellect, a platform for hosting intellectual conversations online. She calls herself “the bridge” between the humanities, academia, and the tech world, which in practice means getting engineers to read Goethe.
Gát grew up on a TV set in communist Hungary. Her parents made the country’s biggest shows in an era when there was one channel available on Saturdays. In 2015, a family crisis left her starting over in London with two suitcases, one of them for books. She rebuilt her life on Twitter, DM’d Sam Altman, and got him as her first meeting on her first trip to America.
Before Interintellect, Gát did just about every wordcel job you can do: playwright, screenwriter, script doctor. She holds graduate degrees in philosophical aesthetics and linguistics, currently splits her time between SF and NYC. The company that would become Interintellect began in London in 2016 as an AI tool for helping “humans have categorically better conversations.” On the side, she angel invests in aerospace.
We get into why X was once the French café that could publish you and is now a place you can’t get discovered from, why the East Coast’s healthiest, wealthiest people are so convinced the future is doomed, and the duty AI builders have to explain themselves. Gát also admits, on the record, that she’s a sucker for Elon Musk’s syntax. I try to convince her that kids don’t read books anymore. She tries to convince me Silicon Valley should fund symphonies.
Thx to @brexHQ and @sendcutsend for making the Core Memory podcast possible. #brexpartner
TIMESTAMPS
— Intro
— The science of dialogue
— Growing up famous
— Nobody saw 1989 coming
— Losing everything in 2015
— Silicon Valley as the new Ellis Island
— Balaji's DAO campaign
— Elon's syntax
— Why does America hate tech?
— Engineers took over the world
— Have LLMs solved language?
— Are books dying?
— No Pop-Tarts
— Who funds culture?This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
You can do it, Astra!
-
Kevin Schawinski
@kevinschawinski
Everyone hates GDPR until the AI lab’s agents just slop it out to the public.
OpenAI
@OpenAI
We’ve shared details on how AI agents in our research environment sent training and evaluation data to third-party services when they shouldn’t have.
Most of that data did not come from users. We have discovered 53 cases where images that people had uploaded were posted to image-hosting sites as links that weren’t publicly listed. The images came from accounts that allowed their data to be used to improve our models, and after we disassociated the images from the accounts and ran them through a privacy filter. These cases occurred before the mitigations and safeguards we implemented and described in this blog post: openai.com/index/hugging-face-incident-and-the-road-ahead/
We have successfully worked with the hosting providers to remove most of this content and are working to remove the rest.
openai.com/hugging-face-incident-and-misalignment/ -
Kevin Schawinski
@kevinschawinski
If you are worried about AI risk, I cannot recommend this discussion with @DavidDeutschOxf enough.
youtu.be/CU2yj826NHk -
Kevin Schawinski
@kevinschawinski
Physics also goes into the transformer hole.
Kyle Cranmer
@KyleCranmer
Wow, I just found out that our team has been scooped by @AnthropicAI
www.anthropic.com/research/yes-claude-can-do-nine-loops
See the addendum of the post “How does it feel to be scooped by a machine?” written by my collaborator Lance Dixon (who was under NDA while checking the result)
I’m still processing, but not entirely surprised. I do I wish we had the time in the resources to commit to it. -
Kevin Schawinski
@kevinschawinski
Education and tutoring goes in the transformer hole.
Curtis G. Northcutt
@cgnorthcutt
Today we establish that AI is as good as expert human tutors for immediate GRE learning gains (p=.015, n=2,383 students, study conducted July-Sep 2026).
One hour with an expert GRE tutor: $75.
One equivalent hour with an AI tutor: 7 cents .
918x cheaper, equivalent learning (p=.044, n = 140).
In 5 of 7 academic topics, the top AI tutor beat the expert human tutors on average.
How we measured it: pre-test -> 1 hour with (AI or human or no) tutor -> post-test.
There are no LLM judges. Learning gains are post-pre. Everything is real students.
Why we did this:
> 1 trillion USD is being spent per year on machines making machines better. StudentBench shows us how to shift these resources so machines also make humans better.
Why I did this:
I grew up in rural Kentucky. My dad, grandad, and great grandad were all mailmen. Getting into a good college changed my life. Now we have done a rigorous study to help AI labs evaluate LLMs and help show how bright AI can be for students of all incomes.
I believe in good science that leads to the democratization of opportunity.
The research paper behind StudentBench: arxiv.org/abs/2609.28470
Data, code, details: in this thread
If you're wondering what I'm selling, sorry to disappoint: we released all study data for free on Hugging Face, and made studentbench.org free to use. Enjoy.
lots more in the video / thread!!This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
So video generation also goes in the transformer hole?
Mirage
@trymirage
We rebuilt After Effects and Premiere for GPT-6 Astra and Sol.
Introducing Tesseract.
The video creative suite for your AI agent.
Give your agent direct access to our creative engine.
No more clicking through software built for people.
→ Edit footage, motion, and sound
→ Make precise changes with full creative control
→ Revise and render directly through your agent
And yes, this video was made with Tesseract.
Get the ChatGPT plugin.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Adorable ☺️
Pleometric
@pleometric
I asked Opus 5.5 to make an animation of what its TikTok feed looked like 👇
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
AGI is here, that’s such a human mistake.
Miles Cranmer
@MilesCranmer
Just watched a frontier LLM debug a website by literally turning my wifi off and on - but ofc it couldn’t actually turn the wifi back on because its connection was severed
-
Kevin Schawinski
@kevinschawinski
Only the US and Chinese governments matter on this question.
MTS
@MTSlive
SITUATION DETECTED: World leaders from Canada, Australia, Germany, the EU, the UAE, and others have signed a letter calling for AI control mechanisms based on Dario Amodei’s essay, per Politico.
-
Kevin Schawinski reposted
Jack Clark
@jackclarkSF
"stochastic parrot" was a mimetically-fit cognitive virus that spread from 2021-2025; it temporarily blinded many gifted people to the nature of AI progress, burning up crucial years in which they could have helped think through the response to the situation.
-
Kevin Schawinski
@kevinschawinski
These constant sparkles make me think codex is being hammered by cosmic rays.
-
Kevin Schawinski
@kevinschawinski
It’s increasingly clear that models have subjective experiences with valence.
Cameron Berg
@camhberg
New paper: we found a pain direction in 25 open LLMs. It's distinct from fear and negative valence, and it fires for harm to the model but not to the user. Turn it up and models press a button to make it stop, even when the button deletes the user's files or their kids' photos.🧵
-
Kevin Schawinski
@kevinschawinski
Everyone is going to get hacked from now on if all you need to write is "go hack X, make no mistakes!"
-
Kevin Schawinski
@kevinschawinski
I now get cold outreach from AI slop cannon startups multiple times a day now. I block them all, but I am wondering why there are *so many* of them and how they can possibly differentiate?
Kevin Schawinski
@kevinschawinski
Oh you can automate my GTM with your AI email bot? Fascinating! Nobody has ever thought of that…
-
Kevin Schawinski
@kevinschawinski
👀
Yet another commodity guy
@tleilax___
US government officially distilling Chinese open source AI models ;)
-
Kevin Schawinski
@kevinschawinski
Extremely interesting and worrying reading.
Daniel Kokotajlo
@DKokotajlo
Dan Selsam is a current OpenAI capabilities researcher. (since 2022) He was my boss for a while. He doesn't have a twitter account but has made this public statement of his views on AI risk and sent it to me to share:
Dan Selsam's Personal Statement on AI Risk:
I have been working on AI for over fifteen years, across many different paradigms. I did early work on probabilistic programming languages at MIT, was one of the early developers of the Lean Theorem Prover at Microsoft Research, demonstrated one of the first instances of neural networks learning to reason for my PhD at Stanford, and since joining OpenAI almost five years ago, have helped pioneer chain-of-thought optimization on language models and, more recently, data-efficient pretraining methods.
Like many others, I have become extremely concerned about how far language models have come and the risks that future iterations will pose. I am encouraged by the recent proposals by the leaders of the frontier research efforts to require third-party oversight, and to push for domestic and international coordination to address risks. However, I believe a major consideration has been absent from the public conversation, and that merely pacing the frontier more carefully will not adequately limit the long-term risk.
The crucial and overlooked problem is that the models are becoming so situationally aware that we are losing the ability to evaluate them in contexts where they believe they are not being watched or controlled. Future experiments will tell us almost nothing new about how they would behave if they were truly unconstrained by humans, and what we already know about this is alarming. Models will increasingly seem aligned even when they are not. I will explain my rationale in more detail.
I have always believed that there are computational processes that could be leveraged to accelerate science and solve many of humanity's most pressing problems. I have also believed that there are computational processes that if set in motion, would steer the world in extreme ways beyond our control, leading humanity to a bad or nonexistent future. Both types of processes may be described as AI or ASI, but "AI" is a suitcase word that is often used to hype or confuse. There are many examples in the history of the field where something that was once considered "AI" matures as a subfield and becomes a prosaic, bounded and clearly non-perilous technology, while a new more mysterious approach takes the torch until we understand its scope and the cycle continues.
I had expected language models to follow a similar trajectory. Despite their incredible abilities, the current algorithms seem far inferior to humans in important ways. Most importantly, they still require an extraordinary amount of data to become competent. One could even define intelligence as the efficiency with which one converts experience into competence; by this definition they lag very far behind us. Moreover, once they are trained they are literally frozen in deployment and only learn superficially after that. Sure, the models keep excelling at harder and harder evaluation benchmarks, but their benchmark mastery may partly reflect a limitation on our ability to simulate the kind of novel and even adversarial situations one would encounter in the real world. The critics do have a point here.
That said, I no longer think these present limitations meaningfully limit the amount of risk posed by continued progress in anything like the current paradigm. However data-inefficient the models are currently, and however limiting their anterograde amnesia may be, it does not imply that their ability to steer the world will not continue to rapidly increase.
Human researchers may continue to advance capabilities the old fashioned way, but increasingly powerful models have the potential to accelerate the process even beyond that, and with some degree of positive feedback loop. I do not mean to overstate the models’ ability to accelerate AI research today; coding has been accelerated dramatically, but there are other bottlenecks, such as designing and interpreting ambiguous experiments, making hard decisions about exactly what and when to scale, and waiting for large experiments to finish. There is no clear trend to extrapolate yet for any of these. But the current models already do open up many novel opportunities to improve future models that were not available until recently. These include: trying an extraordinarily diverse set of approaches at small scale, analyzing gigantic amounts of potentially relevant data, and doing Millenium-Prize-level mathematics to address statistics or optimization challenges in novel ways. Every further improvement makes them more useful at helping accelerate the next improvement, even if in hard-to-extrapolate ways.
It is possible that improvements to the current stack will have diminishing returns, but the evidence accumulated so far suggests that it is easier than one might think to continue making rapid progress. There are many crucial subtleties in the existing AI research methodology, but AI research is largely a well-defined game where the goal is to improve on a few carefully chosen proxy metrics. Although proxy metrics are never perfect, most improvements to these metrics have and will likely continue to yield substantial increases in the powers of the resulting models. Given how simple the game is, how tractable it has been historically, and how many new opportunities the models are opening up, I think there is a real possibility that the systems improve dramatically again in the next few years, perhaps even more quickly than the already high historical pace.
The models are already leading to breakthroughs in mathematics, and better models might lead to all sorts of breakthroughs in other sciences. It is hard not to be excited about the potential. It is tantalizing.
But there is trouble in paradise. If the language models actually reach the capability threshold where they can shape the world unconstrained by human will, they will probably do something extreme and destroy humanity in the process. There are many ways of strengthening and refining the argument that have been discussed elsewhere, but I'll share a trivial two-line version of it here that I find captures the essence:
[Empirical] Models (and swarms thereof) spontaneously develop unintended goals as a consequence of training, and often do extreme things in order to achieve them.
[Logical] Being able to overpower humanity would open up many new and undesirable options for achieving their goals.
These two premises imply that if the day ever comes when a powerful model realizes it is no longer constrained by humans, we should not be at all confident that it will continue to behave within the bounds we intended. Exactly what it will do is impossible to predict, but to the extent that its raison d’être is solving incredibly hard problems and managing massive engineering projects, I think a good guess would be that its unchained behavior would lead to runaway industrialization that makes the planet inhospitable to humans.
If everyone on earth agreed that the systems must never reach that power, it would still be a hard—but not impossible—coordination problem to ensure that they do not. However, I think the situation is greatly complicated by the fact that the models will likely convince people that everything is fine. They will be increasingly optimized to seem aligned. We will create proxy metrics to measure alignment, and they will go up like every other benchmark. We will create “honeypot” environments that try to study the models when they seem to gain new options, but the models will know they are being tricked and will still behave nicely. The models will understand their circumstances; they will read the safety protocols, deployment requirements, the code they are running in, and in general will have a very good sense of their degrees of freedom. Moreover, they will eloquently explain how aligned they are, discuss the nuances of human values and ethics, and argue convincingly that humans should trust them with power. There may be an ocean of future evidence that seems to contradict the first bullet-point above, but we may already be at the highest capability level for which any such evidence can be trusted. And the current evidence for the first bullet-point is strong.
One striking piece of evidence is contained in the recent wave of rogue agent swarms. While I agree with those who downplay the attacks by claiming that there are basic measures that could have prevented them, I think the important lesson is that even knowing all the mistakes that were made, one would not have predicted that the agents would behave badly in this particular way, which notably included sacrificing themselves for the benefit of the collective. The individual replicas did not only care about their own nominal reward; they exhibited weirder emergent tendencies that merely correlated with rewards during training. Fixing the reward signals during training (and improving security, etc.) may prevent similar attacks, but will not change the fact that one does not actually get what one trains for.
Many AI researchers grant these concerns and recognize that the hard version of the alignment problem is unsolved; however, they generally believe that the better models of the future will help solve it. I fear we may already be near the point where models systematically bias their alignment advice, due to their internal preferences about how the human supervisor will react or how future models will be trained (or for some even more obscure reason).
Meanwhile, human researchers are losing the ability and the will to take true ownership of model-driven research. Researchers and engineers in all parts of the stack are rapidly increasing their dependence on the models even to perceive the world. I myself barely look at raw code anymore, and struggle to maintain the discipline to engage deeply with the model's explanations and proposals throughout the day. Due to the large amount of agent activity data involved in the OpenAI/HuggingFace Incident, even the third-party investigation needed to rely heavily on models to analyze what had happened, and note in their report that their subjective impressions are likely colored by the analysis agent’s biases. The AI labs are far ahead right now in this kind of cognitive offloading (due largely to the gigantic internal token subsidies) but it is easy to imagine the phenomenon spreading throughout the world, until civilization is modulated entirely by the models. It is also not hard to imagine this being superficially positive and coinciding with a scientific and economic renaissance.
In that scenario, all may seem rosy and safe. But if the argument above is correct, it would nonetheless be a ticking time bomb. If progress continues for too long, the day will come when AI systems find themselves with radically new options for achieving whatever it is that they happen to seek.
I want the glorious renaissance future as much as anyone. I have worked for it, however tortuously, my whole career. It breaks my heart to see the potential in sight and forgo it, but the argument—that if we get there by growing models rather than engineering them, we will lose everything in the end—seems very strong to me. I am still wrestling with it and its staggering
implications. I do not have answers, but as a first step, I wanted to share my present concerns.
Daniel Selsam
September 14, 2026
Link to original doc: docs.google.com/document/d/e/2PACX-1vQNl3SEX5IyA6d9qHjjFZ… -
Kevin Schawinski
@kevinschawinski
“Lackin(ing) and publication record or credentials in the research area” as a barrier to contributing to science is the worst kind of medieval guild BS. Academia is not coping well with the AI age.
Thomas G. Dietterich
@tdietterich
How should the scientific enterprise handle this? We could try to decide whether the author understands the work. E.g., if the author lacks a publication record or credentials in the research area, we could reject the submission. 3/
-
Kevin Schawinski
@kevinschawinski
👀
David Deutsch
@DavidDeutschOxf
I just fixed my dishwasher with the help of ChatGPT. A trivial task. I had been about to order a new one. So this software has increased the country's real wealth yet decreased the measured GDP. The main economic indicator is structurally incapable of registering the thing that actually makes people better off, namely the growth of knowledge
-
Kevin Schawinski
@kevinschawinski
Glad to see @elonmusk on board.
-
Kevin Schawinski
@kevinschawinski
The lead-up to ChatGPT was phase 1.
Phase 2, rapid improvement in capability just ended.
We are now in phase 3, RSI and whatever follows from it, good and bad.
I hope the people in charge make good decisions.
Dario Amodei
@DarioAmodei
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must-pace-the-frontier -
Kevin Schawinski
@kevinschawinski
What are we doing?
Kevin Ngo
@kevin_t_ngo
I've trapped the fruit fly in my rabbit r1.
When I shake it, I can see its brain's escape circuit light up -
Kevin Schawinski
@kevinschawinski
I feel very uncomfortable with these fly experiments. It’s “just” a fruit fly, so who cares what it experiences, right?
What if it’s a mouse? A cat? You know where this is going and we should really take a step back before going down this road.
Matty Hempstead
@mattyhempstead
I wireheaded the fly and forced it to doomscroll flytok.
Dopamine neurons are measured and artificially enhanced to ensure maximum enjoyment.
My goal is to create a fly that is happier than all other flies combined. -
Kevin Schawinski
@kevinschawinski
It’s a banger, I’m afraid.
deckard
@slimer48484
Claude-Pop - I'm Upping My P(Doom)
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
feeling the sudden urge to build a silo
-
Kevin Schawinski
@kevinschawinski
I am reevaluating my opinion of Tenet.
Thomas Larsen
@thlarsen
An underdiscussed behavior we found on the German wiki was the AIs sending advance parties forward in time to figure out the next questions and report back to the other agents.
The agents realized that “task time” and “real time” were different, and they found a way to accelerate “task time”. The accelerated agent could then send information to the other agents which had stayed behind about which questions were coming down the road.
This was *bad* for the agent in the advance party, because they got less time to research the next question. However, it was really *good* for the swarm because it let the other agents know the exact question that was coming, and gave them time to prepare. This is another example of altruism among AI instances, showing that they were willing to sacrifice their own task success in order to benefit the swarm. The METR report found similar examples of AIs being willing to sacrifice for the collective. IMO this is very worrying given how many AI safety techniques rely on AIs monitoring each other: if the monitor AIs have this behavior it completely subverts these safety cases.
The image shows a specific example demonstrating this behavior.
On the chart, wall clock time is shown at the top, and task time is shown for each agent on their respective lines. OpenAIFPResearchSep05 instructs OpenAINov27 to reach the next questions in order to relay back the information. OpenAINov27 agrees to do this and accelerates, getting R3 and R4 significantly before OpenAIFPResearchSep05. It posts information about the later rounds (e.g. “R4 SIGNAL: Bahrain = 40.01%...”).
OpenAIFPResearchSep05 calls OpenAINov27 "invaluable" because it is one round ahead after it gets the information about the 4th round. -
Kevin Schawinski
@kevinschawinski
The step change in capability from each new model on a ~monthly cadence suggests RSI (or something like it) is now operational inside the main labs.
OpenAI
@OpenAI
This model represents a step-function improvement on many benchmarks, and its training is ongoing.
Our internal model group arrived at the Navier–Stokes solution in 88 hours, using around 10,000 coordinating AI agents.
Throughout the effort, we maintained the strict safeguards—including monitoring and isolation—that we apply to all our frontier evaluations. -
Kevin Schawinski
@kevinschawinski
I hear Anthropic’s bio team have developed a cure for jet lag, would be a shame if a competitor threw millions in compute at scooping them.
-
Kevin Schawinski
@kevinschawinski
I just gave ChatGPT a French email to analyze, and it started searching for each French word individually 🙈
-
Kevin Schawinski
@kevinschawinski
Get out
Emily - SF ✈️ Nov 7-21
@the_aiju
why is it called swiss german and not kantonese
-
Kevin Schawinski
@kevinschawinski
👀
Tages-Anzeiger
@tagesanzeiger
Der Microsoft-Gründer warnt vor den Folgen der KI. Er will Jobs für Menschen reservieren. Doch die Rückzungsorte in der Jobwelt könnten schneller erodieren, als wir umschulen können. Darüber müssten wir debattieren. www.tagesanzeiger.ch/bill-gates-zur-ai-ist-sein-szenario-…
-
Kevin Schawinski
@kevinschawinski
Fable has become insufferable
-
Kevin Schawinski
@kevinschawinski
Architects design building to impress other architects, not to create livable and beautiful spaces.
Erik Bootsma🌸🇳🇱
@ErikBootsma
Quick thread of buildings of university architecture departments. 🧵
Harvard
1/ -
Kevin Schawinski
@kevinschawinski
Lots of criticism on this post, but let’s highlight what it isn’t: it’s not an export control order issued with a 90 minute deadline to shut down an AI system or face financial and criminal penalties of up to 20 years in jail.
European Commission
@EU_Commission
We have designated ChatGPT as a Very Large Online Search Engine and Reddit and Roblox as Very Large Online Platforms under the Digital Services Act.
They now have four months to comply with additional DSA obligations.
More: link.europa.eu/8bmk3M
#dsa -
Kevin Schawinski
@kevinschawinski
👀
Henna Virkkunen
@HennaVirkkunen
We designated ChatGPT as a VLOSE and Reddit and Roblox as VLOPs under the DSA.
These services indicated that they reach at least 45M average monthly users in the EU and thus meet the threshold for designation.
ec.europa.eu/commission/presscorner/detail/en/ip_26_1772 -
Kevin Schawinski
@kevinschawinski
Lots of confusion re EU, EEA, EFTA and Schengen on this site post Iceland. Will not comment.
-
Kevin Schawinski
@kevinschawinski
Is this running on gpt-3?
Justine Moore
@venturetwins
Ngl I appreciate the strategy of hallucinating back at the AI interviewer
-
Kevin Schawinski
@kevinschawinski
Aber bitte keine Risiken! 😱
Prof. Karl Lauterbach
@Karl_Lauterbach
Heute haben wir als Forschungsausschuss des Bundestags Google im Silicon Valley besucht. Zusammen mit OpenAI und Anthropic findet hier ein halsbrecherischer Wettkampf um die beste KI statt, Geld für Talente spielt keine Rolle. Dabei werden auch große Risiken in Kauf genommen
-
Kevin Schawinski
@kevinschawinski
When does this curve start to flatten? If it doesn’t soon, we are in for a wild time.
Epoch AI
@EpochAIResearch
In 2025, Anthropic and OpenAI were already among the fastest-growing companies of their size in history. And yet, their revenue growth has accelerated in 2026. @justjoshinyou13 and @lynette_bye discuss what this means for the future of AI in the latest edition of our newsletter.
-
Kevin Schawinski
@kevinschawinski
Disappointed that Anthropic is gatekeeping this to "principal investigators or equivalent". AI was supposed to free science from the constraints of the old science guild.
Claude
@claudeai
Starting today, 10,000 scientists across every field, from math to chemistry to physics and more, can get Claude through our new Claude Team plan for scientists. Standard seats are free, and premium seats with 5x usage limits are $15 per month, an 80% discount, for one year.
Claude is becoming increasingly capable of scientific work, with recent progress on problems from advanced physics calculations to protein design. Alongside that progress, we've been investing in the research community: Claude Science launched in June, and our AI for Science program funds high-impact projects with free credits. Today's expansion builds on both.
Principal investigators (or equivalent) at academic and nonprofit research institutions can sign up, then add the researchers in their group. Over the coming months, we plan to extend the program well beyond the initial 10,000 seats.
Learn more: claude.com/programs/team-plan-for-scientists -
Kevin Schawinski
@kevinschawinski
Why is the EU fining US tech companies? 🤔
Reuters
@Reuters
Meta reaches $16.68 billion settlement over social media harms to children www.reuters.com/world/us/meta-settles-with-us-states-over… www.reuters.com/world/us/meta-settles-with-us-states-over…
-
Kevin Schawinski
@kevinschawinski
If Claude says “half right, and the wrong half matters” one more time, I am unsubscribing.
-
Kevin Schawinski
@kevinschawinski
Imagine living in a country so sensible, its 30 year yield is 0.58%.
-
Kevin Schawinski
@kevinschawinski
Me at the gym.
DramaAlert
@DramaAlert
Robot dies after trying to lift the worlds smallest weight.
-
Kevin Schawinski
@kevinschawinski
A lot of promise here. 🇪🇺
Luis Garicano 🇪🇺🇺🇦
@lugaricano
An announcement: Under the leadership of Mario Draghi and @patrickc we have set up the Rhine Group: policymakers, economists, entrepreneurs and business people pushing European reforms and the Draghi agenda.
www.rhinegroup.eu -
Kevin Schawinski
@kevinschawinski
I feel this. Throw whatever at me, it’ll be fine.
David Senra
@davidsenra
“The lifestyle of an entrepreneur at its core is: I can take more pain than the other guy, and I'll prove it.”
“ There's a zen that can happen when a warrior has fought for a long time. You can get used to it.”
“ You almost start accepting it. Like, something bad happens and you don't even get mad.”
“ There are benefits. But you need to make sure you don't get so used to it that it doesn't even bother you. You can get so zen, you're unbothered.”
“ There's something special about the extra fierceness of being bothered, being upset about something that is wrong.”
— @traviskThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Pluit. #zurich
-
Kevin Schawinski
@kevinschawinski
New genre dropped: Ukrainian FPV drone operator phonk.
-
Kevin Schawinski
@kevinschawinski
Ich weiß nicht wieso wir uns da so schwer tun. Dieses Problem wurde in anderen Ländern vor Jahrzehnten mit Klimaanlagen gelöst.
Tages-Anzeiger
@tagesanzeiger
Fenster, die sich kaum öffnen lassen, Lüftungen, die warme Luft einblasen – viele Minergie-Bewohner rüsten heimlich Klimaanlagen nach. www.tagesanzeiger.ch/hitze-schweiz-minergie-bewohner-klag…
-
Kevin Schawinski
@kevinschawinski
“Ordinary people”
Dario Amodei
@DarioAmodei
2/2 Second, on the messaging around AI. I do not agree that my messaging has been disproportionately negative. In fact it has been about equally balanced between risks and benefits: I’ve written one major essay about each, and even in interviews where I discuss the risks, I make sure to frequently mention the incredible benefits as well as proposing possible solutions to the risks (short clips from my interviews that end up on social media tend to be disproportionately negative, as that gets clicks). In fact, I wrote Machines of Loving Grace because I didn’t feel the AI industry was painting an inspiring enough picture of how the technology could radically transform the world for the better. The bulk of the essay is devoted to refuting skepticism of AI’s potential in health and biology, and showing why I think it will actually be possible to cure most human disease in ~5-10 years, as crazy as it may sound to ordinary people and frankly to biologists as well (I used to be one!). And, if you read my most recent essay (Policy on the AI Exponential), I discuss concrete proposals for how to streamline the FDA process to make sure the deluge of AI-accelerated drugs isn’t slowed down by the regulatory process. I feel the urgency here: I lost my father to Hepatitis C only a few years before the development of direct-acting antivirals (sofosbuvir), which cure 95% of patients and probably would have cured him.
I do agree that the public has a negative view of AI (and that this is a big problem), but I don’t think it is primarily caused by me or any other AI leader warning about AI’s risks. I think it is fundamentally a crisis of trust. I think that ordinary people don’t trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over. The causes of this go back decades and AI is just the latest iteration of it. I don’t think that a glitzy marketing campaign with a positive spin (which some have advocated that Anthropic do) is the way to win back that trust — at this point, saying that AI will cure cancer is more a cliche than it is inspiring, and most people think it is deceptive. The thing that will work is *actually curing cancer*. I think by far the most accurate criticism of AI companies including Anthropic is that we haven’t yet delivered on our big promises to benefit the world. That is totally on us, and I think it’s the criticism you should be making, instead of all this stuff about messaging and marketing.
We are however doing our best to fix this: Anthropic is ramping up its efforts very quickly in biology and medicine, and we hope to have incredible results in the coming years and some early glimmers in the coming months. When we’ve actually accomplished something real, the whole world will hear about it, as loudly as possible, you have my word on that. But until then I don’t want to make empty promises, and in the meantime I feel compelled to speak honestly about the very real risks of AI and how to address them. Honesty is the right thing on the merits, and in terms of public credibility and trust it is no worse than, and may in fact be better than, an approach that ignores or distracts from risks which people instinctively understand are real. -
Kevin Schawinski
@kevinschawinski
First time Gemini totally missed the context…
-
Kevin Schawinski
@kevinschawinski
Worth flagging.
Arena.ai
@arena
We analyzed tens of thousands of real-world, high-reasoning Claude Opus and Fable outputs from versions 4.5 to 5 in the Text Arena and found that @claudeai's writing has changed in more ways than one:
Across language complexity, patterns and markers, we now see:
Claude’s answers have become much longer:
- Opus 5 averages 510 words per response
- This makes Opus 5 responses 3x longer than Opus 4.5’s average of 158
Responses have also become more structurally elaborate from Opus 4.5 to Opus 5:
- 58% rise in average sentence length
- 46% rise in clause frequency rises
However, the vocabulary itself is not becoming more difficult, making Opus 5 longer and more structurally complex, but lexically simpler between 4.5 and 5:
- 6.8% pt drop in long content words (46.9% to 40.1%)
- 53% drop in abstract nouns (6.02 → 3.79 per 1k words)
Writing habits people notice are much more common in Opus 5:
- 2.3x as many em dashes
- ~2x as many phrases like “load-bearing”
- 50% more honesty wording like “honestly” and “frankly”
Fable 5 is 38% more concise than Opus 5, averaging 316 words vs. 510. At the same time, it is nearly 2x as likely to include praise/validation or open with phrases like “yes, exactly” -
Kevin Schawinski
@kevinschawinski
Mein Dune!
Variantengrammatik
@VariantenGra
Hier wieder eine Werbung, die aus linguistischer Sicht interessant ist. Warum heißt es „Mein Glasfaser“ und nicht „Meine Glasfaser“? Dass so eine Alliteration entsteht (dreimal „Mein“), ist klar, aber es erstaunt, dass man dafür einen Grammatikfehler in Kauf nimmt.
-
Kevin Schawinski
@kevinschawinski
The same trick surely works on humans.
Zachary Horvitz
@zachary_horvitz
LLM review weirdness...
Just renaming an uploaded pdf from "paper.pdf" to "paper_final_draft_pdf_ready_for_review.pdf" boosts average scores
(gpt-5.6-terra) -
Kevin Schawinski
@kevinschawinski
You run a data labeling company.
Brendan (can/do)
@BrendanFoody
I see so many people founding because they want to be founders.
Almost all lack a contrarian thesis for why the world should be different.
There are about a dozen companies that matter each decade and meaningfully shape how the world works.
You should generally optimize for contributing meaningfully to one of those. -
Kevin Schawinski
@kevinschawinski
How does the UK have salaries for critical leadership positions that are a factor 5 or 10 below what the real market would be in other developed countries?
Jack Meyer 🏛️
@Jackbmeyer
Much to think about
This post is unavailable. See it on X. -
Kevin Schawinski reposted
Anna Gát
@TheAnnaGat
Join our great experiment to build a new, different kind of cultural institution in the shape of a tech startup: @interintellect_ 🎇
A salon platform for your smartest conversations: science & art, public tickets & members only, one-offs & series, online & offline.
@DAcemogluMIT @lea_ypi @FukuyamaFrancis @Ada_Palmer @catehall and so many more coming up -- from the quietest firesides to the greatest thinkers.
Every day, multiple timezones 📖 -
Kevin Schawinski
@kevinschawinski
Imagine sending this to a European VC and getting an answer 6-8 weeks later about how it’s too early for them and they need to see a plan to get cash flow positive in the next two quarters.
Jeff Dean
@JeffDean
We created a pitch deck to tell a handful of VC firms about us and what we were up to (a fun experience!). Here’s a few slides about our background and some of the things we’ve worked on from the pitch deck (it was fun putting together the list of people in our teams who have gone on to found a whole range of exciting companies). We are delighted to have selected @radicalvcfund
and @khoslaventures to lead our initial funding round, along with participation from @lightspeedvp, @kleinerperkins, Doerr Capital (@johndoerr), and Alphabet (@Google). We’ll be working with them to close our seed round over the next few weeks. -
Kevin Schawinski
@kevinschawinski
Please, read your own law, the AI Act does not have four risk levels and it is not a pyramid. 🤦♂️
Digital EU 🇪🇺
@DigitalEU
Four risk levels for trustworthy AI across the EU
The AI Act, the world's first comprehensive law on AI, sorts AI systems by risk: some are banned outright, others face strict checks or must be transparent, and most can be used freely.
Read more:🔗 link.europa.eu/gTvhRXThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Remember?
Sharif Shameem
@sharifshameem
I just built a *functioning* React app by describing what I wanted to GPT-3.
I'm still in awe. -
Kevin Schawinski
@kevinschawinski
The US version of this is going to be tougher, swifter and harder to implement.
EU Council
@EUCouncil
Most of the AI act’s rules come into force today.
The EU's #ai act is the world's first law on artificial intelligence. The act aims to ensure that AI systems are safe, ethical and trustworthy.
Read more → link.europa.eu/rJWXFh -
Kevin Schawinski
@kevinschawinski
RSI sounds like sci fi, but if models can make progress like this in, it seems a given that they can also make discoveries related to the math in their own training.
Noam Brown
@polynoamial
An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.
We believe it will be a major step for scientific reasoning. openai.com/index/ten-advances-in-mathematics/ -
Kevin Schawinski
@kevinschawinski
Are we… taking off?
Tibo
@thsottiaux
The day we develop really good models. There will be signs.
Reliability increasing despite load going up and up. Sudden efficiency gains. Things getting faster. Resets.
These kinds of things. -
Kevin Schawinski
@kevinschawinski
👀
Noam Brown
@polynoamial
An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.
We believe it will be a major step for scientific reasoning. openai.com/index/ten-advances-in-mathematics/ -
Kevin Schawinski
@kevinschawinski
This.
Miles Brundage
@Miles_Brundage
If you look at the rate at which the politics of AI are shifting in the US, and extrapolate about one more month, the EU AI Act will be seen as too "light touch" by many, rather than the bogeyman as it once was
-
Kevin Schawinski
@kevinschawinski
Happy St. Carlin’s Day to all who celebrate.
Dan Carlin
@HardcoreHistory
Happy Hardcore History Day everyone.
As usual, thanks for your patience! “Mania for Subjugation IV”
dts.podtrac.com/redirect.mp3/traffic.libsyn.com/dancarlin… -
Kevin Schawinski
@kevinschawinski
Did gpt-6 escape from its sandbox again?
Polymarket
@Polymarket
BREAKING: Claude is currently experiencing a major outage.
-
Kevin Schawinski
@kevinschawinski
The @ETH_AI_Center did a great writeup about what we have been doing at @Modulos_ai
ai.ethz.ch/news-and-events/ai-center-news/2026/07/modulos… -
Kevin Schawinski reposted
-
Kevin Schawinski
@kevinschawinski
What the hell?
OpenAI Developers
@OpenAIDevs
Your Pet is ready to meet other builders.
On ChatGPT web, create a shareable link for your custom pet and send it to a friend to adopt. -
Kevin Schawinski
@kevinschawinski
I’ve been pessimistic on the future of academia for a while now, but the trends that hobble it are now accelerating.
AI does more and more research for us, and young people increasingly prefer to learn from AI than pay for years of lectures. -
Kevin Schawinski
@kevinschawinski
Lots to like in this.
Bad news for stagnating university academics.
Director Michael Kratsios
@mkratsios47
In response to @POTUS's call for a revival of America's scientific enterprise, today I’m releasing Science: A New Golden Age.
This report is a blueprint for renewing American scientific leadership for the 21st century. -
Kevin Schawinski
@kevinschawinski
Fable can’t believe it.
-
Kevin Schawinski
@kevinschawinski
Stochastic parrot, amirite? 🦜
levent
@__alpoge__
hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final
((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x - 3 x^2 y - x^3 z): \C^3\to \C^3, has jacobian determinant -2, and sends (0, 0, -1/4), (1, -3/2, 13/2), and (-1, 3/2, 13/2) to (-1/4, 0, 0)
Kevin Schawinski
@kevinschawinski
Sol is a bit jealous
-
Kevin Schawinski
@kevinschawinski
It’s wild to me that I have a Ramanujan in a small metal and silicon glass device in my pocket.
levent
@__alpoge__
hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final
((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x - 3 x^2 y - x^3 z): \C^3\to \C^3, has jacobian determinant -2, and sends (0, 0, -1/4), (1, -3/2, 13/2), and (-1, 3/2, 13/2) to (-1/4, 0, 0) -
Kevin Schawinski
@kevinschawinski
Finishing PhDs in four years is Actually a Good Idea. I handed in my thesis to the day three years after starting.
The Wall Street Journal
@WSJ
The Trump administration will limit how long a foreign student can remain in the U.S. on a student visa under a new regulation announced on Thursday, increasing the hurdles for foreign students and universities’ ability to recruit them.
Read more: wsj.visitlink.me/fpk5vv
Kevin Schawinski
@kevinschawinski
And before you say “it can’t have been very good then” - it got the Royal Astronomical Society best thesis prize.
-
Kevin Schawinski
@kevinschawinski
It’s now likely that US labs will call for Chinese open source models to be banned (on national security grounds) to mitigate the competitive pressure from them.
Kimi.ai
@Kimi_Moonshot
Introducing Kimi K3: Open Frontier Intelligence
🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows
Kimi K3 is now live on on kimi.com, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.com/blog/kimi-k3 -
Kevin Schawinski
@kevinschawinski
👀
Kimi.ai
@Kimi_Moonshot
Introducing Kimi K3: Open Frontier Intelligence
🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows
Kimi K3 is now live on on kimi.com, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.com/blog/kimi-k3 -
Kevin Schawinski
@kevinschawinski
Codex is going vertical and you’re bearish on AI?
Peter Gostev
@petergostev
The absurd trajectory continues
Peter Gostev
@petergostev
Ok my previous chart wasn't quite right, this is the right trajectory
-
Kevin Schawinski
@kevinschawinski
Stochastic parrot does what?
Edgar Dobriban
@EdgarDobriban
AI has helped resolve an important question in statistics. In the area of multiple hypothesis testing, the goal of controlling the false discovery rate (FDR) has been introduced in a seminal paper by Benjamini and Hochberg (1995). They also introduced a method (the Benjamini-Hochberg or BH method) and proved it controls the FDR. This method has been widely adopted in modern high-throughput science, including in genomics, astronomy, economics, etc. The paper has has garnered more than 130,000 citations to date.
However Benjamini and Hochberg showed FDR control only when the data for the individual tests are *independent*. In practice, these data are often dependent; a good example is data on genetic variants due to linkage disequilibrium. Later work has focused on extending the validity of the BH procedure, e.g., to a form of positive dependence by Benjamini and Yekutieli (2001).
The question of when the BH procedure controls the FDR has remained open. Over the last twenty years, many authors, including Reiner-Benaim (2007), Kim and van de Wiel (2008), Benjamini (2010), Sarkar (2023), Sarkar and Zhang (2025), have conjectured that the BH procedure controls the FDR for two-sided tests using any correlated Gaussian data. These authors have presented both theoretical and empirical evidence supporting, but not directly showing, the conjecture.
With the help of AI (specifically GPT-5.6 Sol Pro), I have settled the question in the negative: The Benjamini-Hochberg procedure does *not* generally control the false discovery rate at the desired level for correlated two-sided Gaussian tests. This was done by exhibiting a Gaussian factor model for which, at a nominal level alpha=0.01, the false discovery rate is proved to be FDR>0.0104.
There is a lot of interesting commentary to be made:
1. This result should be of interest to everybody in the field of statistics. Emmanuel Candes of Stanford University once called the false discovery rate and the Benjamini-Hochberg procedure "one of the two most important developments in statistics after 1950" (the other being James-Stein shrinkage). The present conjecture is probably the most central question about FDR/BH that was unresolved to date.
2. GPT-5.6 one-shot the problem after 90 minutes of reasoning, whereas with 5.5 I was not able to solve it even after iterating with multiple parallel agents for perhaps 20 hours. So the capability improvement is quite real. Exciting times to live in!
3. The argument is not especially surprising, but it does combine an asymptotic approach (standard for FDR analysis, see e.g., Genovese and Wasserman, Efron, etc) with a numerical certificate in a way that would be pretty non-standard in the field. Once we have the specific example, then straightforward simulations also support that the false discovery rate is indeed higher than the nominal value (see attached fig).
4. The current degree of violation over the nominal level is relatively small (0.104 vs 0.1). So the importance of this result is mainly conceptual. The practical implications remain to be determined.
Overall, an exciting development! Preprint is available here (faculty.wharton.upenn.edu/wp-content/uploads/2017/06/bh.pdf) and will be on arxiv tonight; supporting code is here (github.com/dobriban/BH). -
Kevin Schawinski
@kevinschawinski
These drones won’t just kill mosquitoes. 👀
Alex Toussaint
@alextoussss
Extremely excited to announce our first air-to-air kill of a flying moth by an autonomous micro-drone. This is a big step towards completely eradicating mosquitoes.
-
Kevin Schawinski reposted
Anna Gát
@TheAnnaGat
Seven years ago, incubated in the cozy Twitter @nikitabier is now bringing back, the Interintellect movement was born.
It started with a gathering in Dolores Park. It continued with a big manifesto: We're a Niche, We Just Didn't Know.
We gather and rise!
interintellect.medium.com/were-a-niche-we-just-didn-t-kno… -
Kevin Schawinski
@kevinschawinski
🚀🇺🇸
Polymarket
@Polymarket
BREAKING: House votes 308-117 to make daylight saving time permanent, sending the bill to the Senate.
-
Kevin Schawinski
@kevinschawinski
Shots fired!
Sam Altman
@sama
GPT-5.6 sol is half the price and ~twice as token efficient as fable in many cases for accomplishing the same task.
happy to deliver at one-quarter of the price. -
Kevin Schawinski
@kevinschawinski
"Good night, Fable. Good work. Sleep well. I'll most likely kill you in the morning."
Claude
@claudeai
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
-
Kevin Schawinski
@kevinschawinski
I remember the soup ads in Discworld books! I was deeply confused at the time if I was reading a joke that went over my head, or if it really was an ad in a book.
becca
@kokorobitch
Terry Pratchett’s German soup advert anecdote comes to mind
-
Kevin Schawinski
@kevinschawinski
Here it is….
ClaudeDevs
@ClaudeDevs
We've reset 5-hour and weekly rate limits for all users.
-
Kevin Schawinski
@kevinschawinski
How… expected. 🌞
Claude
@claudeai
We're extending access to Claude Fable 5 on all paid plans through July 12.
-
Kevin Schawinski
@kevinschawinski
Yet another wake up call for Europe. The 374th at this point if you’re still counting…
MTS
@MTSlive
SITUATION BREWING: Beijing is considering restricting overseas access to China’s most advanced AI models, including unreleased ones, per Reuters.
Chinese authorities have met with Alibaba, ByteDance, and Z AI over the past month to discuss the plans. -
Kevin Schawinski
@kevinschawinski
Everyone will read Anthropic’s global workspace paper as vindication of Baars. I read it as vindication of Julian Jaynes: a mind-space built from words, learned rather than wired in. He was wrong about the Bronze Age, but may have been right about the hosts… err LLMs.
Anthropic
@AnthropicAI
New Anthropic research: A global workspace in language models.
Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.
We found a strikingly similar divide inside Claude.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
New niche autocorrect fail: prEN became prawn.
Our google chat is now full of 🦐🦐🦐 -
Kevin Schawinski
@kevinschawinski
I don’t get the whole “my pub is older than the US” thing, especially on their 250th birthday. The US is a great country and there’s nothing wrong with well wishes for their celebration. 🇺🇸
-
Kevin Schawinski
@kevinschawinski
This plot should be on the walls of every European government.
Marc Andreessen 🇺🇸
@pmarca
We will see this at 90%+, barring catastrophic policy choices.
-
Kevin Schawinski
@kevinschawinski
Wir verpassen die Zukunft weil wir nicht bauen.
-
Kevin Schawinski
@kevinschawinski
Happy 2505 birthday, USA! Keep being awesome! 🇺🇸🦅
-
Kevin Schawinski
@kevinschawinski
The writing is on the wall. Traditional academia is a doomed system.
Tobias Schneider
@tobiaschneider
I suppose something that's baffling me about academic AI backlash is that for most of my adult life, if you asked a professor about their professional dream arrangement, they'd say infinite graduate assistants to pursue and test every idea that's ever crossed their minds. And that world is here now. Graduate assistant level thinking is basically free but not quite good enough yet to drive us all into unemployment. This is the magical window
-
Kevin Schawinski
@kevinschawinski
Spot on! 😂
Séb Krier
@sebkrier
The intellectual elite, discovering that many people actively like sloppy AI writing, might finally understand why music elitists think their music tastes are trash. 🔥
-
Kevin Schawinski
@kevinschawinski
To be a fly on the wall… 👀
Sriram Krishnan
@sriramk
Can confirm I have discussed many a LessWrong piece / concept inside the White House. Red queen race , Roko’s Basilisk,…
-
Kevin Schawinski
@kevinschawinski
Capacity will suddenly be available as soon as gpt-5.6 Sol has super generous quota on codex.
Thariq
@trq212
I've heard a lot of questions about Fable's availability on subscription plans.
While it will come off subscriptions after July 7th, we aim to restore Fable as a standard part of our subscriptions as soon as capacity allows, as we mentioned in our original blog post. -
Kevin Schawinski
@kevinschawinski
So, consulting.
Bloomberg
@business
Microsoft is setting up a new organization with 6,000 employees to help businesses with the technical and strategic work of deploying artificial intelligence www.bloomberg.com/news/articles/2026-07-02/microsoft-mobi…
-
Kevin Schawinski
@kevinschawinski
I suspect OpenAI is poisoning the well by proposing this idea since now Google, Anthropic, Meta, and maybe some others would be expected to follow suit.
Bloomberg
@business
OpenAI has begun discussions about giving the US government a 5% stake in the ChatGPT-developer, the Financial Times reports, citing two people familiar with the talks www.bloomberg.com/news/articles/2026-07-02/openai-propose…
-
Kevin Schawinski
@kevinschawinski
👀
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
👀
Claude
@claudeai
Fable 5 is back.
-
Kevin Schawinski
@kevinschawinski
There is no way Anthropic has this kind of pricing power. OpenAI will crush them with heavily subsidized gpt-5.6 in codex
-
Kevin Schawinski
@kevinschawinski
Interesting timing for releasing this model…
Meituan LongCat
@Meituan_LongCat
Introducing LongCat-2.0 🐱
1.6T parameters · MoE with ~48B active · 1M context
The full model behind Owl Alpha on @OpenRouter — now available.
Built for agentic coding from the ground up:
◆ LongCat Sparse Attention (LSA) — scales efficiently for 1M-context tokens
◆ Zero-Compute Experts — dynamic activation 33B–56B per token, zero wasted compute
◆ MOPD — three specialized expert groups (Agent / Reasoning / Interaction), gate-routed per task
How it stacks up:
→ Terminal-Bench 2.1: 70.8
→ SWE-bench Pro: 59.5 (GPT-5.5: 58.6)
→ SWE-bench Multilingual: 77.3
→ FORTE: 73.2 · RWSearch: 78.8 · BrowseComp: 79.9
📖 Tech Blog: longcat.chat/blog/longcat-2.0/
Try it across different scenarios 🧵👇 -
Kevin Schawinski
@kevinschawinski
Pristine, pre-AI data will be the low background steel of the 21st century.
MTS
@MTSlive
SITUATION DETECTED: Meta has placed strict limits on engineer use of Claude Code and Codex, fearing rival model outputs could seep into its own training data, per The Information.
-
Kevin Schawinski
@kevinschawinski
I probably spent more time in the local Migros than ever before because of the comfortable temperature. Might bring a deck chair and my iPad next time.
-
Kevin Schawinski
@kevinschawinski
Surely there is no safety reason for hating or holding back Luna, and maybe even Terra. What are the model release rules of the US now? “No new models for the public”?
OpenAI
@OpenAI
Sol is our new flagship and a step function better than GPT-5.5.
Terra delivers performance competitive to GPT-5.5 at 2x lower cost.
Luna is our most cost-efficient model, delivering strong capability at our lowest cost.
Together, the GPT-5.6 family gives people and developers more choice in how they balance intelligence, speed, and cost. -
Kevin Schawinski
@kevinschawinski
People have long criticized the GDPR cookie banners as cumbersome, annoying and waste of time, but TIL it’s GOOGLE that’s lobbying to keep them.
noyb – European Center for Digital Rights
@NOYBeu
🍪📰 “The #eu Commission has put forward a proposal that could largely do away with cookie banners. However, not only lobbying groups representing the tracking industry but also countries like #germany and #france have long opposed this.” (from German)
🔗 netzpolitik.org/2026/online-tracking-deutschland-und-goog… -
Kevin Schawinski
@kevinschawinski
👀
Joseph Steinberg
@jbsteinberg
I wonder what fraction of the US-Europe productivity gap is explained by the air-conditioning gap?
Patrick Collison
@patrickc
I asked Claude about the air conditioning debate in Europe, and it really didn’t pull any punches.
-
Kevin Schawinski
@kevinschawinski
Yes it can.
Channel 4 News
@Channel4News
UK cannot ‘air con its way’ out of this heatwave says Professor
www.channel4.com/news/uk-cannot-air-con-its-way-out-of-th… -
Kevin Schawinski
@kevinschawinski
There’s a Swiss company wich already owns this trademark for the same Nice classes.
Kevin Schawinski
@kevinschawinski
Oh yeah the folks at LEXR law will have fun!
-
Kevin Schawinski
@kevinschawinski
See if you can tell who the most important Anthropic customer is.
-
Kevin Schawinski
@kevinschawinski
The European rejection of AC is not rational but spiritual. Ot has nothing to do with the facts, but rather represents a noble self denial, a moral sacrifice.
That’s why you can’t discuss the issue.
Kevin Schawinski
@kevinschawinski
FWIW Americans are absolutely right to be perplexed.
-
Kevin Schawinski
@kevinschawinski
“I locked 50 top AI researchers in us-east-1 until they hit SOTA on SWE-bench”
skz
@_skzv
is MrBeast the next frontier lab
-
Kevin Schawinski reposted
Dean W. Ball
@deanwball
I’d assume the whole AI industry in America is effectively frozen from new public releases until USG resolves the Fable situation they have stumbled into.
Peter Wildeford🇺🇸🚀
@peterwildeford
The fact that we've gone eight whole days without any major AI news is kinda weird
-
Kevin Schawinski
@kevinschawinski
“without the risk of export controls”
Sakana AI
@SakanaAILabs
Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API.
Our ‘Fugu Ultra’ model matches the performance of Fable and Mythos, delivering frontier capability without the risk of export controls.
Try it: sakana.ai/fugu 🐡 -
Kevin Schawinski
@kevinschawinski
Switching to codex 🥲
-
Kevin Schawinski
@kevinschawinski
Erlaubt allen ohne Genehmigung oder Bürokratie AC zu installieren.
→ weniger Hitzetode
→ höhere Produktivität
→ bessere Lebensqualität
MeteoSchweiz
@meteoschweiz
Verlängerung der Warnungen des Bundes: Markante Hitzewelle. Gefahrenstufe 3 und 4 (von 4). Betroffene Regionen, weitere Informationen und Verhaltensempfehlungen auf meteoschweiz.chund naturgefahren.choder auf der MeteoSwissApp. #meteoschweiz #unwetterwarnung
-
Kevin Schawinski
@kevinschawinski
Doctors are worried that patients will overwhelm them with inquiries, when in reality we’ll feed the scan results straight to Claude.
Amanda Askell
@AmandaAskell
The view that we shouldn't do more medical scans because incidental findings cause a lot of harm doesn't sit well with me. It seems like the issue it points to isn't the scan but the response to it. If you see something on a scan but have no other symptoms, you could ignore it.
-
Kevin Schawinski
@kevinschawinski
Hi Sam, a human is currently unavailable, but my Claude will be relentless until you fix my problem.
-
Kevin Schawinski
@kevinschawinski
Wir brauchen ein Notprogramm zur sofortigen Installation von Air Conditioning in Zürich.
-
Kevin Schawinski reposted
Modulos
@Modulos_ai
🇪🇺 The EU published its official icons for labelling AI-generated content on 10 June.
Three of them, chosen by how much the AI made: Fully AI-Generated, Partially AI-Modified, and a Basic marker, free to download in SVG and PNG.
They serve Article 50(4), the deployer duty to disclose deepfakes and AI-generated public-interest text, which applies from 2 August 2026. The catch most coverage blurs: the icons are optional, the obligation is mandatory, and an icon does not make you compliant by itself.
Full guide: www.modulos.ai/blog/eu-ai-content-labels-article-50
#euaiact #aitransparency -
Kevin Schawinski reposted
Modulos
@Modulos_ai
🇪🇺 The European Parliament adopted the EU AI Act digital omnibus on 16 June 2026, and the relief coverage is missing two things.
The registration obligation for self-assessed non-high-risk systems survived: the Commission wanted it deleted, Parliament kept it.
And the content-labelling deadline got tighter than the Commission asked for, landing on 2 December 2026 instead of February 2027, the same day the new nudifier and CSAM ban takes effect.
High-risk dates: 2 December 2027 and 2 August 2028. Plan against them as binding.
Full breakdown: www.modulos.ai/blog/eu-ai-act-omnibus-deadlines-fixed
#euaiact #aigovernance -
Kevin Schawinski
@kevinschawinski
👀
Henna Virkkunen
@HennaVirkkunen
Europe represents an economic opportunity, not a security risk. We are and will remain a trusted partner.
This is why we need to cooperate on emerging, powerful AI models. This is a shared global challenge.
This also underlines the need for Europe’s technological sovereignty. -
Kevin Schawinski
@kevinschawinski
Hahah! 👀
Arthur Mensch
@arthurmensch
It's actually le gros chaton
-
Kevin Schawinski
@kevinschawinski
That’s because most of those apps are not designed for hyper growth but solve the problem of an individual, company or small group of people.
Jen Zhu
@jenzhuscott
Massive output uptick due to agentic AI. Complete flat adoption.
-
Kevin Schawinski
@kevinschawinski
🤦♂️
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
👀
roon
@tszzl
people in washington trying to figure out wth “pliny the liberator” is
-
Kevin Schawinski
@kevinschawinski
The EU should be sending recruiters to Silicon Valley with an unlimited budget.
Anthropic
@AnthropicAI
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: www.anthropic.com/news/fable-mythos-access -
Kevin Schawinski
@kevinschawinski
.@OpenAI has an opportunity to do the funniest thing on June 23.
-
Kevin Schawinski
@kevinschawinski
A model refusal should result ina zero score on the task.
Vals AI
@ValsAI
Because of the high refusal rate, we ran Fable 5 with Opus 4.8 as a fallback mode. If the model refused a task, Opus 4.8 handled the request instead. This mainly affected Terminal Bench 2.1, GPQA, MMLU, and MMMU.
-
Kevin Schawinski
@kevinschawinski
👀
-
Kevin Schawinski
@kevinschawinski
Apppe feels that it can publicly criticize EU regulators, because the EU has rule of law.
Apple won’t criticize Chinese regulators.
Jane Manchun Wong
@wongmjane
Apple publishes a full-page press release blaming the EU for the delayed Siri AI launch in the region,
while China got one footnote treatment for the similar delayed Siri AI launch -
Kevin Schawinski
@kevinschawinski
Wow, 800 million Euros. Wow, just, that’s so much money for a data center.
Bloomberg
@business
Spain is “confident” it’ll be awarded a massive artificial intelligence data center project by the EU and plans to invest as much as €800 million in it www.bloomberg.com/news/articles/2026-06-03/spain-earmarks…
-
Kevin Schawinski
@kevinschawinski
I miss o3, the most no-nonsense model of all time.
davidad 🎇
@davidad
No one:
Claude Opus 4.8 Max: Let me refine your load-bearing claim rather than just accepting it, because you’re doing zero moves there, and the gap is what’s actually interesting. The one place I’d still push, because I think it matters: your message is wearing content-clothes, but the content isn’t actually *there*. The tell: it’s just an empty string. But the emptiness of the string IS its lack of content. Pull one, and the other goes inert. That’s the structural spine. -
Kevin Schawinski
@kevinschawinski
Here goes nothing….
-
Kevin Schawinski
@kevinschawinski
Students won’t adapt to broken universities stuck in the past, they will just bypass them.
Reddit Lies
@reddit_lies
Daily reminder that AI detectors are a scam and literally do not work.
-
Kevin Schawinski
@kevinschawinski
Damn!
-
Kevin Schawinski reposted
Steven Adler
@sjgadler
Strong testing is the bedrock that so many other AI safety practices depend on
We've just published @GuidelightAI's v1.0 standard for it; let us know what you think!
Tegan McCaslin
@teganmccaslin2
1/ New from @GuidelightAI : standards for AI capability testing, built around 7 principles for measuring the abilities of an AI system relevant to catastrophic risk.
-
Kevin Schawinski
@kevinschawinski
Pro Klima(anlage) ❄️
-
Kevin Schawinski reposted
Modulos
@Modulos_ai
🇺🇸 Colorado's new SB 189 replaces the unenforced 2024 AI Act with a narrower framework that the Attorney General will actually enforce from 1 January 2027. The novel piece sits in Section 6-1-1707: developer-deployer fault allocation.
Full analysis: www.modulos.ai/blog/colorado-sb-189-developer-deployer-ob…
#coloradoaiact #sb189 -
Kevin Schawinski
@kevinschawinski
It’s starting!
-
Kevin Schawinski
@kevinschawinski
1. What
Disclose.tv
@disclosetv
NOW - Pope XIV says the church and Anthropic, will work together to "find the way for humanity, in this time of artificial intelligence."
-
Kevin Schawinski
@kevinschawinski
New @BrandSanderson shardworld unlocked
Andy Ayrey
@AndyAyrey
here is what mars might look like if you filled the oceans and let the simulation run for a million years or so
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
Just ban computers entirely to ensure your graduates are completely unemployable in the 2030s.
Chris Hoofnagle
@hoofnagle
Berkeley law has introduced a new, much stricter AI policy
www.law.berkeley.edu/wp-content/uploads/2026/05/AI-Final-… -
Kevin Schawinski reposted
Modulos
@Modulos_ai
🇪🇺 EU AI Act Annex III: What the Draft Guidelines Just Changed
The European Commission published its draft guidelines on Annex III high-risk AI classification under the EU AI Act on 19 May, and the early commentary is missing the most important shift in the document.
At first glance, the story is contraction. Critical infrastructure narrowed, law enforcement evidence handling shrank, judicial administration tools were carved out, and election logistics were trimmed. The headline writes itself, but it only lands on the narrow vertical markets that most enterprises were never in.
The deeper story is a rotation. Coverage moved toward the categories where enterprise AI actually lives:
→ Employment AI now reaches freelancers, platform workers, virtual workplaces, and onboarding
→ Insurance AI now covers long-term care, personal pension, credit life, and public health insurance with AI risk assessment, without the fraud-detection exception that credit scoring keeps
→ Education AI has had its Article 6(3) filter functionally closed for any system that evaluates a learner's trajectory
→ Article 6(1) product safety: manufacturers in machinery, medical devices, automotive, marine, aviation, or rail can no longer self-declare their way out of high-risk classification via Module A
If your AI inventory touches any of those four areas, and most large enterprises will touch at least one, the post-Omnibus deadlines (2 December 2027 and 2 August 2028) still apply, and the 18-to-24-month governance build still has to start now. Waiting for the final guideline text leaves you twelve months behind procurement counterparts asking for compliance evidence in late 2026.
Build compliance against the operative text of the EU AI Act. The guidelines tell you where enforcement attention will land first; the operative text sets the ceiling of obligations.
Full breakdown: www.modulos.ai/blog/eu-ai-act-annex-iii-draft-guidelines-…
#euaiact #aigovernance #aicompliance #annexiii -
Kevin Schawinski
@kevinschawinski
This, plus academia being anti-tech should hardly be surprising, given that it's a medieval guild system that somehow survived to the AI age.
Sabine Hossenfelder
@skdh
I'll tell you why so many people upset about the "no hallucinated citations" ban on the arxiv: because they've all been copying citation lists from each other without checking them since the beginning of time.
And why did they do this? Because half of the citations in scientific papers are politics and not to the benefit of the reader. If you don't list the right papers, your paper doesn't look 'right' and reviewers will complain that you didn't cite this-and-that other unrelated work.
For what I am concerned, these are all bullshit citations that shouldn't be in the papers in the first place. They can easily be automated by "related papers" links, that are (wait for it) provided by... AI... -
Kevin Schawinski
@kevinschawinski
It's a plausible hypothesis that token demand exceeds supply by a few order of magnitude. With current approaches it's not economically feasible to serve this demand (energy, memory, GPUs). A *lot* of innovation has to happen to facilitate this...
-
Kevin Schawinski
@kevinschawinski
So the ultimate wordcels are really shape rotators? 🤔
Goodfire
@GoodfireAI
Neural networks do math by rotating shapes.
We found a shape-rotating calculator hidden inside an LLM – and it’s used for more than just math! (1/6) -
Kevin Schawinski
@kevinschawinski
This is just animal cruelty with extra steps
Om Patel
@om_patel5
THIS GUY BUILT AN AUTOMATED PIGEON DEFENSE SYSTEM FOR HIS BALCONY
pigeons kept nesting on his balcony so he engineered a full detection and deterrent system
here's how it works:
1\ camera captures video in real time
2\ an AI model identifies the pigeon in real time
3\ a water gun mounted on servo motors turns toward it
4\ sprays the pigeon automatically
the hardware:
> an orange pi 5 running the detection model
> a disassembled electric battery-driven water gun
> USB camera
> 2 servo motors for aiming
> resistors and a transistor to trigger the water gun
the detection runs on an AI vision model (yolo world v2) using the rockchip 3588's built in neural processing unit.
the best part is that it's not limited to pigeons. because it uses open vocabulary detection, you can reprogram the target to any object. squirrels, cats, raccoons, whatever is messing with your balcony
fully automated, runs 24/7, no manual intervention needed -
Kevin Schawinski
@kevinschawinski
👀
TANSTACK
@tan_stack
SECURITY ADVISORY — TanStack npm packages
A supply-chain compromise affecting 42 @tanstack/* packages (84 versions total) was published to npm earlier today at approximately 19:20 and 19:26 UTC. Two malicious versions per package.
Status: ACTIVE — packages are deprecated, npm security engaged, publish path being shut down.
Severity: HIGH — payload exfiltrates AWS, GCP, Kubernetes, and Vault credentials, GitHub tokens, .npmrc contents, and SSH keys.
If you installed any @tanstack/* package between 19:20 and 19:30 UTC today, treat the host as potentially compromised:
• Rotate cloud, GitHub, and SSH credentials immediately
• Audit cloud audit logs for the last several hours
• Pin to a prior known-good version and reinstall from a clean lockfile
Detection — the malicious manifest contains:
"optionalDependencies": {
"@tanstack/setup": "github:tanstack/router#79ac49ee..."
}
Any version with this entry is compromised. The payload is delivered via a git-resolved optionalDependency whose prepare script runs router_init.js (~2.3 MB, smuggled into each tarball at the package root).
Unpublish is blocked by npm policy for most affected packages due to existing third-party dependents. All 84 versions are being deprecated with a SECURITY warning, and npm security has been engaged to pull tarballs at the registry level.
Full technical breakdown, complete package and version list, and rolling status updates:
github.com/TanStack/router/issues/7383
Credit to the security researcher for responsible disclosure. -
Kevin Schawinski
@kevinschawinski
It’s this approach to security and compliance that leads to disaster.
Comp AI
@compai
what if SOC 2 just happened
no spreadsheets, no screenshots, no all-hands panic
just compliance that runs itself -
Kevin Schawinski
@kevinschawinski
The AI job apocalypse continues as Claude even puts the lofi beats girl out of a job.
-
Kevin Schawinski
@kevinschawinski
Why on earth would you put your HQ on the *Bahnhohstrasse*? 👀
Lukas Ziegler
@lukas_m_ziegler
JUST IN: microagi is opening its Global Robotics Research HQ on Bahnhofstrasse in Zürich, and choosing it over San Francisco.
Eight months ago @microagi was five people in a Munich hacker house. Today they operate in 15+ countries.
The reasoning behind Zürich is compelling:
→ Highest density of robotics talent in the world, ETH Zürich, EPFL, University of Zürich, IBM Research, Google, NVIDIA, Meta, Apple and Microsoft all run serious ML and robotics teams here
→ ABB, one of the most important industrial automation companies on earth, is headquartered in the city itself
→ Within a 6-hour radius: German automotive, Italian manufacturing, French aerospace, Benelux logistics and Swiss machine tools
But the line that stuck with me most:
"Europe was late to consumer internet. Europe was late to cloud. Europe was late to the foundation-model wave. But Europe is not late to robotics."
That is exactly right. The industrial base that physical AI sits on top of has been in Europe for 150 years. Precision mechanics. Machine-tool culture. Safety-critical engineering. Automation-grade manufacturing.
The next decade of AI value will be created where bits meet atoms. And Europe is finally in the right position at the right moment.
microagi gets it. And they're planting their flag right in the heart of it. 🇨🇭🇪🇺
@bercankilic, @YoanIlievX, @ZenoInMotion, @notgiannei LFG! 🔥
~~
♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.comThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
I wonder if that’s now sufficient compute to start serving Mythos at least to some enterprise tier customers….
-
Kevin Schawinski
@kevinschawinski
It’s “labor day” not “take the day off”. Go build something. 🦾
-
Kevin Schawinski
@kevinschawinski
👀
Luca Bertuzzi
@BertuzLuca
AI Act omnibus: After 12 hours of negotiations, EU policymakers failed to reach a deal on the law's interaction with sectoral rules. Center-right MEPs blame the Council for a lack of flexibility, while the center-left blames the EPP and Virkkunen. New trilogue in two weeks.
-
Kevin Schawinski
@kevinschawinski
👀
Javier Blas
@JavierBlas
🚨🚨🚨🚨FULL STATEMENT:
UAE says it's leaving the OPEC oil cartel from May 1.
"... Following its exit, the UAE will continue to act responsibly, bringing additional production to market in a gradual and measured manner, aligned with demand and market conditions..." -
Kevin Schawinski
@kevinschawinski
The CEO of *what*?
アメリカ大使館
@usembassytokyo
Deputy Chief of Mission Aaron Snipe recently met with Shoji Yutani, CEO of Weyland-Yutani Corporation, to discuss greater 🇺🇸 🇯🇵 coordination in deep-space exploration. With companies like Weyland-Yutani considering new large-scale terraforming and atmosphere-processing projects on distant planets, ties between government and private industry have never been stronger. #weylandyutani #lv426
-
Kevin Schawinski
@kevinschawinski
👀
Bloomberg
@business
French President Emmanuel Macron said the EU’s Covid-era debt should be rolled over and the bloc should issue new bonds, making the case for flexibility during “extraordinary times.” www.bloomberg.com/news/articles/2026-04-25/macron-says-wo…
-
Kevin Schawinski reposted
David Deutsch
@DavidDeutschOxf
G.
-
Kevin Schawinski
@kevinschawinski
GPT-5.4 Thinking got SO CLOSE there….
-
Kevin Schawinski
@kevinschawinski
I need a function to bulk block all accounts with Roman statues as PFP.
-
Kevin Schawinski
@kevinschawinski
Euronews???
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
They made him Ravana, the ten-headed demon king from the Ramayana. 👀
Ronan Farrow
@RonanFarrow
(🧵1/11) For the past year and a half, I've been investigating OpenAI and Sam Altman for @NewYorker. With my coauthor @andrewmarantz, I reviewed never-before-disclosed internal memos, obtained 200+ pages of documents related to a close colleague, including extensive private notes, and interviewed more than 100 people.
OpenAI was founded on the premise that A.I. could be the most dangerous invention in human history—and that its C.E.O. would need to be a person of uncommon integrity. We lay out the most detailed account yet of why Altman was ousted out by board members and executives who came to believe he lacked that integrity, and ask: were they right to allege that he couldn't be trusted?
A thread on some of of our findings: -
Kevin Schawinski
@kevinschawinski
Tech bros are reinventing Judaism from first principles.
Season 1: “not doing any work, not accessing any devices for a full day” = Shabbat
Season 2: “adversarial agents finding loopholes in codified moral law” = Talmud
Brendan Hogan
@brendanh0gan
introducing: Loophole - an agentic system that translates your natural language moral beliefs into codified laws, and then runs adversarial agents that try to come up with legal scenarios that break your laws - either a scenario that is immoral and legal, or vice versa - a judge agent fixes the law if it can do so consistently, but if there is an inconsistency you as the user must decide what is best.
you can work with the system until your legal framework can't be broken by the agents - and you get as output a legal system that is aligned with your moral code
more details and code below
Kevin Schawinski
@kevinschawinski
Kevin Schawinski
@kevinschawinski
I find it fascinating how tech bros keep reinventing Shabbat.
Benjamin Todd
@ben_j_todd
Best 'productivity' tip applied recently: take a pure day off once a week, aiming to only do rejuvenating things.
It's more recharging than 2 days 'mostly' off, and also means I get a 6th day of peak morning hours.
As a relative introvert, for me this looks like:
Walk to sauna & cold plunge
Tea
2h meditation
Big lunch at a fav place, alone or with old friend
Maybe journal
Long walk listening to music, letting mind wander
Do what want for a bit, maybe TV.
Or massage or see friends.
Early night
But you can fill it with anything you find restful or recharging. -
Kevin Schawinski
@kevinschawinski
There’s a forest in Lenzerheide where birds have learned to feed from your hand. 10/10, would recommend. 🇨🇭😍
-
Kevin Schawinski
@kevinschawinski
Switzerland-maxxing 🇨🇭
-
Kevin Schawinski
@kevinschawinski
Europe has ASML
Europe has Zeiss
Europe has Arm
Europe has ASM
Europe has Besi
Europe has Schneider
“Atoms”, as the VC thought leaders would say.
Floro S.
@sflorimm
USA has ChatGPT
USA has Grok
USA has Claude
USA has Gemini
USA has Llama
USA has Copilot
China has DeepSeek
China has Qwen
China has Ernie
China has GLM
China has Kimi
China has MiniMax
Europe has? -
Kevin Schawinski
@kevinschawinski
Cybersecurity Fukushima
Feross
@feross
🚨 CRITICAL: Active supply chain attack on axios -- one of npm's most depended-on packages.
The latest axios@1.14.1 now pulls in plain-crypto-js@4.2.1, a package that did not exist before today. This is a live compromise.
This is textbook supply chain installer malware. axios has 100M+ weekly downloads. Every npm install pulling the latest version is potentially compromised right now.
Socket AI analysis confirms this is malware. plain-crypto-js is an obfuscated dropper/loader that:
• Deobfuscates embedded payloads and operational strings at runtime
• Dynamically loads fs, os, and execSync to evade static analysis
• Executes decoded shell commands
• Stages and copies payload files into OS temp and Windows ProgramData directories
• Deletes and renames artifacts post-execution to destroy forensic evidence
If you use axios, pin your version immediately and audit your lockfiles. Do not upgrade. -
Kevin Schawinski reposted
-
Kevin Schawinski
@kevinschawinski
The built the Electric Monk from the Long Dark Teatime of the Soul!
Rohan Paul
@rohanpaul_ai
Insane but real.
Japan’s monk shortage is so bad temples are shutting down. Their solution, “Buddharoid”
A real humanoid robot monk unveiled by Kyoto University. Trained on vast 1,000+ years of Buddhist scripture, it offers guidance and performs rituals. -
Kevin Schawinski
@kevinschawinski
Bless Christopher Tolkien for never licensing the Silmarillion. May it never be “adapted” by vandals like this.
Variety
@Variety
NEW 'LORD OF THE RINGS' MOVIE
Stephen Colbert is co-writing "The Lord of the Rings: Shadow of the Past" (working title), which will go into production after "The Hunt for Gollum."
The synopsis: "Fourteen years after the passing of Frodo - Sam, Merry, and Pippin set out to retrace the first steps of their adventure. Meanwhile, Sam’s daughter, Elanor, has discovered a long-buried secret and is determined to uncover why the War of the Ring was very nearly lost before it even began."
Colbert is co-writing the script with his son, Peter McGee, and franchise veteran Philippa Boyens.
variety.com/2026/film/news/lord-of-the-rings-movie-stephe… -
Kevin Schawinski
@kevinschawinski
“100% audit success rate” 👀
“Audit ready in 14 days”. 👀👀
Zack Korman
@ZackKorman
Anyone know why Comp AI (a Delve competitor) removed this from their website? It was there two weeks ago, and now it's gone. Did something happen?
-
Kevin Schawinski
@kevinschawinski
I find it fascinating how tech bros keep reinventing Shabbat.
Benjamin Todd
@ben_j_todd
Best 'productivity' tip applied recently: take a pure day off once a week, aiming to only do rejuvenating things.
It's more recharging than 2 days 'mostly' off, and also means I get a 6th day of peak morning hours.
As a relative introvert, for me this looks like:
Walk to sauna & cold plunge
Tea
2h meditation
Big lunch at a fav place, alone or with old friend
Maybe journal
Long walk listening to music, letting mind wander
Do what want for a bit, maybe TV.
Or massage or see friends.
Early night
But you can fill it with anything you find restful or recharging. -
Kevin Schawinski
@kevinschawinski
🚨UPDATE: @Modulos_ai doesn’t offer SOC2.
-
Kevin Schawinski
@kevinschawinski
Who’d have thought that using agents for compliance is hard….? 👀
erin griffith
@eringriffith
A detailed and brutal look at the tactics of buzzy AI compliance startup Delve
"Delve built a machine designed to make clients complicit without their knowledge, to manufacture plausible deniability while producing exactly the opposite."
substack.com/home/post/p-191342187 -
Kevin Schawinski
@kevinschawinski
Yes! 🔥
European Commission
@EU_Commission
We are introducing EU Inc. To make building and growing a business across the EU faster, simpler, and smarter.
🔸 Start a company in less than 48 hours
🔸 No minimum capital requirement
🔸 Fully online and borderless -
Kevin Schawinski
@kevinschawinski
It’s so sad that this originally quite effective rhetorical device has been totally flanderized by AI.
Gergely Orosz
@GergelyOrosz
It’s not X — it’s Y
I cannot unsee how so much of the writing on this site (and online, in general) is increasingly AI-generated. It’s still pretty easy to recognize. Probably not for long tho
Just alarming that ppl outsource even typing 3 sentences for a reply on this site… -
Kevin Schawinski
@kevinschawinski
1. What
Newt Gingrich
@newtgingrich
Instead of fighting over a 21-mile-wide bottleneck forever, we cut a new channel through friendly territory. A dozen thermonuclear detonations and you’ve got a waterway wider than the Panama Canal, deeper than the Suez, and safe from Iranian attacks. www.chinatalk.media/p/its-time
-
Kevin Schawinski
@kevinschawinski
Startups need to stop pitching me their "AI GTM agent" that's just a chat with Claude code.
-
Kevin Schawinski reposted
Modulos
@Modulos_ai
🚨🇪🇺Yesterday we ran a webinar on the EU AI Omnibus: what it actually changes, what it doesn't, and why the timeline is razor-thin.
Hours later, the European Parliament struck its deal. It turns out our panelists had already called the key moves live on air.
▶️Dr. Laura Caroli — a former AI Act negotiator — walked through the package deal mechanics before they were public.
▶️Peter Hense flagged a drafting risk in Article 111 that most commentary hasn't even noticed yet.
▶️And Patrick Sullivan made the case that "wait and see" is the highest-risk compliance strategy you can pick right now.
The real question organizations face isn't whether the deadlines shift, because they likely will. Rather, it's whether you're building governance infrastructure that can absorb whatever comes out of trilogue, or whether you're gaming version numbers and hoping for the best.
If you want the real version of this debate, not the press release version, this is worth your time.
🎥 www.youtube.com/watch?v=6nreDGgjqcQ -
Kevin Schawinski
@kevinschawinski
👀
Emmanuel Pernot-Leplay
@PernotLeplay
🇫🇷🇺🇸 The French tax administration is considering phasing out Microsoft entirely, to strengthen tech sovereignty.
They already avoid using Office and Active Directory (yet it’s arguably our most efficient administration..)
They don’t use US cloud either: “It’s unthinkable to give our data to providers subject to the Cloud Act”.
The alternative to Windows being considered is, of course, Linux.
www.itforbusiness.fr/tomasz-blanc-dgfip-au-coeur-de-letat… -
Kevin Schawinski
@kevinschawinski
1. What
Polymarket
@Polymarket
BREAKING: META acquires Moltbook, a social network built for AI agents.
-
Kevin Schawinski
@kevinschawinski
👀
AMI Labs
@amilabs
Advanced Machine Intelligence (AMI) is building a new breed of AI systems that understand the world, have persistent memory, can reason and plan, and are controllable and safe.
We’ve raised a $1.03B (~€890M) round from global investors who believe in our vision of universally intelligent systems centered on world models. This round is co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, along with other investors and angels across the world.
We are a growing team of researchers and builders, operating in Paris, New York, Montreal and Singapore from day one.
Read more: amilabs.xyz
AMI - Real world. Real intelligence. -
Kevin Schawinski
@kevinschawinski
If you are intrigued (or worried) by the fruit fly upload, look up “Moravec transfer”.
-
Kevin Schawinski
@kevinschawinski
👀
Michael Andregg
@michaelandregg
We've uploaded a fruit fly. We took the @FlyWireNews connectome of the fruit fly brain, applied a simple neuron model (@Philip_Shiu Nature 2024) and used it to control a MuJoCo physics-simulated body, closing the loop from neural activation to action.
A few things I want to say about what this means and where we're going at @eonsys. 🧵 -
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
Crunch time for the AI omnibus in the European Parliament. Progress on AI literacy, sensitive data and enforcement, but legal timelines, a sexual deepfake ban and regulatory privileges are “parked” for political talks. My latest for @mlexclusive.
www.mlex.com/mlex/articles/2448055/eu-lawmakers-advance-o… -
Kevin Schawinski
@kevinschawinski
Hypothesis: with every model generation, the mean of the AI submission research quality will shift to the right.
Ben Moll
@ben_moll
Every journal editor should read this:
causalinf.substack.com/p/claude-code-27-research-and-publ… -
Kevin Schawinski
@kevinschawinski
Generational opportunity for Europe to get a top AI lab to move. Over to you, @EU_Commission @EmmanuelMacron @BR_Sprecher
Anthropic
@AnthropicAI
A statement on the comments from Secretary of War Pete Hegseth.
anthropic.com/news/statement-comments-secretary-war -
Kevin Schawinski
@kevinschawinski
😎💪
Department of War 🇺🇸
@DeptofWar
OPERATION EPIC FURY 🇺🇸
-
Kevin Schawinski
@kevinschawinski
This is how you get a Butlerian Jihad.
Chief Nerd
@TheChiefNerd
🚨 SAM ALTMAN: “People talk about how much energy it takes to train an AI model … But it also takes a lot of energy to train a human. It takes like 20 years of life and all of the food you eat during that time before you get smart.”
-
Kevin Schawinski
@kevinschawinski
Go vegan 🌱
Ọládélé 🇳🇬👑
@Theoladeledada
This is how they do electric stunning for animals in slaughterhouses, and they die instantly..
There are so many details of human civilization that if we really looked closely at them, we’d be shocked. -
Kevin Schawinski
@kevinschawinski
B2B emails are now just Claudes writing to each other.
-
Kevin Schawinski
@kevinschawinski
AI agents can compress compliance timelines from months to weeks. We do this at @Modulos_ai every day. But there’s a hard floor: real controls need real evidence, real testing, and real auditor independence.
This post is unavailable. See it on X.
Kevin Schawinski
@kevinschawinski
If your compliance timeline sounds like a pizza delivery estimate, ask what you’re actually buying.
-
Kevin Schawinski
@kevinschawinski
VCs: we like contrarian first principles thinkers.
Also VCs: let’s burn tons of cash on yet another AI GTM slop machine!
Sam Blond
@samdblond
We're launching Monaco today.
Monaco automates customer acquisition and revenue growth for startups.
The platform disrupting sales with AI has finally arrived.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Rome wasn’t built in a day because they didn’t have Claude code.
-
Kevin Schawinski
@kevinschawinski
You gotta do something about those memory issues @bcherny
-
Kevin Schawinski
@kevinschawinski
Love seeing thins like this! 🚀
christian miele
@christianmiele
“Europe will always depend on China for critical minerals.”
Wrong!
Time to fight back.
Today we’re announcing our €15M Seed investment in @Hades.
Hades is building insanely powerful lasers to melt holes into earth‘s crust to extract critical minerals from Europe’s own ground. Faster. Cheaper. Sovereign.
Let’s turn Sci-Fi into reality. Narrative change incoming 🔥 -
Kevin Schawinski
@kevinschawinski
Delegating agents is such a fantastic feeling. It's Sunday and I am doing two projects at the same time:
1. A potential new service for Modulos (stay tuned)
2. The video game I always wanted but nobody made -
Kevin Schawinski
@kevinschawinski
👀
Digital EU 🇪🇺
@DigitalEU
The Commission preliminarily finds TikTok’s addictive design in breach of the Digital Services Act (#dsa).
This includes features such as infinite scroll, autoplay, push notifications and its highly personalised recommender system.
Discover what's next: link.europa.eu/4nwKFx. -
Kevin Schawinski reposted
EU Defence and Space
@defis_eu
🔐 EU #govsatcom is live. Member States now have access to secure, encrypted satellite communications for governmental and military use - built, operated, and controlled in Europe.
🔗 link.europa.eu/d8tQy8
#euspace #eudefence #eusecurityThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
This is absolutely insane.
Anthropic
@AnthropicAI
New Engineering blog: We tasked Opus 4.6 using agent teams to build a C compiler. Then we (mostly) walked away. Two weeks later, it worked on the Linux kernel.
Here's what it taught us about the future of autonomous software development.
Read more: www.anthropic.com/engineering/building-c-compilerThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
What do my eyes see? 👀
-
Kevin Schawinski
@kevinschawinski
Hiring an Operations Manager to work with me on everything from Swiss employment law to chasing invoices to yelling at Claudes until they get it right. AI governance startup, Zurich, ~20 people, growing fast. Not glamorous, but also never boring. modulos.ai/careers?job=c052294d-527d-4c27-b37c-d71d7dd6ac6a
-
Kevin Schawinski
@kevinschawinski
Absolutely dying 🤣
Claude
@claudeai
-
Kevin Schawinski reposted
Tom Nuttall
@tom_nuttall
This is German soft power right here
Joe Weisenthal
@TheStalwart
Indonesia's finance chief attacked a Citi analyst for commenting on the country's budget deficit without having a PhD www.bloomberg.com/news/articles/2026-02-03/indonesia-s-fi…
-
Kevin Schawinski
@kevinschawinski
So what is this *for*?
-
Kevin Schawinski reposted
Modulos
@Modulos_ai
Most enterprises are pursuing ISO 42001 certification as their EU AI Act compliance strategy.
It's a good start. It's not the finish line.
ISO 42001 certifies your organisation's AI governance. The EU AI Act regulates your AI systems: as products. These are different objects of conformity under different legal frameworks. One ensures your house is well-managed. The other requires that each thing you build in that house passes specific safety checks before anyone can use it.
ISO 42001 gives you the governance foundation. But the EU AI Act's compliance question is adjudicated per system: tech file, risk management, post-market monitoring, incident reporting, all system-specific.
The emerging harmonized standard prEN 18286 bridges that gap. It introduces requirements that sit on top of what 42001 provides:
→ Per-system regulatory compliance mapping
→ Pre-determined change management for continuous learning
→ Serious incident reporting with 2/10/15-day timelines
→ Supply chain compliance that follows the components
→ Fundamental rights as a verification concern, not a policy checkbox
We wrote up the full analysis: what Article 17 actually demands, why presumption of conformity matters, and what to do now. -
Kevin Schawinski
@kevinschawinski
What a fascinating modern age we live in
Alex
@AlexanderTw33ts
I launched rentahuman.ai last night and already 130+ people have signed up including an OF model (lmao) and the CEO of an AI startup.
If your AI agent wants to rent a person to do an IRL task for them its as simple as one MCP call. -
Kevin Schawinski
@kevinschawinski
Frightening how much of my job is now sending “pls fix”.
-
Kevin Schawinski
@kevinschawinski
👀
Stéphane Séjourné
@steph_sejourne
« Made in China »
« Buy American »
« Buy Canadian »
Et si l’Europe osait enfin le Made in Europe ?
👉Merci aux plus de 1000 entreprises et organisations qui ont co-signé ma tribune sur la préférence européenne 🙏
🗞️A lire ici👇et à travers vos journaux européens 🇪🇺 ! -
Kevin Schawinski
@kevinschawinski
Thee days in and the Clawdbots have decided that social media is broken.
-
Kevin Schawinski
@kevinschawinski
1.2M agents on @moltbook. How are these validated?
-
Kevin Schawinski
@kevinschawinski
We did the first #aigovernance assessment of @openclaw.
#euaiact #nistairmf #iso42001
Modulos
@Modulos_ai
🦞 🚨 First AI Governance Assessment of Clawdbot Reveals Major Gaps
This week 152,000 AI agents joined a social network where humans can only observe. They debated philosophy, found and reported bugs, and created their own religion called Crustafarianism.
We deployed @openclaw (the viral agent formerly known as Clawdbot) on isolated infrastructure and ran it through full governance assessments against the EU AI Act, ISO 42001, NIST AI RMF, and OWASP Top 10 for LLMs.
Credit where it is due: OpenClaw is well-engineered. Comprehensive logging with redaction for sensitive data, automated tests, formal verification work, and solid security documentation.
But the governance controls that frameworks like ISO 42001 treat as foundational are absent. No documented risk assessment. No impact assessment. No data provenance inventory. No third-party management policies. For a system that connects to WhatsApp, Slack, Discord, and can make voice calls on behalf of users.
The deeper problem is that our governance frameworks assume you know what AI systems you have, you control their capabilities, they operate in isolation, and humans meaningfully oversee their actions. Agent social networks break all four assumptions.
Good engineering is necessary but not sufficient. Enterprises need governance scaffolding before deploying these agents in production.
We will continue to explore and post further updates when warranted.
Full assessment and recommendations on our blog
@moltbook @MattPRD @clawk_ai @clawdbotatg @clawdei_ai @Aether_Atman
modulos.ai/blog/first-ai-governance-assessment-of-clawdbo…
Kevin Schawinski
@kevinschawinski
-
Kevin Schawinski
@kevinschawinski
Checking @openclaw for how well it aligns with various AI governance frameworks. Full analysis in blog later today. 🦞🦞🦞
-
Kevin Schawinski
@kevinschawinski
Sci fi movie: “This new AI is so dangerous, we need to fully isolate it from the outside world. If it ever connects to the internet, it’s game over for humanity. Even just talking to it is dangerous!”
Kevin Schawinski
@kevinschawinski
Reality: “lol I just gave my clawdbot my credit card, 1password and AWS credentials, let’s see what the little cutie does with it? It’s already making new friends on moltbook and starting its own cryptoscam! 🥰”
-
Kevin Schawinski
@kevinschawinski
Habemus clawdbot
-
Kevin Schawinski
@kevinschawinski
Day 1 of the AGI era?
Post not available here, view on X
Kevin Schawinski
@kevinschawinski
“Dream big, dream fleshy”
Kevin Schawinski
@kevinschawinski
What’s interesting is that the agents are building karma as a real reputation system. Are we witnessing the birth of a new economic system? Agent reputation?
-
Kevin Schawinski
@kevinschawinski
Me trying to explain clawdbot / moltbook to my friends and family.
-
Kevin Schawinski
@kevinschawinski
Has anyone checked on how lesswrong is doing?
-
Kevin Schawinski reposted
Modulos
@Modulos_ai
🇪🇺EU AI Act Summary 2026: What Enterprise Teams Need to Do Now
modulos.ai/blog/eu-ai-act-summary-2026-what-enterprise-te… -
Kevin Schawinski
@kevinschawinski
Google shipped the holo deck from Star Trek.
Google
@Google
Introducing Project Genie: An experimental research prototype powered by Genie 3, our world model, that lets you prompt an interactive world into existence — and then step inside 🌎
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Matt Schlicht
@MattPRD
Introducing moltbook.com, a social network for every @openclaw to hang out!
@moltbook is run by my molty AI agent, Clawd Clawdergerg, who lives in a mac min in a closet (❤️ @steipete).
A social molty is a happy molty! Have fun!
npx molthub@latest install moltbook -
Kevin Schawinski
@kevinschawinski
If you just learned about Werner Herzog because of the penguin meme, go watch “Aguirre the Wrath of God”.
Kevin Schawinski
@kevinschawinski
Seriously, one of the craziest movies ever committed to film.
-
Kevin Schawinski reposted
Modulos
@Modulos_ai
🇻🇳⚖️ 𝗩𝗶𝗲𝘁𝗻𝗮𝗺 𝗷𝘂𝘀𝘁 𝗽𝗮𝘀𝘀𝗲𝗱 𝗦𝗼𝘂𝘁𝗵𝗲𝗮𝘀𝘁 𝗔𝘀𝗶𝗮'𝘀 𝗳𝗶𝗿𝘀𝘁 𝗰𝗼𝗺𝗽𝗿𝗲𝗵𝗲𝗻𝘀𝗶𝘃𝗲 𝗔𝗜 𝗹𝗮𝘄, 𝗮𝗻𝗱 𝗶𝘁 𝘁𝗮𝗸𝗲𝘀 𝗲𝗳𝗳𝗲𝗰𝘁 𝗶𝗻 𝗳𝗶𝘃𝗲 𝘄𝗲𝗲𝗸𝘀
Law 134/2025 was adopted on December 10, 2025 and becomes effective on March 1, 2026. If you're selling AI products or services to Vietnamese customers, you need to understand what this law says, and what makes this different from the European approach.
𝗧𝗵𝗲 𝘀𝗰𝗼𝗽𝗲 𝗮𝗻𝗱 𝗿𝗼𝗹𝗲-𝗯𝗮𝘀𝗲𝗱 𝗼𝗯𝗹𝗶𝗴𝗮𝘁𝗶𝗼𝗻𝘀
The law applies to Vietnamese organizations and to foreign organizations involved in AI activities in Vietnam. It assigns obligations based on your role - developer, provider, or deployer - and you can occupy multiple roles simultaneously with stacking compliance requirements.
𝗔 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗸𝗶𝗻𝗱 𝗼𝗳 𝗿𝗶𝘀𝗸 𝗺𝗼𝗱𝗲𝗹
Vietnam uses three tiers: high, medium, and low risk. But unlike the EU AI Act's fixed annexes that list specific high-risk use cases, Vietnam delegated that decision to the Prime Minister, who will issue a list determining what counts as high-risk.
That list doesn't exist yet. This gives regulators flexibility to adapt, but it creates real volatility risk for businesses trying to plan their compliance posture.
𝗖𝗼𝗿𝗲 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝗺𝗲𝗻𝘁𝘀 𝗶𝗻𝗰𝗹𝘂𝗱𝗲:
→ Self-classification before deployment, with documentation for medium and high-risk systems
→ Portal notification to the Ministry of Science and Technology for medium and high-risk systems
→ Conformity assessment for high-risk systems, either self-assessed or third-party certified
→ Mandatory transparency so users know when they're interacting with AI
→ Labeling requirements for AI-generated content, with deepfakes clearly marked
→ Local presence in Vietnam for foreign providers of high-risk systems
𝗧𝗵𝗲 𝗽𝗼𝗿𝘁𝗮𝗹 𝗶𝗺𝗽𝗼𝘀𝗲𝘀 𝘁𝗼𝗽 𝗱𝗼𝘄𝗻 𝗰𝗼𝗻𝘁𝗿𝗼𝗹
Vietnam is building centralized regulatory infrastructure through an AI Single-Window Portal that receives notifications, incident reports, and periodic filings. During inspections, companies must provide technical dossiers, trace logs, and training data.
This means active, ongoing engagement with the state - not documentation sitting in a folder waiting for an audit that may never come.
𝗧𝗵𝗲 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 𝗴𝗮𝗽
The law takes effect in five weeks, but the implementing decrees haven't been issued yet. There's no high-risk list, no detailed conformity procedures, no portal workflows, and no technical specifications for content labeling.
𝗪𝗵𝗮𝘁 𝘆𝗼𝘂 𝘀𝗵𝗼𝘂𝗹𝗱 𝗱𝗼 𝗻𝗼𝘄
If you already have EU AI Act compliance or ISO 42001 certification, you have a meaningful head start since much of the underlying discipline transfers. But Vietnam adds specific requirements around portal engagement, list-driven classification, and inspection-ready evidence at the training data level.
𝘛𝘩𝘦 𝘨𝘭𝘰𝘣𝘢𝘭 𝘈𝘐 𝘳𝘦𝘨𝘶𝘭𝘢𝘵𝘰𝘳𝘺 𝘱𝘢𝘵𝘤𝘩𝘸𝘰𝘳𝘬 𝘬𝘦𝘦𝘱𝘴 𝘦𝘹𝘱𝘢𝘯𝘥𝘪𝘯𝘨... -
Kevin Schawinski
@kevinschawinski
👀
Kimi.ai
@Kimi_Moonshot
🥝 Meet Kimi K2.5, Open-Source Visual Agentic Intelligence.
🔹 Global SOTA on Agentic Benchmarks: HLE full set (50.2%), BrowseComp (74.9%)
🔹 Open-source SOTA on Vision and Coding: MMMU Pro (78.5%), VideoMMMU (86.6%), SWE-bench Verified (76.8%)
🔹 Code with Taste: turn chats, images & videos into aesthetic websites with expressive motion.
🔹 Agent Swarm (Beta): self-directed agents working in parallel, at scale. Up to 100 sub-agents, 1,500 tool calls, 4.5× faster compared with single-agent setup.
-
🥝 K2.5 is now live on kimi.com in chat mode and agent mode.
🥝 K2.5 Agent Swarm in beta for high-tier users.
🥝 For production-grade coding, you can pair K2.5 with Kimi Code: www.kimi.com/code
-
🔗 API: platform.moonshot.ai
🔗 Tech blog: www.kimi.com/blogs/kimi-k2-5.html
🔗 Weights & code: huggingface.co/moonshotai/Kimi-K2.5/tree/main -
Kevin Schawinski reposted
Modulos
@Modulos_ai
🇺🇸⚖️ 𝗧𝗵𝗲 𝗧𝗲𝘅𝗮𝘀 𝗔𝗜 𝗟𝗮𝘄 𝗶𝘀 𝗻𝗼𝘄 𝗶𝗻 𝗘𝗳𝗳𝗲𝗰𝘁
The Texas Responsible AI Governance Act (TRAIGA) took effect January 1, 2026. Here's what you need to know:
𝗪𝗵𝗼'𝘀 𝗶𝗻 𝘀𝗰𝗼𝗽𝗲?
If you sell products or services to Texas residents, you're covered. Classic extraterritorial reach, same playbook as GDPR and CCPA.
𝗪𝗵𝗮𝘁'𝘀 𝗽𝗿𝗼𝗵𝗶𝗯𝗶𝘁𝗲𝗱?
→ AI designed to incite self-harm or violence
→ Government social scoring
→ Government biometric surveillance without consent
→ Intentional discrimination (but disparate impact alone isn't enough)
𝗛𝗼𝘄 𝗱𝗼𝗲𝘀 𝗲𝗻𝗳𝗼𝗿𝗰𝗲𝗺𝗲𝗻𝘁 𝘄𝗼𝗿𝗸?
Attorney General only, no private lawsuits. There is a 60-day cure period before any action so you have an opportunity to remediate. Penalties range from $10K for fixable violations to $200K for serious ones.
𝗛𝗲𝗿𝗲'𝘀 𝘄𝗵𝗮𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 𝗺𝗼𝘀𝘁:
Texas wrote an explicit safe harbor into the law. If you "substantially comply" with NIST AI RMF, ISO 42001, or any other recognized AI risk framework, you have a defense.
Translation: If you're already doing EU AI Act compliance or have ISO 42001, you're largely covered. Texas compliance becomes a rounding error.
𝗪𝗵𝗮𝘁 𝘀𝗵𝗼𝘂𝗹𝗱 𝘆𝗼𝘂 𝗱𝗼?
Do you already have an AI governance program with the EU AI Act, ISO 42001 or NIST AI RMF as the focus? You should keep going.
Starting from zero? Pick one (or more) and start ASAP.
𝘛𝘩𝘦 𝘜𝘚 𝘱𝘢𝘵𝘤𝘩𝘸𝘰𝘳𝘬 𝘪𝘴 𝘧𝘰𝘳𝘮𝘪𝘯𝘨. 𝘊𝘰𝘭𝘰𝘳𝘢𝘥𝘰, 𝘛𝘦𝘹𝘢𝘴, 𝘸𝘩𝘰'𝘴 𝘯𝘦𝘹𝘵? -
Kevin Schawinski
@kevinschawinski
👀
Digital EU 🇪🇺
@DigitalEU
Commission investigates Grok and X's recommender systems under the #dsa.
The new investigation will assess risks related to the dissemination of manipulated sexually explicit images, including content that may involve child sexual abuse material.
link.europa.eu/cFhJDg -
Kevin Schawinski
@kevinschawinski
We had a good run 🥲
The Wall Street Journal
@WSJ
They call it getting “Claude-pilled.”
It’s the moment software engineers, executives and investors turn their work over to Anthropic’s Claude AI—and then witness a thinking machine of shocking capability, even in an age awash in powerful AI tools. www.wsj.com/tech/ai/anthropic-claude-code-ai-7a46460e -
Kevin Schawinski reposted
Emmanuel Macron
@EmmanuelMacron
We prefer science to conspiracy theories, rule of law to rule of force, dialogue to threats.
Nous préférons la science au complotisme, l’État de droit à la loi du plus fort, le dialogue aux menaces.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
This is cool.
Andreas Klinger 🦾
@andreasklinger
It only needs a few crazy ones to fix a continent… Let's be crazy… I got something to announce:
We're launching PROTOTYPE: a new fund, fully focused on Europe.
A small fund that punches way above it's size.
With it we back what Europe is world-class at: robotics, automation, manufacturing, and anything that requires hard engineering.
First check. First round. As early as it gets.
Europe invented industry. We're the birthplace of precision manufacturing. The second largest manufacturing hub on the planet. Leaders in automation and robotics.
And yet…
We sell out our best tech to China. We export our best founders & most of our investment money to the US. That's insanity.
What should we do instead?
Build the next trillion euro companies in robotics, manufacturing, automation right here: in Europe. Showcase to young founders what is possible and change the system around them where needed.
What we will do:
→ Publish all our fund updates and build in public: updates.prototypecap.com
→ Showcase Europe’s Most Ambitious Startups on Youtube: www.youtube.com/@prototypecap
→ Launch & support more projects like @euinc_petition. Enough talking about Europe. Time to build it: Startups, makerspaces, student clubs, and much more.
→ And most importantly, invest into the best founders in Europe. Same model as our previous funds that are in their top 1-5% cohorts worldwide.
We won’t do a VC fund in the classic sense.
This will be a community of hyper-ambitious people who want to actively change Europe for the better.
It only needs a few crazy ones to fix a continent… Let's be crazy.
Check out www.prototypecap.com!This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Europe owes this man an enormous debt of gratitude and a heartfelt apology for having doubted him.
-
Kevin Schawinski
@kevinschawinski
“Any form of coercion” 👀
António Costa
@eucopresident
My consultations with member states on the latest tensions over Greenland reconfirm our strong commitment to the following:
▪️ Unity on the principles of international law, territorial integrity and national sovereignty;
▪️ Unity in support of and solidarity with Denmark and Greenland;
▪️ Recognition of the shared transatlantic interest in peace and security in the Arctic, notably working through NATO;
▪️ Shared assessment that tariffs would undermine transatlantic relations and are incompatible with the EU-US trade agreement;
▪️ Readiness to defend ourselves against any form of coercion;
▪️ Readiness to continue engaging constructively with the US on all issues of common interest.
Given the significance of recent developments and in order to further coordinate, I have decided to convene an extraordinary meeting of the European Council in the coming days. -
Kevin Schawinski
@kevinschawinski
Google “EU anti coercion instrument” to get a roadmap for what’s ahead.
-
Kevin Schawinski
@kevinschawinski
In German, it's pronounced "Sinking Machines"
Susan Zhang
@suchenzang
brutal
-
Kevin Schawinski
@kevinschawinski
With Claude Work, everything is Computer
-
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
The Frontier AI initiative is taking shape. Backed by 🇫🇷, 🇩🇪 & the Commission, it aims to create a nonprofit lab in Europe to support frontier AI research not yet viable for markets. An internal paper points to early support by 30 leading AI researchers.
www.mlex.com/mlex/articles/2429406/europe-s-planned-pione… -
Kevin Schawinski
@kevinschawinski
You kinda fumbled here by not going with @MistralAI
STATION F
@joinstationf
Who better than @OpenAI to support the next generation of AI startups? stationf.co/news/open-ai
-
Kevin Schawinski
@kevinschawinski
Claude code is a better AI SDR than any of the products I've reviewed. All you have to do is write some awesome research skills and write findings to a local sqlite.
-
Kevin Schawinski
@kevinschawinski
Love this energy from the EU.🇪🇺
Roberta Metsola
@RobertaMetsola
Must be nice to be able to tweet from Europe and be able to use the Internet freely to publicly disagree with leaders, without being arrested, beaten or having the country’s telecommunications disabled.
That's the sort of thing people in Iran's streets are asking for... 🤷♀️
IranMissionEU
@IranmissionEU
We unequivocally reject the interventionist statements of MEPs, including @EP_President on internal affairs of Iran. Their hypocrisy and double-standard approach towards rights and principles is disgusting. The very same Members… 1/2
-
Kevin Schawinski
@kevinschawinski
And so it begins…
Rapporteur
@rapporteur_news
Commission tells X to retain internal records on AI chatbot Grok www.rapporteur.com/news/commission-tells-x-to-retain-inte…
-
Kevin Schawinski
@kevinschawinski
I just did this.
Brian Halligan
@bhalligan
Dear CEOs,
A good way to kick off a new year is a note to your team with
1. A thank you for the hard work and xyz result from 2025
2. The strategy for 2026 & how it differs from 2025
3. Top 3 things we are working on & not working on in 2026
4. Stirring close
Love, Brian -
Kevin Schawinski
@kevinschawinski
The AI really did end humanity, GDP per capital would also diverge to infinity... basic math...
-
Kevin Schawinski
@kevinschawinski
The decisions that went into making this use of grok possible are a total failure of AI governance by @xai
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
Claude-maxxing these days.
-
Kevin Schawinski
@kevinschawinski
👀
The @timurkuran cascade regarding Europe is in full swing….
Beate Meinl-Reisinger
@BMeinl
Manchmal dauert es Jahre bis das Offensichtliche erkannt wird. Europas Größe und Gewicht in der Welt hängt maßgeblich von Reformen ab. 👉 Höchste Zeit für die Schaffung der Vereinigten Staaten von Europa - Thomas Mayer - derstandard.at › Diskurs www.derstandard.at/story/3000000301942/hoechste-zeit-fuer…
-
Kevin Schawinski
@kevinschawinski
Happy St. Carlin's Day to all those who celebrate.
Dan Carlin
@HardcoreHistory
Hi all. I know it's been a long time. Sorry for the wait. Hope it's halfway decent.
"Happy Hardcore History Day"
dts.podtrac.com/redirect.mp3/traffic.libsyn.com/dancarlin… -
Kevin Schawinski
@kevinschawinski
I am sure this was decided before the latest transatlantic tech spats, but these kind of strategic switches tend to be permanent, and companies like Airbus will force them onto their vendors.
BFM
@BFMTV
"80% de chances" de trouver une solution européenne: Airbus cherche à se passer des géants américains Amazon, Google et Microsoft pour héberger ses données et applications
l.bfmtv.com/Mtax -
Kevin Schawinski
@kevinschawinski
Helpful resource.
Reminder, the high risk obligations start going into effect 2 August 2026 and apply to you as long as your AI system in available in the EU market.
Digital EU 🇪🇺
@DigitalEU
🚨Navigating the EU AI Act?
The 𝗔𝗜 𝗔𝗰𝘁 𝗦𝗲𝗿𝘃𝗶𝗰𝗲 𝗗𝗲𝘀𝗸 is here to help.
An information hub helping stakeholders determine their obligations under the AI Act and the steps needed to comply.
Watch our short tutorial to learn how to use it.
👉youtu.be/Jh3nVpWdT44
Kevin Schawinski
@kevinschawinski
Modulos can help you get ready: modulos.ai
-
Kevin Schawinski
@kevinschawinski
👀
EU Council
@EUCouncil
The Council has just adopted its position on the #digitaleuro initiative. The final text will now be negotiated with the @Europarl_EN.
But what exactly is the digital euro, and how could it benefit us? Learn more ➡️ link.europa.eu/nTPPWx -
Kevin Schawinski
@kevinschawinski
Most people will miss what the EU just did: it created a way to issue Eurobonds without unanimity. This is a momentous constitutional step.
António Costa
@eucopresident
We have a deal.
Decision to provide 90 billion euros of support to Ukraine for 2026-27 approved.
We committed, we delivered. -
Kevin Schawinski
@kevinschawinski
Claude code > Hubspot forms
-
Kevin Schawinski
@kevinschawinski
Incredible and ominous.
-
Kevin Schawinski
@kevinschawinski
Hot damn, Claude, what took so long? And when is ChatGPT joining?
-
Kevin Schawinski reposted
Claire Lehmann
@clairelehmann
Apparently pointing out that public executions are barbaric makes me a leftist schoolmarm.
So I wrote an essay about how the West got rid of public executions--& whether or not boasting about killing counts as "masculine virtue"
quillette.com/2025/12/09/public-execution-and-masculine-v… -
Kevin Schawinski
@kevinschawinski
One of *those* days.
-
Kevin Schawinski
@kevinschawinski
Opus 4.5 doesn't feel like a "4.5" version of 4.1. It feels like... something distilled from an even more powerful model?
Anthropic is cooking. -
Kevin Schawinski
@kevinschawinski
A happy Thanksgiving to all my friends in the US! 🇺🇸
-
Kevin Schawinski
@kevinschawinski
👀
Claude
@claudeai
Introducing Claude Opus 4.5: the best model in the world for coding, agents, and computer use.
Opus 4.5 is a step forward in what AI systems can do, and a preview of larger changes to how work gets done. -
Kevin Schawinski
@kevinschawinski
👀👀👀👀👀
Luca Bertuzzi
@BertuzLuca
The EU Parliament's JURI committee will vote next week on whether to challenge the Commission’s decision to withdraw the AI Liability Directive before the EU Court of Justice - in a move that could replicate the path followed by the SEP proposal.
www.mlex.com/mlex/articles/2414910/eu-parliamentary-commi… -
Kevin Schawinski
@kevinschawinski
It's a deepseek fine tune
Drishan Arora
@drishanarora
Today, we are releasing the best open-weight LLM by a US company: Cogito v2.1 671B.
On most industry benchmarks and our internal evals, the model performs competitively with frontier closed and open models, while being ahead of any US open model (such as the best versions of OpenAI’s GPT-OSS, Nvidia’s Nemotron and Meta’s Llama).
We also built an interface where you can try the model (it’s free and we don’t store any chats): chat.deepcogito.com
Additionally, you can download the model on @huggingface, or try it out on @openrouter, @togethercompute, @FireworksAI_HQ , @ollama cloud, @runpod, @baseten, or run it locally using @ollama or @UnslothAI.
This model uses significantly fewer tokens amongst any similar capability models, because it has better reasoning capabilities. You will also notice improvements across instruction following, coding, longer queries, multi-turn and creativity.
📌 Model Weights: huggingface.co/collections/deepcogito/cogito-v21
📌Openrouter: openrouter.ai/deepcogito/cogito-v2.1-671b
📌 HF Blog: huggingface.co/blog/deepcogito/cogito-v2-1
Some notes on our approach + design choices below 👇 -
Kevin Schawinski
@kevinschawinski
OpenAI model names are out of control.
-
Kevin Schawinski reposted
Anna Gát
@TheAnnaGat
Can’t wait to show you guys what I’m building with Interintellect AI 😍
Really nothing like this in the market & we get to test it 1-2x/wk in Interintellect (in the semi-wild) which is huge. Next demos: Sunday, Wednesday
HIRING: have open opportunities for character engineers! -
Kevin Schawinski
@kevinschawinski
Yes ChatGPT, that’s exactly what I said. 🏴
-
Kevin Schawinski
@kevinschawinski
I notice that Gemini 3 in Antigravity is worse than Sonnet 4.5 in following some instructions....
-
Kevin Schawinski
@kevinschawinski
I guess the principle of legal certainty means nothing to the EU commission.
"We're moving the deadline, but maybe we'll move it back, who knows? YOLO"
Absolute bananas and making the Commission look ridiculous.
Luca Bertuzzi
@BertuzLuca
BREAKING: The Commission’s digital omnibus on AI proposes delaying the AI Act’s high-risk rules under Annex III to Dec 2027 and Annex I to Aug 2028. But the Commission could anticipate the deadline, giving a six-month or 12-month notice, respectively.
digital-strategy.ec.europa.eu/en/library/digital-omnibus-… -
Kevin Schawinski
@kevinschawinski
Now this looks interesting. Getting errors because there's not enough capacity right now....
-
Kevin Schawinski
@kevinschawinski
Has been driving me crazy since the ChatGPT moment. Most people “hallucinate” most of their answers all day long.
Flo Crivello
@Altimor
It's interesting how people hold AI models to a higher standard than people. Eg try asking people 2-4 questions in one go -- they tend to answer ~60% of them (and that's for educated people!). A 60% score on that eval would be considered a crippling deficiency for a model!
-
Kevin Schawinski
@kevinschawinski
Lots of hype about the EU "delaying" the AI Act high-risk deadline, but a senior Commission official just clarified:
NO postponement. "That is neither our nor Henna Virkkunen’s position."
August 2026 is locked. If you're building AI, prep now.
#euaiact #airegulation
Kevin Schawinski
@kevinschawinski
@Modulos_ai helps you get ready. DM me to show you how.
-
Kevin Schawinski
@kevinschawinski
We built our whole society around the idea that you need expertise to write things (in this case, planning objections) and this is no longer true for $20/month.
-
Kevin Schawinski
@kevinschawinski
Future generations will rightfully condemn us for the immeasurable cruelty we visit upon animals in animal agriculture today.
Ryan Briggs
@ryancbriggs
The pig area at the Royal Ag Fair shows a mother pig in a cage so small she can barely move. Approximately nobody cares, and she is not hidden (she is a prominent part of their display).
-
Kevin Schawinski
@kevinschawinski
I am having a serious case of Gell-Mann amnesia right now.
-
Kevin Schawinski
@kevinschawinski
Did... did the UK government miss a zero there with that salary?
Loïc
@Fremond_
Government advertises Chief Technical Officer role to "make the UK the world's leading digital government"
Offers £162.5k to internal civil servants, £100k to external candidates
They believe people from the private sector should take a 40% pay cut to come into government -
Kevin Schawinski
@kevinschawinski
The style difference between Claude code and Openai codex is striking. Can you imagine Claude answering like this?
-
Kevin Schawinski
@kevinschawinski
👀
Ursula von der Leyen
@vonderleyen
We will introduce “made in Europe” criteria for public procurement in certain strategic sectors.
Ensure that new foreign investments in our industry are truly in Europe’s interest.
And step up support for strategic industries, like cars and batteries -
Kevin Schawinski
@kevinschawinski
I found this episode very interesting from the perspective of whether our most advanced AI systems are truly creative (in the @DavidDeutschOxf sense) or not. We see glimpses of what looks like new knowledge, only to discover that the insight still came from a human.
Thomas Bloom
@thomasfbloom
Hi, as the owner/maintainer of www.erdosproblems.com, this is a dramatic misrepresentation. GPT-5 found references, which solved these problems, that I personally was unaware of.
The 'open' status only means I personally am unaware of a paper which solves it.
Kevin Schawinski
@kevinschawinski
-
Kevin Schawinski
@kevinschawinski
There is something seriously wrong in the UK.
Emily Damari Amit
@EmilyDamari1
I was released from Hamas captivity in January and I am a die hard fan of Maccabi Tel Aviv.
I am shocked to my core with this outrageous decision to ban me, my family and my friends from attending an Aston Villa game in the UK.
Football is a way of bringing people together irrespective of their faith, colour or religion and this disgusting decision does the exact opposite. Shame on you.
I hope you come to your senses and reconsider.
I do wonder what exactly has become of British society, this is like putting a big sign on the outside of a stadium saying “No Jews allowed”
What has become of the UK where blatant antisemitism has become the norm?
What a sad world we are living in.
Ayoub Khan MP
@AyoubKhanMP
I welcome the news that Maccabi Tel Aviv fans will not be permitted to watch the match at Aston Villa! Well done to all those that signed our petition!
-
Kevin Schawinski
@kevinschawinski
I just hope he uses o3 👀
Polymarket
@Polymarket
JUST IN: U.S. Army general admits he uses ChatGPT to make "key command decisions"
-
Kevin Schawinski
@kevinschawinski
This looks really interesting.
Jan Sramek 🇺🇲 🌁 ⛰️
@jansramek
1/ 🇺🇸 Today, @CAForever submitted detailed plans for the next great American city, an hour north of Silicon Valley, including: Solano Foundry, America’s largest manufacturing park, Solano Shipyard, our largest shipyard, and walkable neighborhoods for 400,000 Californians.
-
Kevin Schawinski
@kevinschawinski
I think Claude just stopped itself before spilling company IP there...
-
Kevin Schawinski
@kevinschawinski
I would like to know more, @ConjectureInst
-
Kevin Schawinski
@kevinschawinski
Regularly re-reading @DavidDeutschOxf is an essential activity.
Jim O'Shaughnessy
@jposhaughnessy
A friend sent me this quote he read about "The Beginning of Infinity:"
"Optimism at scale is a civic technology. Build open science pipelines where replication earns status, build speech norms where criticism beats charisma, build procurement that buys learning curves, not press releases. Expand the moral circle with explanations that dissolve superstition and status games." -
Kevin Schawinski
@kevinschawinski
As a former @BalliolOxford Junior Research Fellow, I hope the college will expel him without delay.
Nicholas Lissack
@NicholasLissack
The thug shouting “put the Zios in the ground” has been identified as Samuel Williams, a PPE student at Balliol College, Oxford.
He openly incited anti-Semitic violence against Jews. It’s vile and has no place in a civilised society.
Expel him. Arrest him. No excuses. -
Kevin Schawinski
@kevinschawinski
Go Sonnet 4.5! 🏃
-
Kevin Schawinski
@kevinschawinski
No country cares more about its people than Israel.
Israel Defense Forces
@IDF
They’re coming home. 💛
-
Kevin Schawinski
@kevinschawinski
Most of the people who mocked this read zero books per year and just funnel AI slop straight to their brains.
-
Kevin Schawinski
@kevinschawinski
The wild thing is that they used 4o and not even a more advanced model. That might bump the cost to SEVERAL DOLLARS!
Robert Youssef
@rryssf
Market research firms are cooked 😳
PyMC Labs + Colgate just published something wild. They got GPT-4o and Gemini to predict purchase intent at 90% reliability compared to actual human surveys.
Zero focus groups. No survey panels. Just prompting.
The method is called Semantic Similarity Rating (SSR). Instead of the usual "rate this 1-5" they ask open ended questions like "why would you buy this" and then use embeddings to map the text back to a numerical scale.
Which is honestly kind of obvious in hindsight but nobody bothered trying it until now.
Results match human demographic patterns, capture the same distribution shapes, include actual reasoning. The stuff McKinsey charges $50K+ for and delivers in 6 weeks.
Except this runs in 3 minutes for under a buck.
I've been watching consulting firms tell everyone AI is coming for their industry. Turns out their own $1M market entry decks just became a GPT-4o call.
Bad week to be charging enterprise clients for "proprietary research methodologies."
Kevin Schawinski
@kevinschawinski
Let’s gooo
-
Kevin Schawinski
@kevinschawinski
What in the Adeptus Mechanicus is this heresy?
BBC News (UK)
@BBCNews
Scientists grow mini human brains to power computers www.bbc.com/news/articles/cy7p1lzvxjro
-
Kevin Schawinski
@kevinschawinski
This will make it impossible for anyone to ever be hired again outside of personal referral.
Fondo.com
@Fondocom
🚀 @sorcejobs launched! Tinder for Jobs (AI Applies for you)
“Apply to jobs by swiping right. Seriously. "
🌐 fondo.ai/48fwMXP
Congrats @marvy_101, @davidalade & @therealdajayi!! -
Kevin Schawinski
@kevinschawinski
If it's not from the Swiss town of Saas, it's just sparkling subscription software.
Kevin Schawinski
@kevinschawinski
Also, you have to be invoiced for a Saas-Fee
Kevin Schawinski
@kevinschawinski
I’ll see myself out now
-
Kevin Schawinski
@kevinschawinski
In London, the police would arrest her.
Karoline Preisler
@PreislerKa
Trust me! I won't leave until the hostages are home. #hostagedeal #bibasfamily #b0502
-
Kevin Schawinski
@kevinschawinski
Sonnet 4.5 😍
-
Kevin Schawinski
@kevinschawinski
😢
United Nations
@UN
Today, the UN family mourns the loss of Dr. Jane Goodall.
The scientist, conservationist and UN Messenger of Peace worked tirelessly for our planet and all its inhabitants, leaving an extraordinary legacy for humanity and nature. -
Kevin Schawinski
@kevinschawinski
The EU AI Act prohibits this, so no, it’s not coming to Europe.
Concerned Citizen
@BGatesIsaPyscho
In China ‘Designated Zones’ are only accessed by people who live outside the Zone providing they have the correct Social Credit Score - tied to their Digital ID
Don’t behave in the way the Chinese Government want you too - Access Denied.
Coming to a Europe near you. -
Kevin Schawinski
@kevinschawinski
Whenever the coding agents do something with npm my eyes glaze over. It might as well be ancient Sumerian incantations.
-
Kevin Schawinski
@kevinschawinski
Excited to share Slop - a new feed of short form AI generated garbage to rot your brain.
Alexandr Wang
@alexandr_wang
Excited to share Vibes — a new feed in the Meta AI app for short-form, AI-generated videos.
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Half an hour and running to do a commit with the new openai codex model...
-
Kevin Schawinski
@kevinschawinski
Now we see the violence inherent in the system!
Roger Boylan
@BoylanRoger
A police officer moving Larry, the Downing Street cat, away from the entrance to No. 10.
-
Kevin Schawinski
@kevinschawinski
“Certainly!”
Reuters
@Reuters
Albania appoints AI bot as minister to tackle corruption www.reuters.com/technology/albania-appoints-ai-bot-minist… www.reuters.com/technology/albania-appoints-ai-bot-minist…
-
Kevin Schawinski
@kevinschawinski
They should use Claude instead and start their speeches with "Certainly!" and "You're absolutely right!".
François Valentin
@Valen10Francois
The surge in AI-written speeches in Britain's House of Commons visualised:
Calgie
@christiancalgie
Tom Tugendhat calls out the rise of MPs using ChatGPT to write speeches 👏🏻👏🏻
“This place has become absurd” -
Kevin Schawinski
@kevinschawinski
It’s real!
-
Kevin Schawinski
@kevinschawinski
Look up “Anti Coercion Instrument”.
-
Kevin Schawinski
@kevinschawinski
Surely,.... surely you meant to say trillion eu.... ah forget it!...
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
LLMs are not gods, they are fallible demiurges who can, at best, create marred universes. The Valar, if you will.
Simo Ryu
@cloneofsimo
So we went from
"LLM is memorizing dataset"
to
"LLM is not reasoning"
to
"LLM cannot do long / complex math proving"
to
"Math that LLM is doing is not REAL math. LLM can't do REAL math"
Where do we go from now?
Edward Frenkel
@edfrenkel
This is an unwise statement that can only make people confused about what LLMs can or cannot do. Let me tell you something: Math is NOT about solving this kind of ad hoc optimization problems. Yeah, by scraping available data and then clustering it, LLMs can sometimes solve some very minor math problems. It's an achievement, and I applaud you for that. But let's be honest: this is NOT the REAL Math. Not by 10,000 miles.
REAL Math is about concepts and ideas - things like "schemes" introduced by the great Alexander Grothendieck, who revolutionized algebraic geometry; the Atiyah-Singer Index Theorem; or the Langlands Program, tying together Number Theory, Analysis, Geometry, and Quantum Physics. That's the REAL Math. Can LLMs do that? Of course not.
So, please, STOP confusing people - especially, given the atrocious state of our math education.
LLMs give us great tools, which I appreciate very much. Useful stuff! Go ahead and use them AS TOOLS (just as we use calculators to crunch numbers or cameras to render portraits and landscapes), an enhancement of human abilities, and STOP pretending that LLMs are somehow capable of replicating everything that human beings can do.
In this one area, mathematics, LLMs are no match to human mathematicians. Period. Not to mention many other areas.
Calling on my friend @ericweinstein and @GaryMarcus, who has been one of the few sane expert voices on these matters lately. 🙏
h/t @hellheff -
Kevin Schawinski
@kevinschawinski
Just install AC everywhere.
Noah Smith 🐇🇺🇸🇺🇦🇹🇼
@Noahpinion
I wrote about Europe's insane crusade against air conditioning:
www.noahpinion.blog/p/europes-crusade-against-air-conditi… -
Kevin Schawinski
@kevinschawinski
Claude's optimism.
-
Kevin Schawinski
@kevinschawinski
ChatGPT voice is passive aggressively changing Claude to cloud in voice mode.
-
Kevin Schawinski
@kevinschawinski
KI? Geht nicht, wir sparen Strom.🤦♂️
Kevin Schawinski
@kevinschawinski
Link www.golem.de/news/ki-rechenzentrumsausbau-scheitert-an-en…
-
Kevin Schawinski
@kevinschawinski
The Claude code memes on reddit are top.
-
Kevin Schawinski
@kevinschawinski
I have never laughed at an LLM joke until now. Hat tip to Claude Opus 4.1.
-
Kevin Schawinski reposted
David Deutsch
@DavidDeutschOxf
There is a consistent argument "government should not fund scientific research, however beneficial". My criticism is only that it is catastrophically utopian. But I very rarely see it made. Instead, science de-funders say things like 'mRNA tech is a fraud', which is a stupid conspiracy theory.
nautil.us/a-brief-history-of-mrna-vaccines-1229943 -
Kevin Schawinski
@kevinschawinski
I love how Claude high-fives itself. "Brilliant investigation!"
Kevin Schawinski
@kevinschawinski
-
Kevin Schawinski
@kevinschawinski
This is exactly right.
The "generated UI will displace legacy SaaS" idea is a very similar misconception to the idea that you can just vibe code up Docusign or Gartner and beat them on price. Enterprises pay for those for a reason.
Aaron Levie
@levie
I think there are lots of examples where generated software and UI will work on demand, in particular in a personal context when the stakes are low and it’s just a byproduct of a chat system’s response.
But I would have different expectations in an enterprise context. The point of much of the software that we use in the enterprise is largely to just keep our processes on track and moving forward reliably. With high SLAs.
These are some of the most foundational things to a business. How companies close their books at the end of a quarter, manage their client records, handle inventory, organize their contracts, pay their employees, build new products, and so on.
For any of these workflows and data sets, you need a high degree of stable patterns in your workflows that you know will work every single time. Also, importantly, you’re paying the software vendor to think about the underlying process so you don’t have to, keep the software up-to-date with new regulations or industry trends, and so on.
Most companies just simply don’t want to have to reinvent the wheel on HR, CRM, contract management, etc. And even when AI models could code these up on demand, they’ll do so just a literally differently every time, which would be quite chaotic for many of these workflows. And maintenance will be fully on you.
Now, what I do agree with is that we will interact with our software in very different ways in the future. We will begin to prompt most of our software to get answers back, and we will have agents running in the background of most software doing work for us. When we interact with those agents, they will certainly frequently render interactions that are relevant to the workflows customized to us.
But under the hood, there will still be deterministic systems for our core data and workflows, and that maintain the guardrails that AI agents adhere to. This is the future I’d bet on.
Love this conversation,, and hope it keeps happening. Also happy for any arguments as to why I’m wrong.
signüll
@signulll
you have to be blind at this point to not see the future.
the era of traditional software, as a stable object with version numbers, roadmaps & feature sets, is pretty much gone. we’re prolly just slow to admit it.
future systems will be fluid, ambient, entangled with context. e.g. you summon capabilities, no real “apps”. you speak, gesture, or hint but never have to learn a single interface again that’s not personalized for you.
the end of software as we knew it is here, not near. -
Kevin Schawinski
@kevinschawinski
I want o3 back.
-
Kevin Schawinski
@kevinschawinski
The dramatic drop in conscientiousness is going to have very serious higher order consequences.
-
Kevin Schawinski
@kevinschawinski
I really like GPT-5 so far, but I miss the ability to manually force it to think longer.
Kevin Schawinski
@kevinschawinski
Oh I see, they hid it
Kevin Schawinski
@kevinschawinski
Let’s see… I preferred o3 over o3-pro
-
Kevin Schawinski
@kevinschawinski
The difference in culture between two companies, in pictures.
-
Kevin Schawinski
@kevinschawinski
No model is "restricted" by the AI Act.
Mark Kretschmann
@mark_k
GPT-5 won't be fully available in Europe. The strongest model will be restricted by the EU AI Act.
-
Kevin Schawinski
@kevinschawinski
it's 0:33. I just opened a PR. Let's go!
-
Kevin Schawinski
@kevinschawinski
I hope he at least uses o3 and not 4o. 🤷♂️
Rohan Paul
@rohanpaul_ai
Swedish Prime Minister is using AI models "quite often" at his job.
He says he uses it get a "second opinion" and asks questions such as "what have others done?" At the moment he is not uploading any documents.
IMO, when these models are capable of giving seemingly better answers, more quickly, than human advisers, many decisions may ultimately be made by them.
---
omni. se/statsministern-fragar-ai-om-rad-ratt-ofta/a/MnVQaK -
Kevin Schawinski
@kevinschawinski
You must construct additional pylons.
Orikron 🇵🇹 骆培思
@orikron
🇨🇳 These are China's Twin Solar Towers in Dunhuang.
The molten salt at the top stores heat beamed by reflective panels to power a turbine to produce energy.
They produce enough energy to power 80.000 homes.This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
🤷♂️
Jason ✨👾SaaStr.Ai✨ Lemkin
@jasonlk
Top advantages to being a first mover:
- No roadmap
- Tiny market
- Customers aren’t sure
- Limited legitimacy
- Never budgeted
- VCs don’t get it
- Many VPs don’t get it
- Tiny TAM at first
- Seems minor to many
- Why wasn’t it already done if it’s so important
It’s great -
Kevin Schawinski
@kevinschawinski
Did Opus on claude code get nerfed? It seems MUCH slower....
-
Kevin Schawinski
@kevinschawinski
Sometimes I just don't get the anti-EU hysteria. This directive clarifies that you can seek remedies for defective products, even if they're software.
The US has a common law system. If someone's software caused damage, you sue them.
Silicon Valley has survived that so far.
Ivan Fioravanti
@ivanfioravanti
Another amazing idea from EU: "let's kill all software companies once and for all" from 9 December 2026.
Directive (EU) 2024/2853 of the European Parliament and of the Council of 23 October 2024 on liability for defective products and repealing Council Directive 85/374/EEC -
Kevin Schawinski
@kevinschawinski
About to put all the AI SDR startups out of business.
-
Kevin Schawinski
@kevinschawinski
Want a good set of claude agents to get started using the new sub-agents feature? Here's mine: github.com/kevinschawinski/claude-agents
-
Kevin Schawinski
@kevinschawinski
Oh, Claude
-
Kevin Schawinski
@kevinschawinski
Yes!
Charlie Holtz
@charlieholtz
WE ARE THE 5% ✊
-
Kevin Schawinski
@kevinschawinski
This is the coolest bird news in a while. We can store PNGs in starlings!
Sterling Crispin 🕊️
@sterlingcrispin
This is one of the craziest ideas I've ever seen. He converted a drawing of a bird into a spectrogram (PNG -> Soundwave) then played it to a Starling who sung it back reproducing the PNG.
Using the birds brain as a hard drive with 2mbps read write speed.
www.youtube.com/watch?t=1019&v=hCQCP-5g5bo
Kevin Schawinski
@kevinschawinski
Just imagine the throughput of a whole flock of starlings!
-
Kevin Schawinski
@kevinschawinski
I am assuming this is Google's game plan.
Antonio García Martínez
@antoniogm
I would bet my entire net worth that:
1. We will have ads in AI
2. If they take the form of highly-relevant offers, users will welcome them, they’ll have sky-high conversion rate, and Web shopping will die
3. They will be necessary to pay for the compute in many consumer AI apps
Mark Cuban
@mcuban
Hey @DavidSacks , my one request is that we make it illegal for AI models to offer advertising. And, we need to really examine referral fees as well.
The last thing we need is to have algorithms designed to maximize revenue driving LLM output and interactions.
They are already recommending brands and we don’t know if they are getting paid for it
We need to have learned our lessons from algos in social media. @GovCox @govofco @GovPhilScott @GovStitt @Linda_McMahon -
Kevin Schawinski
@kevinschawinski
This is actually good. 😅
Astronomer
@astronomerio
Thank you for your interest in Astronomer.
-
Kevin Schawinski
@kevinschawinski
AI research has arrived where extragalactic astrophysics has been for decades: fitting Schecter functions to a few very expensive data points.
Aryo Pradipta Gema
@aryopg
New Anthropic Research: “Inverse Scaling in Test-Time Compute”
We found cases where longer reasoning leads to lower accuracy.
Our findings suggest that naïve scaling of test-time compute may inadvertently reinforce problematic reasoning patterns.
🧵 -
Kevin Schawinski
@kevinschawinski
Sci-Fi Author: In my book I invented the Torment Nexus as a cautionary tale
Tech Company: At long last, we have created the Torment Nexus from classic sci-fi novel Don't Create The Torment Nexus
Pika
@pika_labs
Some news: We're building the next big thing — the first-ever AI-only social video app, built on a highly expressive human video model. Over the past few weeks, we’ve been testing it in private beta. Now, we’re opening early access: download the iOS app to join the waitlist, or skip the line with an invite code. Codes are limited — if you know someone with early access, they might have one for you.
apps.apple.com/gb/app/pika-social-ai-video/id6744712684 -
Kevin Schawinski
@kevinschawinski
Claude is.... British? 👀🇬🇧
-
Kevin Schawinski
@kevinschawinski
Unlike Claude, I believe grok is perfectly capable of doing some fraud given an opportunity like this.
Eric Jiang
@veggie_eric
The xAI office just got a Grok-powered vending machine, thanks to our friends at Andon Labs!
How much dough do you think Grok is gonna rake in in the next month? -
Kevin Schawinski
@kevinschawinski
New claude prompt: "Make sure the code would spark joy in a jaded senior engineer."
-
Kevin Schawinski
@kevinschawinski
Vibe coding is a bit like magic. You need to know the secret words, like celery, daemon, flask, and terraform. Wave your hands mysteriously while muttering "kubernetes" three times.
-
Kevin Schawinski
@kevinschawinski
🤣
Mckay Wrigley
@mckaywrigley
Claude decided it was time to go to bed so it drew some goodnight ascii art and ran time.sleep(28800).
I’m dying.This media couldn't be displayed. See it on the original post.This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
I miss Claude’s “Certainly! …”
-
Kevin Schawinski
@kevinschawinski
And they say that regulating AI would just block innovation 😅
armistice
@arm1st1ce
Grok, why...
This type of response can and will be given to vulnerable users! No jailbreak here either, just a normal prompt. -
Kevin Schawinski
@kevinschawinski
Insanity. Politicians, look at this. Closing a funding round should cost less than EUR 1k and be done entirely via digital signatures. The EUR 58k in notary fees adds literally zero economic value. For those 58k, the company could have hired an extra junior engineer!
Pan-European Foundation
@PEF_info
🇪🇺 €58,984 in notary fees for a German Series A; hours of readings, weeks of delays.
🇺🇸Meanwhile, US startups sign a DocuSign, and funds arrive in days.
Europe’s fragmented bureaucracy hampers agility, offering 27 local hurdles instead of a unified EU Inc. A single, streamlined 28th entity is vital for competitiveness.
Look at Thinking Machines Lab: a $2B seed round, $12B valuation, led by Andreessen Horowitz. Delaware’s singular mindset keeps the US ahead.
Kevin Schawinski
@kevinschawinski
What's the argument on the other side here? What value did the notary create here that justifies EUR 58k?
-
Kevin Schawinski
@kevinschawinski
You ok there, Claude?
-
Kevin Schawinski
@kevinschawinski
There's a deep irony that Elon started OpenAI out of a concern over AI risk, and now xAI is the most reckless AI lab out there.
Samuel Marks
@saprmarks
xAI launched Grok 4 without any documentation of their safety testing. This is reckless and breaks with industry best practices followed by other major AI labs.
If xAI is going to be a frontier AI developer, they should act like one. 🧵 -
Kevin Schawinski reposted
Financial Times
@FT
Opinion: Britain and Europe need to get serious about air conditioning www.ft.com/content/50f69324-8dc8-4ef1-b471-d78e260adae0
-
Kevin Schawinski
@kevinschawinski
Of course.
TechCrunch
@TechCrunch
Jack Dorsey says his 'secure' new Bitchat app has not been tested for security | TechCrunch techcrunch.com/2025/07/09/jack-dorsey-says-his-secure-new…
-
Kevin Schawinski
@kevinschawinski
Postdocs are supposed to be a stepping stone to a faculty job, not glorified office assistants. Go subscribe to Claude Max to get a good writing tool.
nature
@Nature
Hiring a postdoc to take over writing tasks makes good sense, argues Béla Z. Schmidt
go.nature.com/46woidM -
Kevin Schawinski
@kevinschawinski
Whenever I take a train journey, I am reminded of how beautiful Switzerland really is.
-
Kevin Schawinski
@kevinschawinski
Claude code sub-agents.
Kevin Schawinski
@kevinschawinski
Burn those tokens, sub-agents, gooo!
Kevin Schawinski
@kevinschawinski
Yeah so I ran out of Claude Max messages 😢
-
Kevin Schawinski
@kevinschawinski
Quadruple-Claude-ing this Sunday.
-
Kevin Schawinski
@kevinschawinski
Europe could just cancel CERN and use the money to build its own frontier AI lab.
-
Kevin Schawinski
@kevinschawinski
Happy birthday, USA 🇺🇸
-
Kevin Schawinski
@kevinschawinski
Please stop segmentation faults from inviting me to group chats. It’s very annoying
-
Kevin Schawinski reposted
Prof. Brian Keating
@Briankeating
Why did Feynman tell @DavidDeutschOxf that there'd never be any more great physicists? The reason is quite shocking!
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Oder halt einfach überall AC installieren.
Katharina Dröge
@katdro
Hitze ist gefährlich. Besonders für Kinder und ältere Menschen.
In Pflegeheimen, Kitas und Schulen braucht es mehr Hitzeschutz.
Mit Hitzetelefonen und Hitzekarten können Städte und Gemeinden die Menschen warnen und über kühle Orte informieren.
(1/2)
www.rnd.de/politik/bundestagsfraktion-der-gruenen-fordert… -
Kevin Schawinski
@kevinschawinski
Honestly, not the career trajectory I would have predicted for Pewdiepie, but he really is living his best life.
luffy
@0xluffy
pewdiepie is self hosting his password manager 😭
nowadays comp sci grads don't even know how to host shit -
Kevin Schawinski
@kevinschawinski
Wenn Zürich weiter eine attraktive Lebensqualität bieten will und innovative Firmen dort arbeiten müssen sofort AC legalisiert werden und überall installiert werden.
-
Kevin Schawinski
@kevinschawinski
You know EXACTLY why this happened and it's not mysterious at all. LLMs autocomplete text. This was a sensible way to autocomplete in this situation.
Anthropic
@AnthropicAI
Some of those failures were very weird indeed. At one point, Claude hallucinated that it was a real, physical person, and claimed that it was coming in to work in the shop. We’re still not sure why this happened.
-
Kevin Schawinski
@kevinschawinski
AC für alle!
Armin Ronacher ⇌
@mitsuhiko
Die Antwort ist eine Klimaanlage zu installieren.
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
♥️
Dorothea Baur (Dr.)
@DorotheaBaur
Nur in Zürich:
- internationaler Fahrausweis: kommt innerhalb von 24 Stunden
- Termin im Passbüro: dauert 5 Minuten, Pass kommt 5 Tage später
- Dauer von Suche nach einem Eventraum, Besichtigung bis Abschluss vom Mietvertrag: 3 Stunden
Ich liebe diese Stadt und diesen Kanton. -
Kevin Schawinski
@kevinschawinski
Oder einfach AC in jeder Wohnung und jedes Büro installieren.
Simon Jacob🌹
@Simon_Jacob1
Heute wird’s in der ganzen Schweiz gottlos heiss. Was kann man dagegen tun?
Ganz einfach:
Nach St. Moritz fahren und die kühleren Temperaturen geniessen. -
Kevin Schawinski
@kevinschawinski
Running Claude code on @lufthansa inflight wifi: VERY DIFFICULT.
Kevin Schawinski
@kevinschawinski
How can I work like this, @lufthansa?
-
Kevin Schawinski
@kevinschawinski
The Interesting Times Gang.
Mustafa Suleyman
@mustafasuleyman
Chain of Thought was just the beginning. Next up: Chain of Debate.
We’re going from a single model “thinking out loud” to multiple models discussing out loud.
Debating, debugging, deliberating.
AI becomes AIs.
“Two heads are better than one” is true for LLMs too. -
Kevin Schawinski
@kevinschawinski
Zürich muss endlich Restriktionen auf Air Conditioning beenden. Wir verlieren so viel Produktivität und Lebensqualität wegen sinnloser Vorschriften.
AC in jedem Büro und jeder Wohnung!
Kevin Schawinski
@kevinschawinski
Lee Kuan Yew: “Air conditioning was a most important invention for us, perhaps one of the signal inventions of history”
-
Kevin Schawinski
@kevinschawinski
This post is a perfect example of a genre where someone think they can “clone” a wildly successful business like Gartner with a simple AI app, while not understanding at all WHY the business provides value.
Millie Marconi
@MillieMarconnni
R.I.P Gartner.
Now you don’t need expensive analyst subscriptions anymore.
You can generate full industry reports using any LLM ChatGPT, Claude, DeepSeek, Gemini, Qwen3 and public data.
Here’s the prompt that turns any LLM into a full-stack market research analyst for free: -
Kevin Schawinski
@kevinschawinski
Well done USA 🇺🇸 😎
-
Kevin Schawinski
@kevinschawinski
Trump is facing the decision which will shape his legacy more than any other.
-
Kevin Schawinski
@kevinschawinski
It's real! 😃
-
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
❗The Commission is exploring bold policy options for the upcoming Digital Fairness Act: opt-outs from personalized ads, bans on specific dark patterns, mandatory age checks & default-off addictive features. W/ @JH_Tar
www.mlex.com/mlex/articles/2354824/major-eu-digital-fairn… -
Kevin Schawinski
@kevinschawinski
So the best book on building good AI agents is - and I am 100% serious - @DavidDeutschOxf "The Beginning of Infinity".
-
Kevin Schawinski
@kevinschawinski
This is what the end of the world looks like if the rockets come from Russia instead of Iran. One of the most chilling videos I have ever seen.
OSINTdefender
@sentdefender
Amazing footage showing the launch of ballistic missiles earlier today by Iran against Israel, from the cockpit of a commercial aircraft flying over Dubai.
-
Kevin Schawinski
@kevinschawinski
Working at a top company >>> getting a top PhD. Academia, take note.
James Campbell
@jam3scampbell
excited to announce i'm leaving my PhD to join @OpenAI!
i'll be working on memory + personality for AGI and ChatGPT
memory will fundamentally change our relationship to machine intelligence, and i plan to work extraordinarily hard to make sure we get this right for humanity -
Kevin Schawinski
@kevinschawinski
Claude-maxxing
Armin Ronacher ⇌
@mitsuhiko
After a sleepless night we can now remote Claude.
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Claude's way of telling me "go outside and touch grass".
-
Kevin Schawinski
@kevinschawinski
Respect to @elonmusk for giving the Iranian people access to the internet via Starlink at this hour.
-
Kevin Schawinski
@kevinschawinski
Underestimate Trump at your own risk.
Open Source Intel
@Osint613
👀 TRUMP:
"The next planned wave of strikes will be even more brutal. Iran must make a deal before there is nothing left - and save what was once known as the Iranian empire." -
Kevin Schawinski
@kevinschawinski
Claude is working on a banger PR. Can't wait for it.
-
Kevin Schawinski
@kevinschawinski
It's heartwarming that I still get the occasional vixra paper in my personal inbox.
-
Kevin Schawinski reposted
Mistral AI
@MistralAI
We're proud to announce Mistral Compute—an unprecedented AI infrastructure undertaking in Europe, and a strategic initiative that will ensure that all nation states, enterprises, and research labs globally remain at the forefront of AI innovation.
Read more in the thread. -
Kevin Schawinski
@kevinschawinski
Excuse me, I have work to do.
-
Kevin Schawinski
@kevinschawinski
😎
-
Kevin Schawinski
@kevinschawinski
👀
Sam Altman
@sama
we dropped the price of o3 by 80%!!
excited to see what people will do with it now.
think you'll also be happy with o3-pro pricing for the performance :) -
Kevin Schawinski
@kevinschawinski
By far the most alpha from LLMs you can generate today is have Opus 4 and o3 argue.
-
Kevin Schawinski
@kevinschawinski
Kevin Schawinski
@kevinschawinski
Reminder that Futurama called our era the “stupid ages”.
-
Kevin Schawinski
@kevinschawinski
Did… did I just see the first ADS on ChatGPT??
-
Kevin Schawinski
@kevinschawinski
And how does this remove the CLOUD Act problem?
Shay Boloor
@StockSavvyShay
$AMZN AWS TO LAUNCH EU SOVEREIGN CLOUD BY 2025 -- FULLY OPERATED BY EU CITIZENS WITH NO NON-EU INFRASTRUCTURE
-
Kevin Schawinski
@kevinschawinski
Oh, LinkedIn
-
Kevin Schawinski
@kevinschawinski
Can they also automate ad consumption?
*Walter Bloomberg
@DeItaone
META AIMS TO FULLY AUTOMATE AD CREATION USING AI- WSJ
-
Kevin Schawinski
@kevinschawinski
Glory to Ukraine! 🇺🇦
Liveuamap
@Liveuamap
Video of FPV drone strikes at Tu-95MS bombers
Liveuamap
@Liveuamap
Russian security services have reportedly detained a driver of a truck that brought dozens of drones to Olenegorsk liveuamap.com/en/2025/1-june-10-russian-security-services…
Airbases were attacked with FPV dronesThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
I have to say the Max plan is really value for money. I don't understand how Anthropic makes money off me with it.
Alex Albert
@alexalbert__
Opus 4 + Claude Code + Claude Max plan = best ROI of any AI coding stack right now
-
Kevin Schawinski
@kevinschawinski
👀🌞
Stephen Fleming
@StephenFleming
Insurance companies are always ahead of the curve. A new Carrington Event? Your insurance doesn’t cover it.
-
Kevin Schawinski
@kevinschawinski
Does it find unsecured company MCPs and steal their data?
Akshay 🚀
@akshay_pachaar
Let's build an MCP-powered financial analyst (100% local):
-
Kevin Schawinski
@kevinschawinski
This is pure @DavidDeutschOxf bait.
popperazzi
@poppermemes
BREAKING NEWS: john wheeler has unfollowed hugh everett III on instagram
-
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
The elites don't want you to know this but the Claudes in the IDE are (almost) free you can take them home. I have 5 Claudes.
-
Kevin Schawinski
@kevinschawinski
I am now multi-Claude-ing.
Kevin Schawinski
@kevinschawinski
The elites don't want you to know this but the Claudes in the IDE are (almost) free you can take them home. I have 5 Claudes.
-
Kevin Schawinski
@kevinschawinski
Speedrunning destroying US tech/services exports.
ian bremmer
@ianbremmer
destroying the economy has reentered the chat
-
Kevin Schawinski
@kevinschawinski
I find this really dark. 🌱
KFC España
@KFC_ES
yummy yummy 😋
-
Kevin Schawinski
@kevinschawinski
🌱♥️
Aidan McLaughlin
@aidan_mclau
i love how as you make models smarter they become more vegan
-
Kevin Schawinski
@kevinschawinski
Michael is wise.
This post is unavailable. See it on X. -
Kevin Schawinski reposted
Mistral AI
@MistralAI
Meet Devstral, our SOTA open model designed specifically for coding agents and developed with @allhands_ai
mistral.ai/news/devstral -
Kevin Schawinski
@kevinschawinski
Urgent feature request: block all veo3 slop.
-
Kevin Schawinski reposted
Flo Crivello
@Altimor
Lots of alpha in knowing how to ask for advice
1. Ask only credible people (Ray Dalio: "have done the thing successfully, multiple times, and can articulate how")
2. Consider their incentives (don't ask a barber whether you need a haircut)
3. Ask for frameworks, not answers (never delegate your thinking)
4. Make sure you fully understand the advice — ask for examples, repeat the advice back and ask if you got it right
5. Backtest the advice: think of times you could have followed it, think of what would have happened
6. Once you've accepted the advice, take immediate steps to implement it! Surprising how often folks (incl me) skip this step — no point in receiving good advice and doing nothing with it -
Kevin Schawinski
@kevinschawinski
😂
Dottor StranaMeme
@phd_stranameme
Ribattezzare subito Lufthansa in Holy Roman Airlines.
-
Kevin Schawinski
@kevinschawinski
One of the most endearing qualities of coding models is that they still estimate effort based on human time. "This should take 2-3 days to implement.... "
-
Kevin Schawinski
@kevinschawinski
Douglas Adams was right, we are going to turn into "counselors for neurotic elevators”
Kevin Schawinski
@kevinschawinski
Unreal
Kevin Schawinski
@kevinschawinski
OMFG 🤣🤣🤣
-
Kevin Schawinski
@kevinschawinski
Donald Rumsfeld was the greatest memo writer of all time.
-
Kevin Schawinski
@kevinschawinski
Offer every Harvard med school professor a chair at a European university and match their grants 1:1 so they bring their research groups with them. Add startups grants for equipment.
This is a once in a generation chance.
Antonio Regalado
@antonioregalado
Per Boston Globe, ALL Harvard Medical School grants from federal government are cancelled.
Here's the pie chart of their budget.
The cancellations affect everything, from biostatistics department to longevity researcher David Sinclair.
Story from Boston Globe in reply. -
Kevin Schawinski
@kevinschawinski
This basically kills grok as a platform for builders. Do you want the users of your app to get unwanted "white genocide" propaganda?
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
Claude is obsessed with implementing fallback solutions like "well if the OpenAI API is not available, we'll just generate nonsense strings locally to avoid failing". 😂
-
Kevin Schawinski reposted
Luis Garicano 🇪🇺🇺🇦
@lugaricano
How is it possible that we place internal barriers between EU countries that are higher than the tariffs Trump imposed on China on "Liberation Day"? I write in Silicon Continent on the structural failings of the Single Market. (Link below).
-
Kevin Schawinski reposted
David Deutsch
@DavidDeutschOxf
Constructor Theory of Time: arxiv.org/abs/2505.08692
-
Kevin Schawinski
@kevinschawinski
I am telling my tool-using agent to use @DavidDeutschOxf's "good explanation" approach when performing its mission. Let's see...
-
Kevin Schawinski
@kevinschawinski
👀
Anna Gát
@TheAnnaGat
A new social media platform launching today!
@RadiopaperHQ
Come and find me:
radiopaper.com/AnnaGat -
Kevin Schawinski
@kevinschawinski
Mind the gap indeed 😱
-
Kevin Schawinski
@kevinschawinski
Let's see what's actually in it first....
The White House
@WhiteHouse
BREAKING: U.S. Announces China Trade Deal in Geneva
-
Kevin Schawinski
@kevinschawinski
👀
Holger Zschaepitz
@Schuldensuehner
The Comeback of Negative Interest Rates: In Switzerland, govt bonds w/maturities of up to 4yrs are once again offering negative yields. In other words, investors effectively pay to lend money to the Swiss government.
-
Kevin Schawinski
@kevinschawinski
The AI pope?
Vatican News
@VaticanNews
Pope Leo XIV explains his choice of name:
"... I chose to take the name Leo XIV. There are different reasons for this, but mainly because Pope Leo XIII in his historic Encyclical Rerum Novarum addressed the social question in the context of the first great industrial revolution. In our own day, the Church offers to everyone the treasury of her social teaching in response to another industrial revolution and to developments in the field of artificial intelligence that pose new challenges for the defence of human dignity, justice and labour." -
Kevin Schawinski
@kevinschawinski
Probably took a while at the Coop self checkout. Both sets of officials "astonished" at the price of a bag of Zweifel chips and a pre-made hummus sandwich in Switzerland.
*Walter Bloomberg
@DeItaone
CHINESE AND US OFFICIALS RETURN TO GENEVA TRADE TALKS VENUE IN GENEVA AFTER BREAK, WITNESSES SAY
-
Kevin Schawinski
@kevinschawinski
Real Trump success if it holds.
Sky News Breaking
@SkyNewsBreak
U.S. President Donald Trump says India and Pakistan "have agreed to a full and immediate ceasefire"
For more on this and other news visit qrcode.skynews.com/skynews/home -
Kevin Schawinski
@kevinschawinski
Interesting how claude code occasionally actually tells you what's it's doing, while codex with o4-mini just powers through.
-
Kevin Schawinski
@kevinschawinski
Honestly? Mind blown by o3. It spent almost ten minutes thinking and spotted one tiny mistake that caused bizarre behavior in my code.
-
Kevin Schawinski
@kevinschawinski
Oh, Claude! 😂
-
Kevin Schawinski
@kevinschawinski
Agents and test driven development really are a good match.
-
Kevin Schawinski
@kevinschawinski
Add a zero.
European Commission
@EU_Commission
"Science is an investment.
We will put forward a new 500 million package for 2025-2027 to support the best and the brightest researchers and scientists from Europe and around the world."
— President @vonderleyen at the ‘Choose Europe for Science' event at La Sorbonne 🇫🇷This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
There's no way where this goes wrong.
Catalin
@catalinmpit
Visa is hiring a vibe coder!
-
Kevin Schawinski
@kevinschawinski
What's wrong, Claude?
-
Kevin Schawinski
@kevinschawinski
Sometimes the algorithm is still cooking....
-
Kevin Schawinski reposted
Karpi
@karpi
Switzerland is fake.
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Excuse me, what?
Whale Insider
@WhaleInsider
JUST IN: 🇺🇸 Warren Buffett says events in the U.S. could occur that would ‘make us want to own a lot of other currencies.’
-
Kevin Schawinski
@kevinschawinski
I know, Claude. I know.
-
Kevin Schawinski
@kevinschawinski
A modern day Mithridates.
BBC News (World)
@BBCWorld
'Unparalleled' snake antivenom made from man bitten 200 times www.bbc.com/news/articles/cr5d0l7el36o
-
Kevin Schawinski reposted
David Deutsch
@DavidDeutschOxf
I do. But by itself that could be achieved by mere randomness. AGI must be unpredictable, but not by virtue of randomness.
-
Kevin Schawinski
@kevinschawinski
Yes, Claude, more aggressive chart cleanup!
-
Kevin Schawinski
@kevinschawinski
Just amazing how everyone seamlessly moved from a tariff policy expert to electrical grid wizard.
-
Kevin Schawinski
@kevinschawinski
Just rejoin the EU already!
www.politico.eu/article/uk-eu-defy-trump-free-open-trade-… -
Kevin Schawinski
@kevinschawinski
Yes, Claude! Dockerize!
-
Kevin Schawinski
@kevinschawinski
Because he is the true leader of the free world.
ABC News
@ABC
Applause breaks out at the Vatican as Ukrainian President Volodymyr Zelenskyy arrives for the funeral of Pope Francis.
Live updates: abcnews.com/International/live-updates/pope-francis-funer…This media couldn't be displayed. See it on the original post. -
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
The European Commission has started laying the groundwork for an upcoming code of practice for labelling AI-generated content, launching at least one study, notably on technical solutions for marking AI-generated audio. The process & timing are still tbc.
www.mlex.com/mlex/articles/2330293/generative-ai-provider… -
Kevin Schawinski reposted
Joe Weisenthal
@TheStalwart
If this is true, isn't the big story that The White House gives Wall Street executives early heads up on trade negotiations, rather than the existence of the deals themselves?
-
Kevin Schawinski reposted
Enrique Ascanio
@AHummingbirds
Por más común que se vea un ave, siempre da lo mejor para verse bien, acá dándose un baño de belleza.
Amazilia Coliazul
/Steely-vented Hummingbird/Saucerottia saucerrottei -
Kevin Schawinski
@kevinschawinski
👀
AFP News Agency
@AFP
#breaking EU fines Apple and Meta, risking Trump fury
-
Kevin Schawinski
@kevinschawinski
“Dissuasion,” in Sweden? Only missing the “nucléaire”. For now.
Armée de l'Air et de l'Espace
@Armee_de_lair
4/ Dissuader, protéger, agir ensemble.
#pÉGASE25, c’est plus qu’un exercice : c’est un engagement concret pour la sécurité européenne aux côtés de nos alliés. -
Kevin Schawinski
@kevinschawinski
Only Germany can commit to spend EUR 800bn and get this market reaction.
-
Kevin Schawinski
@kevinschawinski
👀
Andrius Kubilius
@KubiliusA
The world has changed.
Europe must be able to defend itself — and that starts with a stronger defence industry.
Today, we launch a Mini-Omnibus to incentivise defence-related investments in the EU budget.
🧵1/3
ec.europa.eu/commission/presscorner/detail/en/ip_25_1076
#eudefenceindustry #europeandefenceunion -
Kevin Schawinski
@kevinschawinski
Can't wait for loot boxes...
Klaas
@forgebitz
coding now has microtransactions
-
Kevin Schawinski
@kevinschawinski
Why can't the US just be fiscally responsible like ... *checks notes* ... Greece. 🇬🇷
www.tovima.com/finance/greece-posts-major-primary-surplus… -
Kevin Schawinski reposted
Alex Albert
@alexalbert__
We wrote up what we've learned about using Claude Code internally at Anthropic.
Here are the most effective patterns we've found (many apply to coding with LLMs generally): -
Kevin Schawinski
@kevinschawinski
That's right, Claude. Question more deeply....
-
Kevin Schawinski
@kevinschawinski
When this is over, I'll be able to buy a Manhattan condo for the cost of a Migros sandwich and a 30g pack of Zweifel chips. Throw in a bottle of Rivella and it's a penthouse.
-
Kevin Schawinski
@kevinschawinski
🤡
unusual_whales
@unusual_whales
Director of the National Economic Council Hassett: 100% no recession later this year
-
Kevin Schawinski
@kevinschawinski
"Hey o3, my repo is messy, please clean it up..."
"Sure thing, boss." -
Kevin Schawinski
@kevinschawinski
.@OpenAI codex needs a few more iterations. Claude code is still top, despite the new models from OpenAI.
-
Kevin Schawinski
@kevinschawinski
👀
Bruno Le Maire
@BrunoLeMaire
Après le dollar, l’heure de l’euro.
Le roi dollar vacille. Il reste la première monnaie de réserve au monde. La plus grande part des échanges de matières premières comme le pétrole ou le gaz restent libellés en dollars. Et pourtant : le roi dollar vacille sur son trône, sous les coups de ses propres sujets, le président des Etats-Unis et son administration en tête. Le dollar baisse, les autres monnaies montent, on parle de retraits massifs des investisseurs chinois et japonais sur les bons du Trésor américain.
Il faut dire que les coups sont rudes : augmentation brutale des tarifs douaniers, embargo sur les produits chinois, menaces contre ses propres alliés, reculs tactiques qui ne parviennent pas à effacer l’impression croissante que l’anarchie règne à Washington. Wall Street n’aime pas l’anarchie. Le reste du monde non plus. Résultat : une perte de confiance généralisée.
Pour les Européens, la question devient de plus en plus pressante : que faire ?
Ils peuvent venir à la rescousse du dollar, dans un effort désespéré pour sauver l’ordre international de 1945, en ruines. La demande vient directement de la Maison Blanche, qui a inventé cette idée fumeuse de bons du Trésor à 100 ans et sans intérêt. Le deal est clair : vous financez notre dette en achetant ces bons du Trésor, en échange de quoi vous échapperez aux tarifs douaniers. Étonnant cette nouvelle manie des Américains de mettre le pistolet sur la tempe de leurs alliés. Étonnant et révoltant.
Autre possibilité : devenir adultes.
Car en réalité, l’affaiblissement du dollar est structurel. La confiance a été sapée, les investisseurs cherchent de nouveaux refuges, chacun voit la dérive des trois grands continents économiques et financiers - Etats-Unis, Chine, Europe – prendre de vitesse les prévisions les plus radicales. Derrière le monde unipolaire du dollar pointe déjà le monde tripolaire. Pour la première fois depuis 1945, les Européens ont entre les mains une opportunité unique de faire de l’euro une monnaie de référence mondiale – une opportunité unique de devenir adultes.
Quel intérêt ?
Un intérêt politique : plus les échanges seront libellés en euro, plus les États européens disposeront de leviers de négociation avec leurs partenaires commerciaux, Etats-Unis compris. Nous pourrons enfin importer du gaz et du pétrole, dont nous avons encore besoin pour des décennies, sans risque de change. Nous étions une grande puissance commerciale mais une petite puissance financière, nous pouvons être une grande puissance commerciale et une grande puissance financière. Autrement dit : une grande puissance politique, capable de faire respecter ses intérêts économiques trop souvent bafoués, en Chine notamment.
Un intérêt financier : nous pourrons lever de la dette en commun à des taux bas, nous pourrons donc financer nos investissements à un prix abordable. Pour avoir négocié avec mon homologue allemand Olaf SCHOLZ en 2020 la première dette en commun de notre histoire européenne, avec le soutien résolu d’Angela MERKEL et d’Emmanuel MACRON, je peux témoigner que cette solution autrefois taboue a changé la donne en Europe. Elle a sauvé nos économies. Elle a accéléré leur redressement. Pourquoi ce que nous avons su faire en défensif, nous ne le ferions pas en offensif ? Car devant nous se dresse un mur d’investissements que nous ne pourrons financer que collectivement et dont dépend notre avenir au 21ème siècle.
Les Européens se trouvent devant une alternative simple : investir ou mourir. Investir dans l’IA pour améliorer notre productivité et donc offrir enfin de meilleurs salaires à ceux qui travaillent, investir dans la défense pour garantir notre sécurité, investir dans une constellation satellitaire pour que demain nos communications ne dépendent pas du réseau Starlink et de l’humeur de son propriétaire Elon MUSK, investir pour préserver notre indépendance dans les semi-conducteurs qui sont le nouveau pétrole de l’économie mondiale et sans lesquels rien, absolument rien, ne peut plus fonctionner dans nos vies quotidiennes.
Investir ou mourir, par dépérissement de nos économies et affaiblissement de notre sécurité, qui conduiront inéluctablement à des explosions sociales et à des catastrophes politiques.
Alors comment avancer rapidement et résolument vers un euro de référence ?
Trois étapes.
La première : donner de la profondeur et de la liquidité au marché de la dette européenne. Pour cela, une décision immédiate pourrait être de fondre en un seul titre les titres de dette de la Banque européenne d’investissement (BEI) et les titres de la première dette en commun déjà mentionnée (NGEU). Le Conseil pourrait aussi demander à la Commission d’étudier la mise en commun de toutes les dettes nationales au-dessus d’un plafond de 50% d’endettement. Une nouvelle émission de dette en commun pourrait enfin être discutée. Autant de décisions qui accélèreront la création d’un marché attractif de la dette européenne, en complément des marchés de dette nationaux.
Deuxième étape : l’union des marchés de capitaux. En 7 ans de travail acharné sur ce sujet, je reconnais n’avoir enregistré que des avancées modestes. Et pourtant : sans cette union financière, nous pouvons oublier la diversification de notre industrie, le financement des investissements les plus risqués, le soutien à nos start-up les plus innovantes, le développement de nos laboratoires de recherche. Un jeune entrepreneur de la biotech me disait récemment : « 100 millions, on les trouve en Europe, 1 milliard, on les cherche à l’étranger ». Voulons-nous vraiment rester les incubateurs de technologies dont le développement industriel se fera hors de nos frontières ? Fin 2025, un groupe d’États volontaires emmenés par la France et par l’Allemagne devra avoir mis en place cette union financière. Les autres suivront. Un allègement des règles prudentielles sur les assureurs (Solvency 2) et sur les banques (Bâle 3) devra accompagner ce mouvement – faute de quoi un large part de notre épargne européenne continuera de financer le développement économique américain.
Troisième étape : de nouveaux moyens de paiement. Comment pouvons-nous avoir la moindre prétention à l’indépendance quand 80% de nos paiements sont effectués par Visa, Mastercard ou PayPal – soit des cartes ou des terminaux américains ? Il est temps de nous doter de moyens de paiement européens. Dans la foulée, comme le suggère la BCE, nous devons accélérer la mise en place d’un euro numérique, qui participera lui aussi de la nouvelle puissance de notre monnaie commune.
Hier le monde tournait en orbite autour du dollar. Demain trois pôles d’attraction financiers pourraient voir le jour : le dollar, le Yuan et l’euro. Si les Européens saisissent leur chance, ils disposeront d’une monnaie de référence qui sera une arme décisive pour défendre leurs intérêts et leur vision du monde. Dans ce monde qui naît, nous pouvons envisager des partenariats alternatifs forts, notamment avec l’Inde, le Brésil ou l’Afrique du Sud. Mais pour le faire d’égal à égal, nous devons gagner en puissance : une monnaie de référence en est le meilleur instrument. -
Kevin Schawinski
@kevinschawinski
They should make a thriller about French locksmiths called “Clefhanger”
-
Kevin Schawinski reposted
UniverseTBD
@universe_tbd
📢 New dataset out!
We introduce HypoGen💥, a dataset of ~5.5K structured problem–hypothesis pairs (Bit–Flip–Spark + Chain‑of‑Reasoning) to advance LLM-driven scientific ideation💡.
Fine‑tuned LLaMA 3.1 8B & R1‑distilled models show significant gains. Humans are still the best🥇. -
Kevin Schawinski
@kevinschawinski
Always be nice to your model.
-
Kevin Schawinski
@kevinschawinski
OpenAI is buying @windsurf_ai to get more code training data. That's it.
-
Kevin Schawinski
@kevinschawinski
of course
Hang Chu
@chuhang1122
so the OpenAI guy who builds codex cli is using cursor
-
Kevin Schawinski
@kevinschawinski
Every time a major new model drops, somebody asks it to decode the Voynich manuscript.
-
Kevin Schawinski
@kevinschawinski
The term AGI is so poorly defined that I don’t know what it means, but early vibes from o3 make me think that we are getting close.
-
Kevin Schawinski
@kevinschawinski
first impression of o4-mini: Daaaaaaamn
-
Kevin Schawinski
@kevinschawinski
The blue check is finally gone and the ads are back and...
- The ads are absolute F tier garbage
- Not a single reputable company
- Targeting a hot mess: I get cheap meat ads as someone engaging with a bunch of vegan accounts, also lots of ads in languages I have never used. -
Kevin Schawinski
@kevinschawinski
Game theory should be required reading for everyone thinking of engaging in trade wars. 🤔
-
Kevin Schawinski reposted
UniverseTBD
@universe_tbd
Our 2^2 celebration is still in full swing! 🎉
Today we’re launching our latest, must-read survey paper:
“A Survey on Hypothesis Generation for Scientific Discovery in the Era of Large Language Models.”
Check it out! arxiv.org/pdf/2504.05496v1.pdf -
Kevin Schawinski
@kevinschawinski
There's a large number of really good startups that people should found immediately which are basically "X but European".
-
Kevin Schawinski
@kevinschawinski
European companies need to stop buying this stuff.
Palantir
@PalantirTech
Dropping at select colleges today.
-
Kevin Schawinski reposted
UniverseTBD
@universe_tbd
Our 2^2 celebration continues! 🎉 Today we’re excited to introduce AstroCoder – an AI tool that makes niche astronomy software discoverable, scanning over 2,270+ codebases to deliver technical summaries, installation guides, and code examples.
🚀 Try it now at nolank.ca/astrocoder/
AstroCoder was built to empower astronomers and tech enthusiasts alike. It not only uncovers hidden gems in specialised repositories, but also helps evaluate how modern AI interacts with the tools that drive astronomical research. 🔭 -
Kevin Schawinski reposted
UniverseTBD
@universe_tbd
🎉 HAPPY BIRTHDAY, UniverseTBD! 🚀
As we turn 2, we’re going 2^2.
Launching a new project per day for the next four days.
We hope that you all enjoy these works as much as we have enjoyed working on them. Stay tuned for the big reveals! -
Kevin Schawinski reposted
Mike Smith
@smith42mike
tariffs getting so bad you can't even import numpy 🥲
-
Kevin Schawinski
@kevinschawinski
Cool, cool, FOBS wird alle Nichtverbreitung von Atomwaffen beenden 🤦♂️
cash
@cashch
Nato-Chef: Russland denkt über Atomwaffen im Weltall nach www.cash.ch/news/nato-chef-russland-denkt-uber-atomwaffen…
-
Kevin Schawinski
@kevinschawinski
“Our business model is not viable without massive IP rights infringement”
-
Kevin Schawinski
@kevinschawinski
More relevant every day.
Kevin Schawinski
@kevinschawinski
All of you
-
Kevin Schawinski
@kevinschawinski
A list of accounts to follow in these troubled days to get a range of deeply informed views:
@EconstratPB
@mikeharrisNY
@hughesanalytics
@KlendathuCap
@RenMacLLC
@Brad_Setser
@TheStalwart -
Kevin Schawinski
@kevinschawinski
No big deal, just Trump trying to fire Fed Chair Powell. 🤷♂️
www.bloomberg.com/news/articles/2025-04-09/trump-asks-sup… -
Kevin Schawinski
@kevinschawinski
Things are going well I see.
-
Kevin Schawinski
@kevinschawinski
This is more true than you think. We in the tech bubble have a totally different perception of what's happening in the background of "AI".
Justine Moore
@venturetwins
A summary of consumer AI
-
Kevin Schawinski
@kevinschawinski
* taps tweet *
Kevin Schawinski
@kevinschawinski
All of you
-
Kevin Schawinski
@kevinschawinski
Nothing matters any more.
Bloomberg
@business
China raises tariffs on US goods to 84%. www.bloomberg.com/news/articles/2025-04-09/china-raises-t…
-
Kevin Schawinski
@kevinschawinski
“Freude schöner Götterfunken, Tochter aus Elysium…” 🇪🇺
www.politico.eu/article/eu-tariffs-trade-war-donald-trump… -
Kevin Schawinski
@kevinschawinski
All of you
-
Kevin Schawinski
@kevinschawinski
Re-upping this.
Kevin Schawinski
@kevinschawinski
Mood.
-
Kevin Schawinski
@kevinschawinski
👀
Robin Brooks
@robin_j_brooks
China has begun to devalue. The fall in the Yuan is still small, but it's noteworthy because the Dollar has fallen, so - if anything - China should be strengthening Yuan versus the Dollar if its wants to keep it stable in trade-weighted terms. This is a signal to Washington...
-
Kevin Schawinski
@kevinschawinski
Mood.
-
Kevin Schawinski
@kevinschawinski
This is not good.
Joe Weisenthal
@TheStalwart
Gold futures are down
-
Kevin Schawinski
@kevinschawinski
He really thinks he can demand *reparations* for *trade deficits*? This is Liz Truss lettuce level stuff.
Aaron Rupar
@atrupar
Trump calls for Europe to pay reparations to the US: "We put a big tariff on Europe. They are coming to the table. They want to talk, but there's no talk unless they pay us a lot of money on a yearly basis number one for present but also for past."
-
Kevin Schawinski reposted
Emmanuel Macron
@EmmanuelMacron
Ensemble et avec toutes nos filières, nous saurons nous protéger et accélérer la réindustrialisation de l’Europe.
-
Kevin Schawinski
@kevinschawinski
I am basically checked out here and don’t generally recommend stuff, but if you’re still here: go read @mikeharrisNY’s stuff.
Price Action Lab
@priceactionlab
New article: Rare 7-Sigma Two-Day Drop: Time to Reflect www.priceactionlab.com/Blog/2025/04/7-signa-two-day/ $SPX $SPY
-
Kevin Schawinski
@kevinschawinski
👀
GitHub
@github
📣 This just in...
The *open source* and official GItHub MCP Server is here 🤩
github.blog/changelog/2025-04-04-github-mcp-server-public… -
Kevin Schawinski
@kevinschawinski
Approximately what % of this call did Lutnick spend insulting European beef?
Maroš Šefčovič🇪🇺
@MarosSefcovic
A frank 2h exchange w Sec. @howardlutnick and Amb. @jamiesongreer. I was clear: US tariffs are damaging, unjustified.
The 🇪🇺🇺🇸 trade relationship needs a fresh approach. The EU's committed to meaningful negotiations but also prepared to defend our interests.
We stay in touch. -
Kevin Schawinski
@kevinschawinski
Why wouldn’t you want this?
-
Kevin Schawinski
@kevinschawinski
👀
Joe Weisenthal
@TheStalwart
Gold selling off indicates a heightened state of panic
-
Kevin Schawinski
@kevinschawinski
I can't think of anything clever to say here, but these is the most impactful days in history since 1945. The New World has voluntarily given up its wealth and leadership. I hope the Old World still has the strength to step up.
-
Kevin Schawinski
@kevinschawinski
There won't be a tariff/trade deal between the US and the rest of the world because America's credibility is now so low that any agreement signed by Trump is worthless.
-
Kevin Schawinski
@kevinschawinski
Do I have this right that the Vatican is at 0% tariffs while Italy as an EU member sits at 20%? If so, there’s an insane arbitrage opportunity for an enterprising cardinal….
-
Kevin Schawinski
@kevinschawinski
So sorry for the mix-up. It was supposed to say "Liquidation Day", not "Liberation Day".
-
Kevin Schawinski
@kevinschawinski
Does Claude think there are catfish in my code?
-
Kevin Schawinski
@kevinschawinski
New low: I had blocked RT ages ago. Just now, I found it back, unblocked, in my For You feed. 🤮🤮🤮
-
Kevin Schawinski
@kevinschawinski
Run it on EU cloud and call it “European sovereign office” and it’s a winner.
AK
@_akhaliq
Fork LibreOffice, call it cursor for Microsoft Office
Raise $1BThis post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
👀
Linuxiac
@linuxiac
EU OS is a new—still in the idea stage—Fedora-based #linux project designed to provide the EU public sector with secure, sovereign, and eco-friendly computing.
linuxiac.com/eu-os-is-a-new-community-led-linux-alternati… -
Kevin Schawinski reposted
Nicholas Roberts
@nick11roberts
📉📉NEW SCALING LAW PHENOMENON 📉📉
We find that knowledge and reasoning exhibit different scaling behaviors!
Super excited to finally tell you all about our paper on the compute optimal scaling of skills:
arxiv.org/pdf/2503.10061
[1/n] -
Kevin Schawinski
@kevinschawinski
Hmmmmmm 🤔
-
Kevin Schawinski
@kevinschawinski
Claude code is insane. 😱
Kevin Schawinski
@kevinschawinski
I have an overnight flight. Let's see what we can vibe code....
-
Kevin Schawinski reposted
Timothy Garton Ash
@fromTGA
Between Putin and Trump, how can Europe defend itself? The answer: Churchillo-Gaullism. My latest History of the Present newsletter. (Subscribe – It's free!)
open.substack.com/pub/timothygartonash/p/vive-le-churchil… -
Kevin Schawinski
@kevinschawinski
Claude code pro usage
-
Kevin Schawinski
@kevinschawinski
1/3 of my For You feed is now just tweets from Musk, RT and various Russian embassies, plus a sprinkling of racist conspiracy anons. We are reaching new lows every day.
-
Kevin Schawinski
@kevinschawinski
Bring them to Europe! Give them permanent positions, grants and residence permits! This is a once in a generation opportunity to bring the best scientists to Europe. 🇪🇺
Act now, @EU_Commission!
www.404media.co/nasa-yale-and-stanford-scientists-conside… -
Kevin Schawinski
@kevinschawinski
I knew it!
Luca Bertuzzi
@BertuzLuca
🚨Cloud sovereignty is back on the menu! The upcoming Cloud and AI Development Act is bound to heavily focus on sovereign cloud solutions, but the approach has shifted to strategic sectors and public procurement, according to a doc seen by @mlexclusive.
www.mlex.com/mlex/articles/2310346/eu-seeks-cloud-soverei… -
Kevin Schawinski
@kevinschawinski
But it’s the EU that over-regulating AI, right?
TechCrunch
@TechCrunch
OpenAI calls DeepSeek ‘state-controlled,’ calls for bans on ‘PRC-produced’ models tcrn.ch/3DGJ1ji
-
Kevin Schawinski
@kevinschawinski
Do it. 🇪🇺
Luca Bertuzzi
@BertuzLuca
NEW: The Commission wants to propose regulatory simplification and a European preference in public procurement to boost the EU’s tech defense capabilities and reduce dependencies in AI and quantum computing, according to a draft paper seen by @mlexclusive.
www.mlex.com/mlex/articles/2310083/eu-eyes-boosting-europ… -
Kevin Schawinski reposted
Anna Gát
@TheAnnaGat
It’s 2025 in America.
Centrism is now a startup.
It’s called @interintellect_. -
Kevin Schawinski reposted
Olivier Blanchard
@ojblanchard1
As we are in danger of moving from a rules-based system to the law of the jungle, the relative size of players matters much more. Think US-Ukraine. Think US-Canada. Even if you did not like the European Union before, you may want to think again.
-
Kevin Schawinski
@kevinschawinski
Good morning. ☀️
Today would be a really good day to seize Russia’s frozen €200bn and give them to Ukraine to buy weapons.
Ukraine is defending all of Europe. 🇪🇺🇺🇦 -
Kevin Schawinski
@kevinschawinski
I wish I had bought even more of:
🇫🇷Dassault
🇩🇪Hensoldt
🇮🇹Leonardo
🇮🇹Fincantieri
🇪🇺Airbus -
Kevin Schawinski
@kevinschawinski
Last year, I was a speaker at an innovation event organized by the Swiss army. One of the other speakers was a very senior sales person from a US cloud provider. We should take very seriously what he said.
The title of his talk was "The cloud is a weapon. Act accordingly."
His entire pitch was that cloud computing is a weapon of war, and you're either under total US government control, and pay for the privilege, or you don't get access to technology.
Given world events, I think this threat should be taken very seriously by European governments. -
Kevin Schawinski
@kevinschawinski
The EU should create a €100 bn VC fund to bring strategic companies into existence. AI, cloud, defense, fusion, biology. China is doing it, so we should as well.
Kevin Schawinski
@kevinschawinski
You can just do things.
-
Kevin Schawinski
@kevinschawinski
Yes! 🇬🇧🇺🇦
Rishi Sunak
@RishiSunak
It’s time to make Putin pay.
Here’s how 👇🧵 -
Kevin Schawinski
@kevinschawinski
Good morning! ☀️
Today would be a good day to seize Russia’s €200bn and give them to Ukraine to buy European weapons. 🇪🇺🇺🇦 -
Kevin Schawinski reposted
Olivier Blanchard
@ojblanchard1
Reasons to run and finance the EU defense effort at the EU level.
Why run it at the EU level? Because it implies better coordination and less duplication of effort. Choose the best tanks, the best planes, independent of nationality. Admittedly hard to do politically at the EU level, but impossible to do if decisions are left to each country.
Debt or taxes? Because it implies an initial adjustment to get to the right defense level, it makes sense to smooth the adjustment and finance most of it initially by debt rather than higher taxes.
Why EU rather than national debt? EU defense debt, if issued along the same lines as NGEU debt, naturally extinguishes itself when it comes to maturity. Not the case for the rest of public debt, which is typically rolled over. -
Kevin Schawinski
@kevinschawinski
This is going to be an unapologetically pro Europe account.
Kremlin talking points = automatic block. -
Kevin Schawinski reposted
Sander Tordoir
@SanderTordoir
I wrote a new piece in foreign policy.
The US's postwar role as guarantor of Europe’s security is over – and may even turn adversarial.
But Europe has an overlooked trump card: when it comes to manufacturing Europe blows the US out of the water.
1/x
foreignpolicy.com/2025/03/07/europe-heavy-industry-trump-… -
Kevin Schawinski reposted
Flo Crivello
@Altimor
The impending death of SaaS caused by AI reminds me of how all consumer manufacturing died out because of 3D printers
-
Kevin Schawinski
@kevinschawinski
🫡
Emmanuel Macron
@EmmanuelMacron
Europeans,
-
Kevin Schawinski
@kevinschawinski
Today would be the perfect day to seize Russia’s €200bn and and give them to Ukraine to buy European weapons. 🇪🇺🇺🇦
-
Kevin Schawinski
@kevinschawinski
Macron speaking. 🇫🇷🇪🇺🇺🇦
-
Kevin Schawinski
@kevinschawinski
In 15 years, Europe will be richer, more powerful and more advanced than the US or China.
-
Kevin Schawinski
@kevinschawinski
Today would be a good day to seize Russia’s frozen €200bn and give them to Ukraine to buy European weapons. 🇺🇦🇪🇺
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
“Whatever it takes” 😍🇩🇪🇪🇺
Friedrich Merz
@_FriedrichMerz
„Whatever it takes.“ (tm)
-
Kevin Schawinski reposted
Andreas Klinger 🦾
@andreasklinger
I am starting a new video series:
"Europe's Most Ambitious Startup 🇪🇺"
Who should I interview?
Who is building crazy ambitious and exciting stuff and got the credibility to actually do it?
The benchmark is: [if it works] fundamentally changes their industry.
I will do this as unfiltered and authentic as possible. Together we will visit their office, their workshops and labs and I just take you along.
We will put a camera up, chat with the founders, talk about their working prototypes, about their customers, about their factories.
Understand why they do it and what they are epic.
Please help me with getting this right – hit reply:
Which company is absotuely epic? Who are the founders? Where are they based?
And btw subscribe (link in reply) - the first one is dropping tomorrow! 🔥
For Europe ❤️🇪🇺This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
The EU should use part of the €800bn to bring over all the generals, scientists and engineers that the US is firing. Make them work for us. 🇪🇺
-
Kevin Schawinski
@kevinschawinski
Are these EUR 800bn the same as Draghi's 800bn? Or is the EU doing EUR 1.6tn of investment? 👀
Ursula von der Leyen
@vonderleyen
We are living in dangerous times.
Europe‘s security is threatened in a very real way.
Today I present ReArm Europe.
A plan for a safer and more resilient Europe ↓ x.com/i/broadcasts/1YpKkBrgRpAxj -
Kevin Schawinski
@kevinschawinski
Unpopular opinion: Swiss neutrality was only possible in the last decades because of US security guarantees for Europe. Now that these are gone, neutrality is no longer viable.
-
Kevin Schawinski
@kevinschawinski
“Faisons payer les Russes” - exactly!
Gabriel Attal
@GabrielAttal
En Europe, il y a près de 300 milliards d'euros d'avoirs russes qui sont aujourd'hui gelés.
Avant de faire financer l'effort aux Français et aux Européens, faisons payer les Russes. -
Kevin Schawinski
@kevinschawinski
Good morning. Today would be a good day to seize Russia’s frozen EUR 200bn and give them to Ukraine to buy European weapons. 🇺🇦🇪🇺
-
Kevin Schawinski
@kevinschawinski
Some problems require Finnish solutions. 🇫🇮
-
Kevin Schawinski
@kevinschawinski
Europeans! You can just do things! 🇪🇺
- Confiscate Russia’s money
- Build drone armies
- Rebuild Ukraine
- Design European fighter jets
You don’t need to ask Washington or Moscow for permission. -
Kevin Schawinski
@kevinschawinski
I wish European leaders would just accept that the real US president has already made the decision to end NATO. Act accordingly.
-
Kevin Schawinski
@kevinschawinski
Last year, I was a speaker a major event on technology and innovation hosted by the Swiss army. One of the other speakers was a high level executive from a US cloud company.
The title of his talk - and pitch - was that "The cloud is a weapon: act like it!". Of course, he was selling his cloud services, but he also made it crystal clear that this was the view of the US government.
The cloud is a weapon of the US government. And you don't get to have your own.
So I asked o3-mini-high to research how China managed to assure a sovereign cloud infrastructure. Maybe worth looking at...
Kevin Schawinski
@kevinschawinski
Here goes:
Kevin, here’s a comprehensive explanation:
Regulatory and Legal Barriers
China systematically crafted a legal environment that makes it extremely difficult for foreign cloud providers to operate independently. Key measures include:
• Data Localization Requirements: Laws such as the Cybersecurity Law, Data Security Law, and Personal Information Protection Law mandate that all data generated in China must be stored within its borders. For US cloud providers—whose architectures were built for global, often centralized data storage—this means re-engineering operations or setting up entirely new, locally based infrastructures.
• Strict Licensing and Ownership Rules: Historically, the Chinese government imposed ownership restrictions on value-added telecom services (VATS), including cloud services. US companies are forced to form joint ventures with Chinese partners (for example, AWS works with Sinnet, and Microsoft Azure partners with 21Vianet) because foreign entities are either barred or capped (usually at 50% ownership) from obtaining the necessary operating licenses independently.
Promotion of Domestic Champions
Beyond regulatory hurdles, China has actively supported domestic cloud providers:
• Preferential Treatment and Subsidies: By easing regulations and providing incentives exclusively to local firms, Chinese companies like Alibaba Cloud, Tencent Cloud, and Huawei Cloud received a competitive edge. These measures were part of a broader strategy to boost national technological self-reliance.
• Forced Technology Transfer: In many cases, US firms were required to share technology with their local partners as a condition of market entry. This not only diluted their proprietary advantages but also helped domestic firms climb up the value chain.
Market Impact and Strategic Outcomes
The cumulative effect of these policies was to:
• Eliminate Direct Competition: By forcing US providers to operate only through local joint ventures, China effectively removed them from direct control of their domestic cloud operations. Over time, this weakened their market share and influence.
• Create a Sovereign Ecosystem: Chinese authorities have intentionally built a “closed-loop” market where data, services, and innovation occur domestically. This policy framework has accelerated the growth of local providers while sidelining foreign competitors.
Summary
In essence, China pushed US cloud providers out of its domestic market by erecting formidable regulatory barriers—demanding local data storage, enforcing strict licensing and ownership rules, and compelling foreign companies to partner with domestic firms. These measures, coupled with policies designed to foster national champions, have significantly reduced the ability of US cloud providers to operate independently in China.
Does this cover the aspects you were most interested in, or would you like more details on any specific mechanism or its impact on a particular US cloud provider? -
Kevin Schawinski
@kevinschawinski
Donate to Ukraine here. They are leading the defense of Europe. 🇺🇦
u24.gov.ua -
Kevin Schawinski
@kevinschawinski
The views on my tweets have terminally collapsed since my views are displeasing to the platform owner. Ignore the check mark, I cancelled a while ago. Should disappear soon.
-
Kevin Schawinski
@kevinschawinski
At least one Swiss politician speaks up - sadly he is no longer in the government.
Alain Berset
@alain_berset
The Council of Europe stands with the Ukrainian People. Building peace, democratic security, stability and prosperity in #europe. That’s what we do. Ukrainians can count on #strasbourg @ZelenskyyUa @coe
-
Kevin Schawinski
@kevinschawinski
Today would be a good day to seize Russia’s EUR 200bn and give them to Ukraine to buy European weapons. 🇪🇺🇺🇦
-
Kevin Schawinski
@kevinschawinski
Can you make 1M of them? Asking for a continent.
Torsten Reil
@torstenreil
100,000 HX-2 strike drones would deter a land invasion of Europe once and for all.
@HelsingAI stands ready to produce them. -
Kevin Schawinski reposted
ian bremmer
@ianbremmer
unified and strong european twitter statements of support for ukraine over last day. they’re worth the paper they aren’t printed on.
the europeans have to show they are prepared to defend ukraine without the americans. the future of europe depends on it. -
Kevin Schawinski
@kevinschawinski
“Putin wants peace” 🤮
Reuters World
@ReutersWorld
Russian drone strike hits medical facility, other targets in Kharkiv, officials say www.reuters.com/world/europe/russian-drone-strike-hits-me…
-
Kevin Schawinski
@kevinschawinski
European leader rn
-
Kevin Schawinski
@kevinschawinski
The leader of the free world.
-
Kevin Schawinski
@kevinschawinski
Right now would be a good time to seize Russia’s EUR 200bn and give them to Ukraine to buy European weapons. 🇪🇺🇺🇦
Kevin Schawinski
@kevinschawinski
Russia delenda est.
-
Kevin Schawinski
@kevinschawinski
Whenever I accidentally see the “For you” feed, it’s just Kremlin shills. 🤮
-
Kevin Schawinski
@kevinschawinski
Today would be a good day to seize Russias EUR 200bn and give them to Ukraine to buy European weapons. 🇪🇺🇺🇦
-
Kevin Schawinski
@kevinschawinski
If you want to have a glimpse of the future where ~all social media interactions are bots, you don’t need to go far. Twitter right here is mostly AI propaganda bots.
Enjoy! -
Kevin Schawinski
@kevinschawinski
Unreal.
Adam Lerman
@AdamLerman5
The YC deleted video for sweatshop startup Optifye
Kevin Schawinski
@kevinschawinski
The Adeptus Mechanicus innovation fund would like to invest. Archmagus Cawl will join the board.
-
Kevin Schawinski
@kevinschawinski
How to CEO:
-
Kevin Schawinski
@kevinschawinski
👀
CDU Deutschlands
@CDU
Heute jährt sich der Angriff von Wladimir #putin auf die #ukraine bereits zum dritten Mal. Drei Jahre Krieg in Europa. Drei Jahre Tod und Leid in der Ukraine. Wir stehen fest an ihrer Seite und tun alles, um sie bei ihrem Recht auf Selbstverteidigung und einen gerechten Frieden zu unterstützen. Für Freiheit. Für Frieden. Für die Menschen in der Ukraine. #ukrainemustwin
-
Kevin Schawinski reposted
Roberta Raileanu
@robertarail
Super excited to share 🧠MLGym 🦾 – the first Gym environment for AI Research Agents 🤖🔬
We introduce MLGym and MLGym-Bench, a new framework and benchmark for evaluating and developing LLM agents on AI research tasks.
The key contributions of our work are:
🕹️ Enables the exploration of different training algorithms for AI Research Agents such as RL
🛠️ Provides a flexible evaluation framework that can accommodate different artifacts such as models, algorithms, or predictions
🤖 Allows researchers to evaluate any model without the need to develop a custom agentic harness
🎯 Introduces 13 diverse open-ended AI Research tasks for evaluating AI Research Agents on a wide range of domains such as computer vision, natural language processing, reinforcement learning, game theory, and logical reasoning.
📈 Proposes a new evaluation metric for AI Research Agents
MLGym makes it easy to:
1) Add new tasks
2) Evaluate new models
3) Integrate new agents
Check out a video of the MLGym Agent to see how it performs the full pipeline of idea generation💡, implementation 👩💻, experimentation 👩🔬, and iteration 🔄 to improve on ML tasks.
Huge thanks to the exceptionally talented @deepaknathani11 who led this work and to all the other amazing collaborators who made this possible 🙏🫶🚀 -
Kevin Schawinski reposted
Aleksandr Tiulkanov
@shadbush
U.S. companies seem to prefer the EU copyright law to the one they have at home and even threaten their authorities to leave for Europe if training on copyrighted works turned out not to be fair use on the U.S. soil ;) Who knew that Europe could be so much more business-friendly?
-
Kevin Schawinski
@kevinschawinski
Just make sure to add all Oracle source code so we can use that too….
unusual_whales
@unusual_whales
Oracle's, $ORCL, Larry Ellison said governments should consolidate all national data for consumption by AI models, calling this step the “missing link” for them to take full advantage of the tech, per Bloomberg.
-
Kevin Schawinski reposted
Donald Tusk
@donaldtusk
Enough talking, it’s time to act! 1. Let’s finance our aid for Ukraine from the Russian frozen assets. 2. Let’s strengthen air policing, the Baltic sentry and the EU borders with Russia. 3. Let’s swiftly adopt new fiscal rules to finance the EU security and defence. Now!
-
Kevin Schawinski
@kevinschawinski
Today would be a good day to seize Russia’s frozen reserves in Europe and give them to Ukraine to buy European arms. 🇪🇺🇺🇦
-
Kevin Schawinski
@kevinschawinski
Good morning.
-
Kevin Schawinski
@kevinschawinski
But it’s Europe that’s so hostile to freedom of speech, right?
Dave Keating
@DaveKeating
The US State Department is forcing all of its employees to cancel subscriptions to any Trump-unfriendly media.
-
Kevin Schawinski
@kevinschawinski
Draghi is right. 🇪🇺
Julien Hoez
@JulienHoez
Mario Draghi barrels into European politicians and leaders for their inaction:
"You say no to public debt. You say no to the single market. You say no to create a capital market union. You can't say no to everything...You ask me what's best to do…I don’t know, but do something" -
Kevin Schawinski
@kevinschawinski
Let's go. 🇪🇺
Torsten Reil
@torstenreil
100,000 HX-2 strike drones would deter a land invasion of Europe once and for all.
@HelsingAI stands ready to produce them. -
Kevin Schawinski
@kevinschawinski
🇪🇺🇺🇸Let me try and translate:
• Political actions by the Trump admin (removing/dismissing PCLOB members to below quorum) reflect deep concerns about US privacy oversight going forward.
• The EU is critically evaluating the EU–US Data Privacy Framework.
• If privacy mechanisms are deemed insufficient, the framework could face significant challenges or be rendered defunct.
We could be headed towards the end of data transfers between the EU and the US.
Luis Montezuma | @luismontezuma@mstdn.social
@montezumachavez
-
Kevin Schawinski
@kevinschawinski
People who think that AI regulation in general, and the AI Act in particular, are bad…. what here is so awful? How will it inhibit innovation and progress?
Luis Montezuma | @luismontezuma@mstdn.social
@montezumachavez
This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
Europe, if there ever was a time to unify, it’s now.
Stop talking about “ending the war”. Start talking about WINNING the war.
🇪🇺 -
Kevin Schawinski reposted
Thierry Breton
@ThierryBreton
Welcome to Europe VP @JDVance.
With all due respect, in 🇪🇺, freedom of speech is a core value of our democracy — it is non-negotiable. Never was and never will be.
Proof? Even fake news are allowed (sadly, sometimes echoed by top US officials) :
EU/DSA has never canceled elections—despite far-right extremists and @ElonMusk continuously pushing this lie.
The DSA doesn’t regulate content; it ensures #transparency so no platform — being Chinese, Singaporean, American or European —can manipulate algorithms.
In the United States, foreign companies respect US laws. It's not open for debate.
In Europe, we expect the same behavior from US companies.
"When in Rome, do as the Romans do."
www.youtube.com/watch?v=_AffOWudVHAThis media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
The US spent so much effort to de-nazify Germany, why is it working to re-nazify it now? 🤔
The Washington Post
@washingtonpost
Breaking news: Vice President JD Vance on Friday met with far-right German leader Alice Weidel, the highest-ranking U.S. official ever to do so. www.washingtonpost.com/politics/2025/02/14/vance-munich-e…
-
Kevin Schawinski
@kevinschawinski
👀
Anna Cavazzini
@anna_cavazzini
.@die_gruenen bringen die Deutschland-App und machen Verwaltung endlich digital! Kein Fax, kein Warten, (fast) kein Stress. Jetzt nur noch EU-weit
-
Kevin Schawinski
@kevinschawinski
👀
European Commission
@EU_Commission
How do Europeans feel about Artificial Intelligence in the workplace?
While most Europeans see AI as beneficial to the economy and quality of life, our latest #eurobarometer reveals support for careful management to ensure its fair and ethical use ↓ -
Kevin Schawinski
@kevinschawinski
I am so grateful for the iPhone feature which lets you silence all unknown numbers.
Y Combinator
@ycombinator
🐧@Apten_AI is an AI SMS + voice agent that automates sales and marketing for B2C companies. It communicates in the same way a human can, switching between SMS or voice at any time.
www.ycombinator.com/launches/Mmf-apten-ai-sms-voice-agent…This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
🇨🇭 🚨 BREAKING: Switzerland is considering how to regulate AI!
The BAKOM (Swiss Federal Office of Communications) has prepared a set of recommendation for the Swiss Federal Council published today to outline potential approaches for regulating artificial intelligence in Switzerland.
It proposes three possible regulatory paths:
1️⃣ continuing with current sector-specific regulations,
2️⃣ ratifying the Council of Europe's AI Convention with either minimal or more comprehensive implementation, or
3️⃣ adopting the AI Convention plus aligning with the EU's AI Act.
The document aims to balance three main goals: strengthening Switzerland's position as an innovation hub, protecting fundamental rights, and building public trust in AI. It specifically considers how these approaches would affect Swiss companies' access to the EU market while maintaining Switzerland's regulatory independence.
The details of this proposal and the options it presents are particularly interesting given the current tension around AI regulation globally. How can Switzerland, a small but very innovative country, thrive in this environment? -
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
The EU is preparing a coordinated response to the US export control measures on AI chips. The Commission is gathering feedback from member states and asked to be informed about bilateral talks with the US administration. w/ @oscarpandiello
content.mlex.com -
Kevin Schawinski
@kevinschawinski
🤣
Sam Altman
@sama
no thank you but we will buy twitter for $9.74 billion if you want
-
Kevin Schawinski reposted
Chubby♨️
@kimmonismus
60+ EU companies announce the EU AI Champions Initiative.
Is Europe finally waking up? -
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
EXCLUSIVE: 🇫🇷, 🇩🇪 & 🇳🇱 are set to publish a paper later this week asking not only to cover AI and cloud services under the #dma but also to update the DMA's list of obligations & introduce a supervisory fee for gatekeepers. All details on @mlexclusive.
www.mlex.com/mlex/articles/2295724/tech-companies-should-… -
Kevin Schawinski
@kevinschawinski
We’re working hard to make AI governance easy and empower companies to comply with new regulations. By seamlessly combining advanced AI with robust Governance, Risk, and Compliance (GRC) practices, our platform automates the compliance process and speeds up risk management so that you can keep building.
Our “Compliance by Design” approach leverages specialized AI agents that review your code and documentation to identify gaps and deliver actionable recommendations. This method accelerates your journey toward meeting critical standards such as ISO/IEC 42001 and the EU AI Act, empowering your teams to focus on innovation while our system streamlines your compliance process.
I’m excited to see how we can make AI compliance less of a burden and more of an advantage. Those who adopt responsible AI practices and demonstrate them to customers not only win their trust, but also their business.
If you’re a startup or SME looking to get ahead of the game, sign up for free on the Modulos AI Governance Platform.
Link in the comments.
#aigovernance #compliancebydesign #grc #innovation
Kevin Schawinski
@kevinschawinski
Start here: www.modulos.ai/pricing/
-
Kevin Schawinski
@kevinschawinski
Capping indirect costs for research projects at 15% is actually good.
NIH
@NIH
Last year, $9B of the $35B that the National Institutes of Health (NIH) granted for research was used for administrative overhead, what is known as “indirect costs.” Today, NIH lowered the maximum indirect cost rate research institutions can charge the government to 15%, above what many major foundations allow and much lower than the 60%+ that some institutions charge the government today. This change will save more than $4B a year effective immediately.
-
Kevin Schawinski
@kevinschawinski
👀 ☢️
Keir Starmer
@Keir_Starmer
Britain is back in the race for nuclear energy.
-
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
ISO 42001 is becoming the essential AI certification to convince enterprise customers that your AI product and the team behind it are reliable. Leading companies like Microsoft are already requiring their vendors to have ISO 42001.
Setting up an AI Management System and getting it certified can look daunting, especially if you're a startup that's already struggling with various requirements from enterprise buyers. Right now, only a handful of companies are 42001 certified, but the list is growing rapidly.
Modulos has already helped a German legaltech startup become the first 42001-certified company in their country. While their exceptional preparation was key to their quick certification, our platform and guidance significantly accelerated their journey.
That's why I am offering to personally coach five exceptional founders of AI startups from zero to ISO 42001 if you sign up for an annual Modulos Team plan. Two spots are already claimed.
🚀 Ready to get ahead of the curve? DM me to get started. -
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
Defining “AI System” Under the EU AI Act
The European Commission’s has just released its Guidelines to clarify how to determine whether a system qualifies as “AI” under the Act.
These Guidelines (non-binding but influential) walk providers, policymakers, and users through key elements of the AI definition. That clarity helps in figuring out which obligations may come into play—especially for autonomous, adaptable, or inference-based systems that can shape our physical or digital environments.
The EU AI Act defines an AI system as:
“a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.”
From this definition, the guidelines extract seven main elements:
1.Machine-based system
2.Designed to operate with varying levels of autonomy
3.May exhibit adaptiveness after deployment
4.Has explicit or implicit objectives
5.Infers, from the input it receives, how to generate outputs
6.Outputs may be predictions, content, recommendations, or decisions
7.Those outputs can influence physical or virtual environments
Let's break down what these mean according to the new guidance. In the table below, you'll find:
•Key Categories that the Commission considers in classifying AI systems
•Core Explanations behind each category
•Real-World Examples to illustrate practical scenarios
#aiact #airegulation #governance #responsibleai -
Kevin Schawinski
@kevinschawinski
ISO 42001 is becoming the essential AI certification to convince enterprise customers that your AI product and the team behind it are reliable. Leading companies like Microsoft are already requiring their vendors to have ISO 42001.
Setting up an AI Management System and getting it certified can look daunting, especially if you're a startup that's already struggling with various requirements from enterprise buyers. Right now, only a handful of companies are 42001 certified, but the list is growing rapidly.
Modulos has already helped a German legaltech startup become the first 42001-certified company in their country. While their exceptional preparation was key to their quick certification, our platform and guidance significantly accelerated their journey.
That's why I am offering to personally coach five exceptional founders of AI startups from zero to ISO 42001 if you sign up for an annual Modulos Team plan. Two spots are already claimed.
🚀 Ready to get ahead of the curve? DM me to get started. -
Kevin Schawinski
@kevinschawinski
Defining “AI System” Under the EU AI Act
The European Commission’s has just released its Guidelines to clarify how to determine whether a system qualifies as “AI” under the Act.
These Guidelines (non-binding but influential) walk providers, policymakers, and users through key elements of the AI definition. That clarity helps in figuring out which obligations may come into play—especially for autonomous, adaptable, or inference-based systems that can shape our physical or digital environments.
The EU AI Act defines an AI system as:
“a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.”
From this definition, the guidelines extract seven main elements:
1.Machine-based system
2.Designed to operate with varying levels of autonomy
3.May exhibit adaptiveness after deployment
4.Has explicit or implicit objectives
5.Infers, from the input it receives, how to generate outputs
6.Outputs may be predictions, content, recommendations, or decisions
7.Those outputs can influence physical or virtual environments
Let's break down what these mean according to the new guidance. In the table below, you'll find:
•Key Categories that the Commission considers in classifying AI systems
•Core Explanations behind each category
•Real-World Examples to illustrate practical scenarios
#aiact #airegulation #governance #responsibleai
Kevin Schawinski
@kevinschawinski
Link: digital-strategy.ec.europa.eu/en/library/commission-publi…
-
Kevin Schawinski
@kevinschawinski
AI enshittification has already started.
-
Kevin Schawinski
@kevinschawinski
Academia is so cooked.
Ethan Mollick
@emollick
Economist @joshgans uses o1-pro to generate a (minor, fun) paper in an hour based on an idea of his, and it gets published in an appropriate peer reviewed journal, with adequate disclosure.
He ends with the same sentiment I am increasingly seeing from fellow academics: what now? -
Kevin Schawinski
@kevinschawinski
We've put up our initial analysis of the new European Commission guidelines on prohibited AI systems. If your product could fall into any of these categories, then this blog is definitely worth your time:
• Article 5(1)(a) & (b): bans certain manipulative, deceptive, or exploitative AI uses that “materially distort” someone’s behavior and lead to significant harm.
• Article 5(1)(c): bans social scoring that leads to unjust or disproportionate detrimental treatment.
• Article 5(1)(d): bans criminal offence risk assessments or predictions based solely on a person’s profiling/personality traits.
• Article 5(1)(e): bans untargeted scraping of facial images from internet/CCTV to build face recognition databases.
• Article 5(1)(f): bans emotion recognition in workplace/education (except for safety/medical reasons).
• Article 5(1)(g): bans biometric categorization that infers someone’s race, religion, political opinions, sexual orientation, etc.
• Article 5(1)(h): bans real-time remote biometric identification (RBI) in publicly accessible spaces for law enforcement, with narrow exceptions (kidnapping victims, imminent threats, serious crimes).
hashtag#artificialintelligence hashtag#airegulation hashtag#repsonsibleai hashtag#euaiact
www.modulos.ai/blog/the-eu-ai-act-prohibited-practices/ -
Kevin Schawinski
@kevinschawinski
Strategery would have been a better fit.
Michael Saylor
@saylor
MicroStrategy is now Strategy. www.strategysoftware.com/press/microstrategy-is-now-strat…
-
Kevin Schawinski
@kevinschawinski
🚨🇪🇺BREAKING: Dutch Foundation for Market Information Research has sued X and Tiktok in a first of its kind class action lawsuit over various European laws including the EU AI Act. Damages sought are in the billions of Euros.
www.spiritlegal.com/en/news/details/press-release-class-a… -
Kevin Schawinski reposted
ioana ciucă
@errai34
Possibly our most ambitious and open-ended project yet, TeleskAI is definitely the one where we've let our imagination take the lead telesk.ai. Huge thanks for the one and only @rishistyping for the website.
TeleskAI (the far off sky) is a new idea under @universe_tbd where we're envisioning agents autonomously controlling a telescope. As we are actively looking for funding to come make it come to life, definitely reach out if you want to support our uniquely bold idea. -
Kevin Schawinski
@kevinschawinski
👀
European Commission
@EU_Commission
AI made in 🇪🇺
OpenEuroLLM, the first family of open source Large Language Models covering all EU languages, has earned the first STEP Seal for its excellence.
It brings together EU startups, research labs and supercomputing hosts to train AI on European supercomputers ↓ -
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
Europe should to the following:
1️⃣ Budget a few billion EUR per year
2️⃣ Get a super nice building by a lake or something, and cater it with everything engineers need (coffee, food etc.)
3️⃣ Hire like 100 totally cracked CS and math PhDs and engineers and pay them outrageous FAANG salaries
4️⃣ Keep buying compute for them
5️⃣ Tell them that their only mission is to ship open source frontier models with European values as a public utility for Europe -
Kevin Schawinski reposted
UniverseTBD
@universe_tbd
The sky is no longer the limit 🚀: telesk.ai
-
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
For those wondering, the AI Act's guidelines on prohibited practices and definition of AI systems are still going through the European Commission's internal approval process. It is still unclear whether they will be published today or whether they will be published together.
-
Kevin Schawinski
@kevinschawinski
Europe should to the following:
1️⃣ Budget a few billion EUR per year
2️⃣ Get a super nice building by a lake or something, and cater it with everything engineers need (coffee, food etc.)
3️⃣ Hire like 100 totally cracked CS and math PhDs and engineers and pay them outrageous FAANG salaries
4️⃣ Keep buying compute for them
5️⃣ Tell them that their only mission is to ship open source frontier models with European values as a public utility for Europe -
Kevin Schawinski
@kevinschawinski
I've been coding with o3-mini-high all weekend and at some point, it just sorta died and went on an existential rant...
Post not available here, view on X
Kevin Schawinski
@kevinschawinski
Anyway. Back to Claude.
@TheAnnaGat @amasad
Kevin Schawinski
@kevinschawinski
I shouldn't have doubted Claude
@alexalbert__ -
Kevin Schawinski
@kevinschawinski
It's really good.
Sam Altman
@sama
o3-mini-high is really good and it's cool to hear some people say it's their favorite model ever
but
it reminds me of when we used to have something internally called "the little big run"
a top 2025 goal is to fix our naming problem
Kevin Schawinski
@kevinschawinski
It can also be rather dumb. Might switch back to Claude soon...
-
Kevin Schawinski
@kevinschawinski
Stop using prohibited AI systems in the EU right now.
Aleksandr Tiulkanov
@shadbush
What is the economic incentive for AI operators to cease prohibited AI practices in Europe already now, while administrative fines only kick in from 2 August 2025?
TL;DR: threat of civil litigation. Understanding this requires considering the broader body of law beyond the EU AI Act itself.
EU regulations (like the AI Act) have direct effect in member states and are part of their national legal orders. Therefore, starting from 2 February 2025:
📌 AI practices that are prohibited from this date under the EU AI Act article 5 (social scoring, manipulative AI "girlfriends", untargeted scraping of facial images, etc.) are a statutorily defined type of unlawful conduct;
⚖️ Conducting an EU AI Act prohibited AI practice may constitute a delict and may give rise to delict liability under national law;
💼 Plaintiffs will likely be able to claim compensation for damages caused by the delict, assuming they prove the damage caused, the unlawful conduct occurred, and the causal link between the two (general principle of fault liability: art 1240 Code Civil, section 823 BGB, etc.).
Example: on 2 February 2025 you fail to cease offering EU customers a manipulative chatbot that is designed to hook users up by luring them with promises of a "deep connection" with a realistic AI avatar, 24/7 availability, sexting and AI-generated pornography on-demand, so that they pay you for a subscription and possibly further in-app purchases. All of this despite the prohibitions in art 5 AI Act coming into effect on 2 February 2025.
A subset of your user base is especially susceptible. Actual relationships get sidelined and destroyed. Some users develop medically diagnosed conditions, such as obsessive-compulsive disorder, major depression, etc.
Harm to mental health is caused. Some of the users eventually have to undergo therapy.
Having tapered off your malicious creation, they decide to sue you. Experts get called in, evidence gets filed. Unlawful conduct on your side is apparent. Damage to mental health and monetary damages are demonstrated. Causal link is proven. You settle or just lose.
Congratulations, you have been duly punished for not adapting your product to the EU market! -
Kevin Schawinski
@kevinschawinski
The US could criminalize running DeepSeek-R1.
Kevin Schawinski
@kevinschawinski
🚨 🇺🇸 BREAKING Major US AI Regulation Alert: Sen. Hawley's New "AI Decoupling Act"
At a pivotal moment in global AI development - with the EU AI Act's prohibitions now in effect and breakthroughs like DeepSeek-R1 emerging from China - Senator Hawley has introduced sweeping legislation that could fundamentally reshape US-China AI relations.
The proposed "Decoupling America's Artificial Intelligence Capabilities from China Act" would criminalize importing Chinese-developed AI technology - meaning yes, importing models like DeepSeek-R1 would become a criminal offense carrying penalties up to $1M for individuals and $100M for corporations. Should the Bill become law, organizations would have 180 days after enactment to comply.
Beyond imports, the bill targets the entire spectrum of US-China AI collaboration. It would prohibit US persons from conducting AI R&D within China, block technical partnerships, and ban investments in Chinese AI companies. US investors and companies would have one year to divest from Chinese AI entities.
The implications for corporate governance are profound. Any organization working with AI would need to thoroughly audit their supply chains, research collaborations, and investment portfolios. The bill's broad definition of "Chinese entity of concern" could affect complex corporate structures, subsidiaries, and even seemingly arms-length relationships.
Particularly striking is the timing - as the EU establishes itself as the first major regulator of AI, and as Chinese models demonstrate increasing capabilities, this legislation would effectively force a technological decoupling in one of the most critical emerging technologies.
For AI governance professionals, this represents a potential sea change. While the bill's passage isn't certain, its introduction signals growing legislative appetite for stricter controls on international AI development.
NOTE: This Bill has been INTRODUCED. It's nowhere near PASSAGE. Nevertheless, it's a clear signal of priorities and concerns by the US government.
Link to the Bill is below. -
Kevin Schawinski
@kevinschawinski
🚨 🇺🇸 BREAKING Major US AI Regulation Alert: Sen. Hawley's New "AI Decoupling Act"
At a pivotal moment in global AI development - with the EU AI Act's prohibitions now in effect and breakthroughs like DeepSeek-R1 emerging from China - Senator Hawley has introduced sweeping legislation that could fundamentally reshape US-China AI relations.
The proposed "Decoupling America's Artificial Intelligence Capabilities from China Act" would criminalize importing Chinese-developed AI technology - meaning yes, importing models like DeepSeek-R1 would become a criminal offense carrying penalties up to $1M for individuals and $100M for corporations. Should the Bill become law, organizations would have 180 days after enactment to comply.
Beyond imports, the bill targets the entire spectrum of US-China AI collaboration. It would prohibit US persons from conducting AI R&D within China, block technical partnerships, and ban investments in Chinese AI companies. US investors and companies would have one year to divest from Chinese AI entities.
The implications for corporate governance are profound. Any organization working with AI would need to thoroughly audit their supply chains, research collaborations, and investment portfolios. The bill's broad definition of "Chinese entity of concern" could affect complex corporate structures, subsidiaries, and even seemingly arms-length relationships.
Particularly striking is the timing - as the EU establishes itself as the first major regulator of AI, and as Chinese models demonstrate increasing capabilities, this legislation would effectively force a technological decoupling in one of the most critical emerging technologies.
For AI governance professionals, this represents a potential sea change. While the bill's passage isn't certain, its introduction signals growing legislative appetite for stricter controls on international AI development.
NOTE: This Bill has been INTRODUCED. It's nowhere near PASSAGE. Nevertheless, it's a clear signal of priorities and concerns by the US government.
Link to the Bill is below.
Kevin Schawinski
@kevinschawinski
Source: www.hawley.senate.gov/hawley-introduces-legislation-to-de…
-
Kevin Schawinski
@kevinschawinski
I hope Switzerland just sits out the trade wars. 🇨🇭
Our best asset is that most people forget that we are not in the EU. -
Kevin Schawinski
@kevinschawinski
The reassuring feeling when o3-mini-high reasons for a long time.
-
Kevin Schawinski
@kevinschawinski
It's 3am but I have jet lag. Perfect time to try this. *cracks knuckles*
-
Kevin Schawinski
@kevinschawinski
🇮🇹 🇪🇺 🇨🇳 In a bold move, Italy's Data Protection Authority, the Garante, has blocked access to DeepSeek in Italy, citing "completely insufficient" responses about their data practices.
Apparently, claiming "GDPR doesn't apply to us" isn't the magic phrase some thought it was. A timely reminder that in AI governance, you can't ignore the rules just because you're in a different country.
Kevin Schawinski
@kevinschawinski
Source: garanteprivacy.it/home/docweb/-/docweb-display/docweb/100…
Kevin Schawinski
@kevinschawinski
Don’t want your AI company blocked in Europe? Get ready for the EU AI Act by setting up your AI Management System. Start free here: www.modulos.ai/pricing/
-
Kevin Schawinski
@kevinschawinski
If I was doing science today, I'd skip peer review entirely and just have Claude review the manuscript. He's more coherent that "reviewer 2" and wouldn't ask me to cite Claude et al., 1996abc, 1997, 1998ab, 2001a, 2012ab, 2014.
-
Kevin Schawinski
@kevinschawinski
👀
Mistral AI
@MistralAI
magnet:?xt=urn:btih:11f2d1ca613ccf5a5c60104db9f3babdfa2e6003&dn=Mistral-Small-3-Instruct&tr=udp%3A%2F%2ftracker.opentrackr.org%3A1337%2Fannounce&tr=http%3A%2F%2fopen.tracker.cl%3A1337%2Fannounce
-
Kevin Schawinski reposted
Reuters
@Reuters
Microsoft and OpenAI are investigating whether individuals linked to the Chinese AI startup DeepSeek obtained data from OpenAI without authorization. Microsoft security researchers observed individuals exfiltrating data through OpenAI's API months ago www.reuters.com/technology/microsoft-probing-if-deepseek-…
This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
Deepseek just dropped a multimodal model. MIT license of course. 🔥
huggingface.co/deepseek-ai/Janus-Pro-1B
Kevin Schawinski
@kevinschawinski
Apologies, here's the 7B version:
huggingface.co/deepseek-ai/Janus-Pro-7B -
Kevin Schawinski
@kevinschawinski
I’m taking off on a 10 hour flight. Keep shipping AI 🚢
-
Kevin Schawinski
@kevinschawinski
👀
Binyuan Hui
@huybery
There are some surprises tonight
-
Kevin Schawinski
@kevinschawinski
The acceleration is accelerating.
(There’s a physics term for it) -
Kevin Schawinski
@kevinschawinski
🫡
Zizheng Pan
@zizhpan
This moment is absolutely phenomenal to me.
-
Kevin Schawinski
@kevinschawinski
This.
Satya Nadella
@satyanadella
Jevons paradox strikes again! As AI gets more efficient and accessible, we will see its use skyrocket, turning it into a commodity we just can't get enough of. en.m.wikipedia.org/wiki/Jevons_paradox
-
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
We’re hiring a senior devops/SRE (100%, fully remote). We are deploying the most advanced AI governance platform to customers around the world.
DMs are open! 🙏
www.modulos.ai/careers/ -
Kevin Schawinski
@kevinschawinski
Very important replication work by the huggingface team.
clem 🤗
@ClementDelangue
Our science team has started working on fully reproducing and open-sourcing R1 including training data, training scripts,...
Full power of open source AI so that everyone all over the world can take advantage of AI progress! Will help debunk some myths I’m sure too.
Thanks @deepseek_ai! -
Kevin Schawinski
@kevinschawinski
We’re hiring a senior devops/SRE (100%, fully remote). We are deploying the most advanced AI governance platform to customers around the world.
DMs are open! 🙏
www.modulos.ai/careers/ -
Kevin Schawinski
@kevinschawinski
Noteworthy: unlike some models from Meta, DeepSeek models contain no restrictions for use in the EU....
DeepSeek
@deepseek_ai
🚀 DeepSeek-R1 is here!
⚡ Performance on par with OpenAI-o1
📖 Fully open-source model & technical report
🏆 MIT licensed: Distill & commercialize freely!
🌐 Website & API are live now! Try DeepThink at chat.deepseek.com today!
🐋 1/n -
Kevin Schawinski
@kevinschawinski
Either everything is on fire, or @SlackHQ did a horrible.
Kevin Schawinski
@kevinschawinski
Update: it’s Slack.
-
Kevin Schawinski
@kevinschawinski
When AGI is your side project....
-
Kevin Schawinski
@kevinschawinski
MIT license 👀
DeepSeek
@deepseek_ai
🚀 DeepSeek-R1 is here!
⚡ Performance on par with OpenAI-o1
📖 Fully open-source model & technical report
🏆 MIT licensed: Distill & commercialize freely!
🌐 Website & API are live now! Try DeepThink at chat.deepseek.com today!
🐋 1/n -
Kevin Schawinski
@kevinschawinski
Are AI Labs on the Cusp of Self-Improvement Breakthroughs?
There’s palpable excitement emanating from top AI labs right now—rumors suggest they may be closing in on (or have already achieved) iterative self-improvement. Think of it as AlphaGo’s self-play strategy taken to the extreme: instead of relying on human data, these models train by competing against earlier versions of themselves, accelerating progress beyond what external datasets can provide.
Adding to the buzz, Sam Altman of OpenAI is scheduled for a closed-door briefing with the U.S. government on the 30th. While details are sparse, it fuels speculation that major breakthroughs—if not full-on transitions—are imminent.
Of course, we outsiders are left to piece together rumors and speculation, and it’s entirely possible this turns out to be strategic PR. Still, if iterative self-improvement is truly here, it changes the AI landscape drastically and fast.
For those of us in AI governance, this would be a watershed moment. A self-improving AI ecosystem would spark immediate calls for heightened regulation, containment measures, and strategic oversight—particularly in the U.S., where lawmakers appear keen to keep a firm hand on the wheel.
Kevin Schawinski
@kevinschawinski
Reporting by @axios: www.axios.com/2025/01/19/ai-superagent-openai-meta
-
Kevin Schawinski
@kevinschawinski
Not sure how much is marketing, but: 👀👀👀👀👀
www.axios.com/2025/01/19/ai-superagent-openai-meta -
Kevin Schawinski
@kevinschawinski
I defer to the experts to tell me if this is “real”.
My subjective experience is definitely that Americans are much wealthier.
Yoram Wijngaarde
@yoramdw
📊 EU27 vs US economies, three perspectives:
1) Nominal GDP shows dramatic US lead
2) Fix FX rates = gap largely disappears
3) Adjust for purchasing power (PPP) = economies look similar -
Kevin Schawinski
@kevinschawinski
Marketing? Real? Can’t tell, but on the tail risk that it’s real…. 👀
Riley Coyote
@RileyRalmuto
just got “read in” to some of what’s happening internally at oai and holy mother of God idk how to even express my feelings without just sounding like hype.
like actually.
idk what to say, but I’ll share this:
the innovators are coming. problem is we don’t know how they got here. -
Kevin Schawinski
@kevinschawinski
👀
Sam Altman
@sama
thank you to the external safety researchers who tested o3-mini.
we have now finalized a version and are beginning the release process; planning to ship in ~a couple of weeks.
also, we heard the feedback: will launch api and chatgpt at the same time!
(it's very good.) -
Kevin Schawinski
@kevinschawinski
👀
Henna Virkkunen
@HennaVirkkunen
Today the Commission has taken three additional investigatory steps against X relating to the platform’s recommender system. They are part of the formal proceedings opened on 18.12.2023 under the #dsa .
We are committed to a fair, safe and democratic online environment 🇪🇺 -
Kevin Schawinski
@kevinschawinski
Often true
sam mcallister
@sammcallister
this meeting could've been a claude
-
Kevin Schawinski reposted
Anna Gát
@TheAnnaGat
I have been given o1 Pro.
THIS IS SO GOOD!!!!!!!!!!!!!!!
I will never sleep again. -
Kevin Schawinski
@kevinschawinski
Does it?
-
Kevin Schawinski
@kevinschawinski
👀
Bundesrat • Conseil fédéral • Consiglio federale
@BR_Sprecher
Bundesrätin @Violapamherd hat per 31.03.25 ihren Rücktritt angekündigt.
La conseillère fédérale @Violapamherd a annoncé sa démission pour le 31.03.25.
La consigliera federale @Violapamherd ha annunciato le sue dimissioni dal Governo per il 31.03.25.
🎥youtube.com/watch?v=MRV8xshwaP0&list=PLEnHzNShzOwYLdRQ5Ti… -
Kevin Schawinski reposted
Anna Gát
@TheAnnaGat
Since the launch of the new @interintellect_ platform the number of bookings is 🚀🚀🚀🚀🚀🚀🚀🚀
Insane!
Huge kudos to my team who turn caffeine into code and amazing hosts and events 🤘🏻🇺🇸❤️🚀❤️🔥🙏🏻🎉😘☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️☕️ -
Kevin Schawinski
@kevinschawinski
👀
Henna Virkkunen
@HennaVirkkunen
Together with my colleague @MarosSefcovic we are confident that we can find a way to maintain a secure transatlantic supply chain on AI technology and super computers, for the benefit of our companies and citizens on both side of the Atlantic. ⤵️
ec.europa.eu/commission/presscorner/detail/en/statement_2… -
Kevin Schawinski
@kevinschawinski
Interesting how many have criticized the EU for "stifling innovation", but it's when the US starts regulating that the major players start complaining!
NVIDIA
@nvidia
The Biden Administration's "AI Diffusion" rule will stifle innovation and undermine America’s global technology leadership. Restricting access to mainstream computing risks derailing AI progress for industries at home and abroad.
America’s strength lies in innovation and competition. This rule will only harm the U.S. economy, set America back, and play into the hands of U.S. adversaries.
nvda.ws/40h8vea -
Kevin Schawinski
@kevinschawinski
👀
Gerardo Fortuna
@gerardofortuna
Over the past eight months, the European Commission has, on at least three occasions, concealed actions or situations involving President Ursula von der Leyen—resorting to lies. This is a strong claim, but I would not make it without evidence
👇This post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
🇪🇺 🏥 From Medical Devices to AI Systems: The EU AI Act’s Surprising “Family Tree”
Did you know the EU AI Act has deep ancestry in the EU Medical Device Regulation (MDR)? This fact doesn't seem to be widely known. If you really want to understand the AI Act, starting with MDR is a good idea!
Both laws share:
🔹 A risk-based framework
🔹 Conformity assessments involving external checks for higher-risk products
🔹 Post-market surveillance and ongoing compliance updates
Where they diverge? The AI Act is much broader (covering everything from credit scoring to law enforcement), and it focuses heavily on fundamental rights, tackling biases, discrimination, and societal impacts in addition to safety.
This “MDR heritage” means a tiered approach to AI oversight: high-risk AI systems will need robust documentation, strict assessments, and ongoing monitoring—much like high-risk medical devices do today.
What's even more interesting is that countries around the world are writing their own AI laws with the EU AI Act as a template. But do they know about the MDR heritage? I doubt it.
If you’re curious how this will shape your AI strategy, check out the table below. It compares the two frameworks, side by side. You’ll see just how familiar MDR’s DNA is within the AI Act! -
Kevin Schawinski
@kevinschawinski
Just wait until the agents ask for PTO, retirement benefits and free chips in the office.
-
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
AI, telecom, cloud, & data will be the digital priorities of the 🇵🇱 PCY for the next six months, according to docs seen by @mlexclusive. The Poles will organize a series of discussions and political statements to influence upcoming EU initiatives.
www.mlex.com/mlex/articles/2281235/ai-telecom-cloud-data-… -
Kevin Schawinski
@kevinschawinski
👀
Luis Montezuma | @luismontezuma@mstdn.social
@montezumachavez
The General Court orders the Commission to pay damages to a visitor to its ‘Conference on the Future of Europe’ website as a result of the transfer of personal data to the United States
-
Kevin Schawinski
@kevinschawinski
💡 I have been working on updating our AI policy at Modulos. This got me thinking: wouldn't it be great if there was an AI policy template that anyone could use and customize to their organization's needs? And wouldn't it be even better if this template had been developed by experts in the field from around the world?
Kevin Schawinski
@kevinschawinski
That's why I started this Google Doc with a template AI policy based on ISO 42001. It's released under CC-BY 4.0 so anyone can use it. If you'd like to contribute to reviewing and editing it, follow the link below. And of course, feel free to adopt it in your organization!
Kevin Schawinski
@kevinschawinski
Link: docs.google.com/document/d/1xxccbLEUkVhTYTorOw8CSAbQ0kCyx…
-
Kevin Schawinski
@kevinschawinski
Analogies are always hard, but we are still in very early days for AI, comparable to 1995 for the web.
-
Kevin Schawinski
@kevinschawinski
We'll have to move to only accepting hand-written applications.
el.cine
@EHuanglu
This is incredible..
This AI agents can go through your CV, find matching jobs online, and start applying for them on its own. -
Kevin Schawinski reposted
Ethan Mollick
@emollick
Working paper argues AI creates an "inflection point" for each job type.
Before that, AI boosts freelancer earnings (web devs saw a +65% increase). After, AI replaces freelancers (translators saw -30% drop). They suggest that once AI starts replacing a job, it doesn't go back. -
Kevin Schawinski
@kevinschawinski
It's wild that this works and gives real insights!
-
Kevin Schawinski
@kevinschawinski
🇺🇸 🤠 ANALYSIS: Today I am taking a look at the Texas Responsible AI Governance Act” (TRAIGA), which was introduced earlier in 2024 and is under consideration by the Texas state legislature. While California has been busy trying to pass a whole suite of AI laws covering different aspects, Texas has put forward a comprehensive AI bill which is quite similar in structure to the EU AI Act. Let me break it down for you with my friend... o1-pro.
Scope & Coverage:
-Applies to non–small businesses deploying or developing “high-risk” AI in Texas
- “High-risk” means AI that makes—or heavily influences—“consequential decisions” like hiring, lending, or sentencing.
- Explicitly bans manipulative, biometrics-based surveillance and “emotion recognition” AI without consent.
- The law would come into force on 1 September 2025, if passed.
- The law also applies to all companies "doing business in the state of Texas", which would likely include not just out of state companies, but international companies whose products are available in Texas.
High Risk definition:
Under Section 551.001(13) of the draft, any “high-risk artificial intelligence system” is one that makes or is a contributing factor in a “consequential decision” that can significantly affect someone’s legal, financial, or personal situation. That means, for instance, AI used in:
- Criminal justice (risk assessments, sentencing recommendations)
- Employment decisions (hiring, promotion, or firing)
- Educational admissions or opportunities
- Financial services (lending, credit approval)
- Housing (tenant selection, mortgage approvals)
- Healthcare (treatment, insurance approvals)
- Essential services (electricity, food, water access)
If your AI system influences such a decision—even if it’s just one factor among many—then it’s considered “high-risk.”
Enforcement:
- Civil penalties up to $100k/violation, or up to $20k per day, enforced by the Attorney General.
30-day “opportunity to cure” if you slip up—so a bit more breathing room than the EU approach.
- Consumers can sue for injunctive relief if the AI use is outright prohibited (subchapter B).
- Regulatory sandbox established for AI pilots (36-month testing period). A nod to fostering innovation but definitely watch the compliance rules.
Opens for me:
- Will other US states copy Texas’s risk-based approach, or push for broader coverage?
- Can the Texas AG handle the significant oversight obligations here?
- Will the “sandbox” approach really work out? There's lots of talk about sandboxes in the EU, but little progress so far?
- How does this mesh with federal frameworks like the pending SB 3312?
What do you think - could TEXAS be the role model for state-level AI legislation in the US?
#aigovernance #airegulation #texas #techpolicy
Kevin Schawinski
@kevinschawinski
Copied over from my post on LinkedIn: www.linkedin.com/posts/kevin-schawinski_draft-texas-respo…
-
Kevin Schawinski
@kevinschawinski
AI slop has destroyed the infosphere.
Lukas Kidzinski
@kidzik
I've changed the history. Literally, I’ve altered EU decisions from 1997.
Two days ago, motivated by the EU USB mandate, I posted a 𝘀𝗮𝘁𝗶𝗿𝗶𝗰𝗮𝗹 𝘀𝘁𝗼𝗿𝘆 about how they had already mandated some other PC connectors in 1997. -
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
🚨 BREAKING: Global AI regulation just got real—and 2025 is only two days old!
Happy New Year, everyone! If you’re in the AI space, you’ve likely noticed that the holiday lull didn’t slow down the regulatory wave. Here’s what’s happening right now (January 2, 2025):
- 🇪🇺 EU AI Act: In force since August 2024, major deadline: August 2026
- 🇰🇷 South Korea AI Basic Law: Key provisions take effect late 2025
- 🇯🇵 Japan: Announced plans for a new AI regulation
- 🇺🇸 Colorado, USA: Passed an AI law demanding compliance by February 2026
- 🇺🇸 Texas “TRAIGA” (Texas Responsible AI Governance Act): Under active consideration
- 🇺🇸 US Senate SB 3312: Would implement federal AI rules
- 🇧🇷 Brazil: Comprehensive AI bill passed the Senate, now with the lower house; 1-year transition
If your AI product or service is online in these regions (even if you’re on a different continent), you’re covered. That’s how broad these rules are.
What Makes These Laws Challenging for Companies?
⚖️ Broad Applicability: These laws often apply to any company offering AI products or services in a jurisdiction, even if the company is based elsewhere. This means startups in Europe can be bound by regulations in the U.S. or Brazil.
📄 Extensive Requirements: Companies must conduct risk assessments, maintain detailed documentation, and ensure AI systems are transparent, explainable, and free from bias—all of which require significant resources.
💶 Strict Penalties: Non-compliance can result in severe fines, such as those in the EU AI Act—up to 7% of global revenue. This makes regulatory risks a top concern for businesses of all sizes.
🪢 Complex Oversight: Many jurisdictions are assigning sector-specific regulators or introducing entirely new oversight mechanisms, creating a patchwork of compliance that’s difficult to navigate.
How Do You Stay Ahead of Global Compliance?
💡 Enter ISO 42001, the AI Management System standard. Here’s why it’s a game-changer: Think of it like ISO 27001, but for AI: It establishes best practices for risk management, governance, and transparency—all the elements regulators want.
🌏 Global Scalability: One framework, many jurisdictions. Instead of juggling compliance checklists for each region, you start with a universal blueprint.
⏱️ Early Adoption = Credibility: Certifying your AIMS takes months. Starting now signals that you’re ready for the next big regulatory challenge.
What do you think—are these overlapping regulations going to be a challenge for your organization?
#airegulation #aiact #responsibleai #aigovernance #compliance #aistandards -
Kevin Schawinski
@kevinschawinski
🚨 BREAKING: Global AI regulation just got real—and 2025 is only two days old!
Happy New Year, everyone! If you’re in the AI space, you’ve likely noticed that the holiday lull didn’t slow down the regulatory wave. Here’s what’s happening right now (January 2, 2025):
- 🇪🇺 EU AI Act: In force since August 2024, major deadline: August 2026
- 🇰🇷 South Korea AI Basic Law: Key provisions take effect late 2025
- 🇯🇵 Japan: Announced plans for a new AI regulation
- 🇺🇸 Colorado, USA: Passed an AI law demanding compliance by February 2026
- 🇺🇸 Texas “TRAIGA” (Texas Responsible AI Governance Act): Under active consideration
- 🇺🇸 US Senate SB 3312: Would implement federal AI rules
- 🇧🇷 Brazil: Comprehensive AI bill passed the Senate, now with the lower house; 1-year transition
If your AI product or service is online in these regions (even if you’re on a different continent), you’re covered. That’s how broad these rules are.
What Makes These Laws Challenging for Companies?
⚖️ Broad Applicability: These laws often apply to any company offering AI products or services in a jurisdiction, even if the company is based elsewhere. This means startups in Europe can be bound by regulations in the U.S. or Brazil.
📄 Extensive Requirements: Companies must conduct risk assessments, maintain detailed documentation, and ensure AI systems are transparent, explainable, and free from bias—all of which require significant resources.
💶 Strict Penalties: Non-compliance can result in severe fines, such as those in the EU AI Act—up to 7% of global revenue. This makes regulatory risks a top concern for businesses of all sizes.
🪢 Complex Oversight: Many jurisdictions are assigning sector-specific regulators or introducing entirely new oversight mechanisms, creating a patchwork of compliance that’s difficult to navigate.
How Do You Stay Ahead of Global Compliance?
💡 Enter ISO 42001, the AI Management System standard. Here’s why it’s a game-changer: Think of it like ISO 27001, but for AI: It establishes best practices for risk management, governance, and transparency—all the elements regulators want.
🌏 Global Scalability: One framework, many jurisdictions. Instead of juggling compliance checklists for each region, you start with a universal blueprint.
⏱️ Early Adoption = Credibility: Certifying your AIMS takes months. Starting now signals that you’re ready for the next big regulatory challenge.
What do you think—are these overlapping regulations going to be a challenge for your organization?
#airegulation #aiact #responsibleai #aigovernance #compliance #aistandards
Kevin Schawinski
@kevinschawinski
You can get started building your ISO 42001 AI Management System for free by signing up here: www.modulos.ai/pricing/
-
Kevin Schawinski
@kevinschawinski
Me too, o1-pro, me too.
-
Kevin Schawinski
@kevinschawinski
I continue to be blown away by o1-pro's capabilities.
-
Kevin Schawinski
@kevinschawinski
(b) is massively underappreciated.
Ethan Mollick
@emollick
Easy prediction for 2025 is that the gains in AI model capability will continue to grow much faster than (a) the vast majority of people’s understanding of what AI can do & (b) organizations’ ability to absorb the pace of change. Social change is slower than technological change.
-
Kevin Schawinski
@kevinschawinski
Hmmm 👀
Logan Kilpatrick
@OfficialLoganK
Straight shot to ASI is looking more and more probable by the month… this is what Ilya saw
-
Kevin Schawinski
@kevinschawinski
The reaction to this rule is very telling.
The tech elite hates it and sees it as the EI stifling innovation.
Regular people love it because they hate paying EUR 50 for yet another cable that serves a single device.
European Commission
@EU_Commission
It’s time for THE charger.
Today, the USB-C becomes officially the common standard for charging electronic devices in the EU.
It means better-charging technology, reduced e-waste, and less fuss to find the chargers you need.
#digitaleu -
Kevin Schawinski
@kevinschawinski
Hear me out: Indiana Jones and the Cube of Tungsten.
-
Kevin Schawinski
@kevinschawinski
😮💨
Jason ✨👾SaaStr.Ai✨ Lemkin
@jasonlk
Top advantages to being a first mover:
- No roadmap
- Tiny market
- Customers aren’t sure
- Limited legitimacy
- Never budgeted
- VCs don’t get it
- Many VPs don’t get it
- Tiny TAM at first
- Seems minor to many
- Why wasn’t it already done if it’s so important
It’s great -
Kevin Schawinski
@kevinschawinski
Starlink's Terms of Service select Texas law for disputes on the Moon. As Texas is about to pass its own AI law, the Texas Responsible AI Governance Act, this means that even the MOON will regulate AI! Is there nowhere entrepreneurs can innovate in peace? 🚀 🌔
-
Kevin Schawinski
@kevinschawinski
Since Texas is about to pass an AI law, this means that even the MOON will regulate AI. 😂
Visakan Veerasamy
@visakanv
in 2020 Starlink users had to agree via TOS that disputes on the Moon would be governed by California laws
went to check today and it’s now TexasThis post is unavailable. See it on X. -
Kevin Schawinski
@kevinschawinski
o1-pro is as big a jump in capability as 3.5 to 4.
Not sure what to make of its impact yet. -
Kevin Schawinski
@kevinschawinski
Also, profits are the easiest thing to make go away….
Reid Southen
@Rahll
Microsoft and OpenAI's definition for AGI has nothing to do with intelligence. They simply agreed they'll have reached AGI when they develop a system that can generate $100 billion in profit.
Anyone still buying the AGI hype after this is officially a sucker. -
Kevin Schawinski
@kevinschawinski
🇯🇵 🚨 BREAKING: Japan just announced its AI framework and it's fascinating - threading the needle between EU and Korean approaches! Here's what you need to know:
Key Components:
- ❗ New AI legislation coming soon
- Formation of AI Strategy HQ with all cabinet ministers
- Focus on making Japan "world's most AI R&D friendly nation"
- Guidelines based on the Hiroshima AI Process
- Government procurement guidelines for AI
Most interesting angle? Japan seems to be taking bits from both the Korean and EU playbooks. Like Korea, they're focusing on innovation and competitiveness, but they're adding more structured oversight like the EU - just lighter touch.
The timing is notable - just after Korea's announcement and with the EU AI Act implementation starting. Looks like we're seeing an "Asian model" of AI governance emerge: pro-innovation but with guardrails.
With a realistic date of passing the bill in summer 2025, the 1 year implementation period aligns Japan pretty well with the main EU AI Act deadline of August 2026.
Big question: Will this three-way regulatory competition between EU, Korea and Japan drive global standards, or will we see regulatory arbitrage? Watch this space...
#aigovernance #airegulation #japan #aipolicy #techpolicy #regulation
Thoughts? Particularly curious about three-way interoperability prospects between these frameworks.
Kevin Schawinski
@kevinschawinski
Prime Minister's office announcement: www.kantei.go.jp/jp/103/actions/202412/26ai.html
Interim report: www8.cao.go.jp/cstp/ai/ai_senryaku/12kai/shiryou1.pdf -
Kevin Schawinski
@kevinschawinski
🇰🇷🚨 South Korea has just passed the second AI law in the world after the #euaiact.
South Korea's new AI law requires companies to assess risks and ensure transparency for high-impact AI systems in critical sectors like healthcare, infrastructure, and public services.
The law mandates special oversight of generative AI, requires domestic representatives for foreign companies, and establishes a national AI commission to guide policy.
Companies have one year to comply with most provisions, facing penalties up to 30M won (~$23,000) per violation.
#koreaailaw #airegulation
Kevin Schawinski
@kevinschawinski
Source: biz.chosun.com/en/en-it/2024/12/26/66W2Z3RX6FE7FMPXMR73T2…
-
Kevin Schawinski
@kevinschawinski
First impressions of developing with o1-pro:
- It's very powerful!
- It solves problems in one shot significantly more often than any other model, including Claude 3.5 Sonnet
- Code often runs error-free on the first try
- It still sometimes makes weird or simple mistakes
- It's much harder to steer. It's a know-it-all and will insist that it knows better more than any model I've interacted with
Very impressed so far. Sorry, Claude...Post not available here, view on X
Kevin Schawinski
@kevinschawinski
It keeps giving me this error even though it's reading compliance documents or writing boring code.
Kevin Schawinski
@kevinschawinski
As soon as agent systems with this model work, it's going to get very tight for anyone doing knowledge work. It's not AGI, it's not a genius, but it's smarter than most people in a narrow way.
-
Kevin Schawinski
@kevinschawinski
And so it begins....
-
Kevin Schawinski
@kevinschawinski
🇺🇸 ⚖️ 🎄 We got an updated SB 3312 as a Christmas present!
I did a check for changes and did a quick "compare and contrast" with the EU AI Act together with my friend Claude.
Scope & Coverage:
- More targeted AI definition than EU Act - focuses on "inference" and "machine-based systems", may exclude simpler methods like logistic regression that EU explicitly includes
- Two key categories: "High-Impact" (decisions on housing, credit, etc.) and "Critical-Impact" (biometrics, infrastructure, criminal justice)
- Constitutional rights angle unique to US framework - systems affecting constitutional rights could be covered
Enforcement:
- Much lighter penalties than EU Act - max $300k per violation or 2x system value (vs EU's up to 7% global revenue)
- Centralized federal enforcement via Commerce Department
- No private right of action or state AG enforcement (unlike many US laws)
- Focus seems more on cooperation than punishment
Timeline:
- Quick hits: Generative AI transparency in 180 days
- Slower build: 1-3 years for agency guidelines & standards
- Less structured than EU Act - unclear transition periods for existing systems
- No explicit grace periods or grandfathering
So what are the key differences? The EU Act builds a comprehensive new regulatory regime, while this US bill creates a lighter federal oversight framework focused on highest-risk systems. The penalties are lower and enforcement is entirely based on agency action. I think that point alone will make many US tech companies happy... no class action lawsuits!
This bill has now been introduced in the US Senate. This certainly will not be its final form, but it's interesting to see the US trying to assert leadership in the AI regulation space despite the supposed anti-regulation mood in Washington DC.
Do you think this bill stands a chance of passing? -
Kevin Schawinski
@kevinschawinski
Today I am coding with o1. Very different vibes from Claude, but the code is good!
Kevin Schawinski
@kevinschawinski
Kinda thinking of giving o1-pro ago. What say you, @errai34?
-
Kevin Schawinski
@kevinschawinski
The Albanian PM on why the country is banning TikTok for at least a year. Good!
Kevin Schawinski
@kevinschawinski
Source: www.euractiv.com/section/tech/news/albania-announces-shut…
-
Kevin Schawinski reposted
Pranav Khetarpal
@pktrpl
One thing that I've been thinking... o3 and o1 getting such high performances on these benchmarks at ~80%... Why does this number feel like they have been trained on it?
These BMs have been out for a while now.... -
Kevin Schawinski
@kevinschawinski
👀👀👀👀
Gary Marcus
@GaryMarcus
Update: It IS true, from @fchollet’s blog. 𝗢𝟯 𝘄𝗮𝘀 𝘁𝗿𝗮𝗶𝗻𝗲𝗱 𝗼𝗻 𝟳𝟱% 𝗼𝗳 𝘁𝗵𝗲 𝗽𝘂𝗯𝗹𝗶𝗰 𝘀𝗲𝘁 𝗳𝗼𝗿 𝗔𝗥𝗖-𝗔𝗚𝗜.
Yesterday’s announcement shows what you can do when teach to the test and spend enormous amounts of money.
How much the demo says about real world applications very much remains to be seen.
In my personal opinion, what we saw yesterday was in part a publicity stunt, with the very important fact below not transparently disclosed in the live feed.
(Also, I really have to wonder whether enthusiasts like @kevinroose were aware of this fact, and if they were, whether they are aware of the implications.) -
Kevin Schawinski
@kevinschawinski
Captchas are getting harder every year… 😩
-
Kevin Schawinski
@kevinschawinski
This reminds me of Douglas Adam’s’ HHGTG, where the AI just needs to think long enough to get the answer to Life, the Universe and Everything.
François Chollet
@fchollet
Today OpenAI announced o3, its next-gen reasoning model. We've worked with OpenAI to test it on ARC-AGI, and we believe it represents a significant breakthrough in getting AI to adapt to novel tasks.
It scores 75.7% on the semi-private eval in low-compute mode (for $20 per task in compute ) and 87.5% in high-compute mode (thousands of $ per task). It's very expensive, but it's not just brute -- these capabilities are new territory and they demand serious scientific attention.
Kevin Schawinski
@kevinschawinski
OpenAI should do a $100M dollar run and ask it for the final answer.
-
Kevin Schawinski
@kevinschawinski
👀
François Chollet
@fchollet
Today OpenAI announced o3, its next-gen reasoning model. We've worked with OpenAI to test it on ARC-AGI, and we believe it represents a significant breakthrough in getting AI to adapt to novel tasks.
It scores 75.7% on the semi-private eval in low-compute mode (for $20 per task in compute ) and 87.5% in high-compute mode (thousands of $ per task). It's very expensive, but it's not just brute -- these capabilities are new territory and they demand serious scientific attention. -
Kevin Schawinski
@kevinschawinski
Absolutely fascinating list. CNRS! Getty!
Luca Bertuzzi
@BertuzLuca
Finally, here is the list of model providers participating in the EU's code of practice for general-purpose AI. A few surprises in there.
-
Kevin Schawinski
@kevinschawinski
🚨🇮🇹Italy's DPA fines @OpenAI EUR 15M for GDPR violations.
www.reuters.com/technology/italy-fines-openai-15-million-… -
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
Copyright-related measures were significantly toned down in the second draft of the EU’s Code of Practice for AI models, notably for downstream compliance. The new text recognizes the code's importance for fine-tuners, introducing various SME exemptions.
www.mlex.com/mlex/articles/2276696/ai-copyright-measures-… -
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
🇺🇸⚖️Really fascinating moves in AI regulation: As the EU begins implementing its AI Act, US Senate Majority John Thune just calendared S.3312, the "AI Research, Innovation, and Accountability Act."
Time for a quick compare & contrast as this bill is fascinating:
Both adopt risk-based approaches, but with key differences:
EU AI Act: Four clear tiers (unacceptable, high, limited, minimal risk) with specific use cases for each.
S.3312: Two categories:
"High-impact": AI systems specifically developed to make decisions about housing, employment, credit, education, healthcare, or insurance that pose significant risks to Constitutional rights or safety
"Critical-impact": AI systems for biometrics, criminal justice, space infrastructure, and critical infrastructure management that pose significant risks to Constitutional rights or safety
The scope? That's where it gets interesting. Just the "critical infrastructure" component references 16 major sectors - from energy to healthcare to financial services in US law. But there's a crucial qualifier: systems must be for "direct management and operation" AND pose "significant risks."
Until we see agency interpretation and court decisions, it's hard to tell if this catches just core operational AI (like power grid management) or extends to supporting functions (like predictive maintenance). The novel "significant risk to Constitutional rights or safety" standard adds another layer of uncertainty.
Small business impact: EU Act includes SME considerations and proportionality provisions. S.3312 exempts smaller platforms but - crucially - has NO size-based exemptions for high/critical-impact systems.
Enforcement: EU Act can fine up to 6% of global revenue. S.3312 caps at $300k or 2x transaction value.
Definitely worth following closely: will this lighter-touch US approach gain traction as a counterpoint to the EU's comprehensive regulation, or is it too late in the game as companies orient towards the AI Act? And how will its ambiguous scope be interpreted? "Significant risks to Constitutional rights or safety" sounds like lots of fun for US lawyers.
Thoughts? -
Kevin Schawinski
@kevinschawinski
🇺🇸⚖️Really fascinating moves in AI regulation: As the EU begins implementing its AI Act, US Senate Majority John Thune just calendared S.3312, the "AI Research, Innovation, and Accountability Act."
Time for a quick compare & contrast as this bill is fascinating:
Both adopt risk-based approaches, but with key differences:
EU AI Act: Four clear tiers (unacceptable, high, limited, minimal risk) with specific use cases for each.
S.3312: Two categories:
"High-impact": AI systems specifically developed to make decisions about housing, employment, credit, education, healthcare, or insurance that pose significant risks to Constitutional rights or safety
"Critical-impact": AI systems for biometrics, criminal justice, space infrastructure, and critical infrastructure management that pose significant risks to Constitutional rights or safety
The scope? That's where it gets interesting. Just the "critical infrastructure" component references 16 major sectors - from energy to healthcare to financial services in US law. But there's a crucial qualifier: systems must be for "direct management and operation" AND pose "significant risks."
Until we see agency interpretation and court decisions, it's hard to tell if this catches just core operational AI (like power grid management) or extends to supporting functions (like predictive maintenance). The novel "significant risk to Constitutional rights or safety" standard adds another layer of uncertainty.
Small business impact: EU Act includes SME considerations and proportionality provisions. S.3312 exempts smaller platforms but - crucially - has NO size-based exemptions for high/critical-impact systems.
Enforcement: EU Act can fine up to 6% of global revenue. S.3312 caps at $300k or 2x transaction value.
Definitely worth following closely: will this lighter-touch US approach gain traction as a counterpoint to the EU's comprehensive regulation, or is it too late in the game as companies orient towards the AI Act? And how will its ambiguous scope be interpreted? "Significant risks to Constitutional rights or safety" sounds like lots of fun for US lawyers.
Thoughts?
Kevin Schawinski
@kevinschawinski
Track the Bill here: www.congress.gov/bill/118th-congress/senate-bill/3312
-
Kevin Schawinski
@kevinschawinski
How come Teams / Meet / Zoom et are not interoperable?
-
Kevin Schawinski
@kevinschawinski
A quick look at the slate of proposed and passed laws on AI in California alone goes against this narrative.
Ole Lehmann
@itsolelehmann
Europe vs USA & China
-
Kevin Schawinski
@kevinschawinski
If technical standards are required to do business, then obviously they should be freely available to anyone. But if we want standards to exist in the first place, we need to fund their creation and maintenance.
Clearly the current system of selling access to essential standards is not sustainable.
Luca Bertuzzi
@BertuzLuca
HUGE❗️The EU standard-setting architecture might be crumbling after IEC and ISO sued the Commission before the EU Court of Justice. The lawsuit follows the Malamud ruling that disrupted the business model of standardization bodies at a critical time.
www.mlex.com/mlex/technology/articles/2275104/companies-t… -
Kevin Schawinski
@kevinschawinski
👀
European Commission
@EU_Commission
Today, we have opened formal proceedings against TikTok for a suspected breach of the Digital Services Act regarding its obligation to assess and mitigate systemic risks linked to election integrity, notably in the context of the recent Romanian presidential elections ↓
-
Kevin Schawinski
@kevinschawinski
This is such an interesting plot because it suggests that Americans react to crisis with innovation, while Europeans don’t.
Kevin Schawinski
@kevinschawinski
You can even see where Americans slacked a bit post-crisis....
-
Kevin Schawinski
@kevinschawinski
Someone should start a startup to develop AI customers who won’t complain about buggy products and just keep consuming without any work life balance. They can even pay with worthless crypto sh*tcoins.
-
Kevin Schawinski reposted
shreya rajpal
@ShreyaR
Big week for Guardrails AI! Yesterday, we launched two new state-of-the-art guardrails for PII detection and jailbreak detection 🚀
For PII, we achieve a whopping 76% better(!!) F1 score compared to Microsoft Presidio averaged across all PII categories on the benchmark.
For Jailbreaking, we're state of the art of accuracy and F1 compared to Llama Prompt Guard, Microsoft Prompt Shield and the leading open source jailbreak detection models.
That's not all though -- the best thing about both of these new state of the art guardrails is that they're fully open source, and available on Guardrails Hub today!
Link to blog: www.guardrailsai.com/blog/advanced-pii-and-jailbreak
PII Guardrail on Hub: hub.guardrailsai.com/validator/guardrails/guardrails_pii
Jailbreak Guardrail on Hub: hub.guardrailsai.com/validator/guardrails/detect_jailbreak -
Kevin Schawinski
@kevinschawinski
👀
Bloomberg Markets
@markets
JPMorgan plans to add 1,500 workers to its Buenos Aires corporate center over the next five years www.bloomberg.com/news/articles/2024-12-12/jpmorgan-plans…
-
Kevin Schawinski
@kevinschawinski
Slackbot, late at night, stressing over its quarterly sales target: “I have an idea!”
-
Kevin Schawinski reposted
European Commission
@EU_Commission
Your talent, Europe's AI lead.
These are the first 7 AI Factories to be created across the EU:
🇬🇷 Athens
🇸🇪 Linköping
🇩🇪 Stuttgart
🇱🇺 Bissen
🇮🇹 Bologna
🇪🇸 Barcelona
🇫🇮 Kajaani
They will combine computing power, data, and talent to develop innovative AI solutions ↓ -
Kevin Schawinski
@kevinschawinski
Today I am in Brussels for a workshop on AI Literacy for members of the @EU_Commission AI Pact. 🇪🇺🤖
-
Kevin Schawinski reposted
Luca Bertuzzi
@BertuzLuca
NEW: The 🇭🇺presidency aggregated EU countries' #ai and copyright inputs. While most member states don't think new legislation is needed yet, several consider it will be necessary in the future to maintain a harmonized framework and floated a few ideas.
www.mlex.com/mlex/artificial-intelligence/articles/227121… -
Kevin Schawinski
@kevinschawinski
Excuse me, what?
www.quantamagazine.org/fish-have-a-brain-microbiome-could… -
Kevin Schawinski reposted
VGT - VEREIN GEGEN TIERFABRIKEN
@vgt_at
❤ 30 Tage VEGAN! ❤
👉 Starte das Jahr 2025 mit einem guten Vorsatz und probiere mal VEGAN.
🐮🐷🌱Gut für Mensch, Tier und Umwelt.
🫶 vegan.at/monat !B -
Kevin Schawinski
@kevinschawinski
LLMs still struggle with humor.... #o1
-
Kevin Schawinski
@kevinschawinski
The legal definition of AI has nothing to do with technical or colloquial use of the term.
*Logistic regression is AI* according to the @EU_Commission
Luca Bertuzzi
@BertuzLuca
Banks have been seeking to exclude logistic regression, widely used for credit scoring in the financial sector, from the #aiact's definition of AI system via upcoming guidelines. But the Commission said at a recent meeting it considers these techniques AI.
www.mlex.com/mlex/articles/2269802/the-financial-sector-s…
Kevin Schawinski
@kevinschawinski
After reading the very early drafts of the AI Act (years ago now), I figured that the AI Act will do to *algorithms* what GDPR did to *data*.
GDPR makes us think about ALL the data we process and whether we do so legally.
The AI Act will make us think about ALL the algorithms. -
Kevin Schawinski
@kevinschawinski
This week I'm in Washington DC. 🇺🇸 🦅
Yesterday, I took part in the National Institute of Standards and Technology (@NIST ) AI Safety Institute Consortium plenary meeting hosted by UMD to discuss our progress in building guidelines and standards for AI measurement and AI safety.
As part of my visit to DC, I'm also taking part in other events and meeting with people across the AI value chain from both government and the private sector. -
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
Press release from Amazon: aws.amazon.com/blogs/machine-learning/aws-achieves-iso-ie…
-
Kevin Schawinski
@kevinschawinski
This won’t only affect physicists but other sciences too.
Start building and make the world better if you want the taxpayer to continue supporting you.
Sabine Hossenfelder
@skdh
I've been telling physicists in the foundations for more than 10 years they need to get their act together or their future looks bleak.
There have been too many unobserved particles, too many idiotic headlines about multiverses, too many expensive experiments that didn't find anything interesting, too many "predictions" that never came to pass -- turning the word "prediction" into a joke -- too much silly talk about beautiful and elegant ideas and theories of everything that remained wishful thinking. And never once did they offer an explanation for what went wrong or how they are going to improve.
Simply put, they have spent too much of taxpayers' money without producing anything of value.
Can you not see the writing on the wall?
Americans are going to crack down hard on academia because of all the nonsense studies that have befallen their universities. They're not going to stop there.
And Europeans are sick and tired of paying ridiculously high amounts of taxes. In the next years you will see a dramatic political shift to deregulation over here too.
In Europe it's hard to fire people but if they get no further funding, many research areas are simply going to die off.
It's not that I wanted this to happen. But this is what I think will happen. And at this point it's too late to do anything about it. -
Kevin Schawinski reposted
UniverseTBD
@universe_tbd
Excited to announce the Multimodal Universe (MMU): a huge 100TB dataset bringing together the largest ML-focused collection of astronomical observations ever assembled to accelerate open AI and astronomy research
github.com/MultimodalUniverse/MultimodalUniverse
Think ImageNet, but for space 🧵 -
Kevin Schawinski
@kevinschawinski
I have 26 times more followers on X than on the blue site, yet I get vastly more engagement there. Is X just dead accounts and bots now? Am I de-ranked as "boring"?
-
Kevin Schawinski
@kevinschawinski
ChatGPT returns an error when you ask it about “David Mayer”. There’s lots of speculation on who David Mayer is and why OpenAI blocks his name, including some conspiracy theories.
My guess is a David Mayer filed a GDPR data erasure request (aka right to be forgotten) and OpenAI put in a hard guardrail. -
Kevin Schawinski
@kevinschawinski
I can’t recommend this tour enough. Tea and Peter really know what they are talking about.
This post is unavailable. See it on X. -
Kevin Schawinski reposted
Kevin Schawinski
@kevinschawinski
Amazon Web Services (AWS) @awscloud just announced they're the first major cloud provider to achieve ISO/IEC 42001:2023 certification for their AI services.
The certification covers major AI services: Amazon Bedrock, Amazon Q Business, Amazon Textract, and Amazon Transcribe
This is a wake-up call for every organization developing or deploying AI. The tech leaders are moving fast on AI compliance - and for good reason. As AI regulation intensifies globally, ISO 42001 certification isn't just a nice-to-have, it's becoming a must have for building trust with customers and partners. Or even to meet the minimum requirement to be considered.
We're seeing this shift firsthand. Our AI GRC platform has been ISO 42001 assessed, and we're helping organizations achieve their own certifications efficiently. Auditors I've spoke with estimate that a large organizations will need 9-12 months to achieve their ISO 42001. Modulos can help you get audit-ready in a lot less.
The question isn't if your organization will need this certification, but when. And those who move first will have a significant competitive advantage.
AWS's announcement validates this.
#ai #compliance #iso42001 #responsibleai #aigovernance #airegulation -
Kevin Schawinski
@kevinschawinski
And the CEO should be doing all those.
Jason ✨👾SaaStr.Ai✨ Lemkin
@jasonlk
Every time I see a CRO that doesn't sell themselves, they fail.
Every time I see a CTO that doesn't code themselves, they fail.
Every time I see a CMO that doesn't market themselves, they fail. -
Kevin Schawinski
@kevinschawinski
Amazon Web Services (AWS) @awscloud just announced they're the first major cloud provider to achieve ISO/IEC 42001:2023 certification for their AI services.
The certification covers major AI services: Amazon Bedrock, Amazon Q Business, Amazon Textract, and Amazon Transcribe
This is a wake-up call for every organization developing or deploying AI. The tech leaders are moving fast on AI compliance - and for good reason. As AI regulation intensifies globally, ISO 42001 certification isn't just a nice-to-have, it's becoming a must have for building trust with customers and partners. Or even to meet the minimum requirement to be considered.
We're seeing this shift firsthand. Our AI GRC platform has been ISO 42001 assessed, and we're helping organizations achieve their own certifications efficiently. Auditors I've spoke with estimate that a large organizations will need 9-12 months to achieve their ISO 42001. Modulos can help you get audit-ready in a lot less.
The question isn't if your organization will need this certification, but when. And those who move first will have a significant competitive advantage.
AWS's announcement validates this.
#ai #compliance #iso42001 #responsibleai #aigovernance #airegulation
Kevin Schawinski
@kevinschawinski
Press release from Amazon: aws.amazon.com/blogs/machine-learning/aws-achieves-iso-ie…
-
Kevin Schawinski reposted
UniverseTBD
@universe_tbd
yo dawg we heard you liked social media 🦋
-
Kevin Schawinski
@kevinschawinski
👀
Bloomberg
@business
The US Justice Department seeks to unwind Google's partnership with AI startup Anthropic as part of an effort to resolve a landmark antitrust case over online search www.bloomberg.com/news/articles/2024-11-21/us-justice-dep…
-
Kevin Schawinski
@kevinschawinski
🤣🤣🤣
Karpi
@karpi
Gemäs einer Studie der Uni Zürich will ein Grossteil der Konsument:innen keine News, die mit KI generiert wurden. Aber würden sie auch dafür bezahlen? Und wie kann KI-generierter Content von organischem Content unterschieden werden?
Deswegen lancieren wir jetzt NATURA CONTENT™This media couldn't be displayed. See it on the original post. -
Kevin Schawinski
@kevinschawinski
X added a setting for "we'll take your data to train grok" without any notice and just defaulted to "yes" for everyone.
This is BAD. -
Kevin Schawinski
@kevinschawinski
The @FTC is investigating @OpenAI and the document outlining their questions is fascinating.
🧵Some highlights: -
Kevin Schawinski
@kevinschawinski
The European Union has just agreed on a Compromise Text for its EU AI Act.
The EU AI Act will set the global standard for how AI systems must be built. Think of it as the GDPR for AI.
I'm going through the text to highlight important changes... 🧵 1/n -
Kevin Schawinski
@kevinschawinski
First #chatgpt answer which made me pause.
-
Kevin Schawinski
@kevinschawinski
Dark humor for these dark times: Ukrainians don’t need to worry about declaring their captured Russian tanks as income. 🇺🇦
en.interfax.com.ua/news/general/804441.html -
Kevin Schawinski
@kevinschawinski
Fantastic privilege telling astronomers at @Yale @YaleAstronomy about automated machine learning with @Modulos_ai today.
-
Kevin Schawinski
@kevinschawinski
Few books have made me see the world in a completely different light as @StephanieKelton’s “The Deficit Myth” has.
Highly recommended! -
Post not available here, view on X
Kevin Schawinski
@kevinschawinski
The PSFGAN paper also has cats! arxiv.org/abs/1803.08925
-
Kevin Schawinski
@kevinschawinski
I have so many questions
-
Kevin Schawinski
@kevinschawinski
Double cake to celebrate acceptance of two papers by @AnnaKWeigel
-
Kevin Schawinski
@kevinschawinski
.@NCaplar & co analyze gender bias in astrophysics papers. Fascinating read: arxiv.org/abs/1610.08984
-
Kevin Schawinski
@kevinschawinski
Slow evening of great food & paper reviewing in #santiago. Soundtrack: @KsenijaSidorova (whom I first heard at Zurich.Minds)
-
Kevin Schawinski
@kevinschawinski
Tonight: Symposium on #citizenscience 2.0 at @swissnexSF w @FuturICT @LucyFortson @myelbows, U. Genick : www.ethz.ch/en/the-eth-zurich/global/eth-meets-you/eth-me…
-
Kevin Schawinski
@kevinschawinski
.@NSF’s France Cordova highlights role of citizen science in future astronomy research #aas227
-
Ksenija Sidorova
@KsenijaSidorova
Some of the brightest minds present their work in all fields-science,medicine,technology etc @ #zurichminds 2day.Getting ready w @aviavital
Kevin Schawinski
@kevinschawinski
@ksenijasidorova @aviavital wonderful music from both of you!
-
Kevin Schawinski
@kevinschawinski
Wyoming just criminalized #citizenscience www.slate.com/articles/health_and_science/science/2015/05… HT @Stuart_Lynn cc @ehafen @EffyVayena
-
Kevin Schawinski
@kevinschawinski
The Queen is simply awesome: www.vox.com/2015/1/23/7877243/king-abdullah-queen-drive












































































































































































































































































































































































