Professor at NYU & Executive Chairman at AMI Labs.
Ex-Chief AI Scientist at Meta.
Researcher in AI, Machine Learning, Robotics, etc.
ACM Turing Award Laureate.
I do not write posts on X.
I tweet links to posts on other platforms.
I like and retweet (occasionally)
I comment on friends' tweets (rarely)
Follow me on...⬇️⬇️⬇️
We trained language models that compress massive contexts into tiny latent representations. Latent Context Language Models (LCLMs) outperform existing KV cache compression methods on the latency/accuracy frontier. 🧵1/10
It's confusing that Trump constantly warns about the security of our elections & foreign interference, yet his administration has taken several steps to dismantle or divert resources away from key safeguards that were in place:
- Gutted the DOJ's Public Integrity Section (went from ~30-40 lawyers down to just 2)
- Left the Election Crimes Branch director position vacant
- Canceled all election integrity training for prosecutors & FBI agents
- Deleted the official 281-page Federal Prosecution of Election Offenses guide from the DOJ website
He disbanded the FBI's Foreign Influence Task Force, which targeted foreign efforts to meddle in U.S. politics & elections.
He also severely cut CISA (the agency for election cybersecurity), firing or transferring dozens of election security specialists, freezing key support programs for states, and proposing to eliminate the entire federal Election Security Program in future budgets.
You would think that if he were ACTUALLY concerned about the integrity & security of our elections, he would be allocating MORE resources, not less — or at least wouldn't be leaving these critical positions vacant & programs dismantled right before the midterms.
As many of you were interested in the technical details of the model, here is a followup thread to go more into technical details about VLA-JEPA.
1. Architecture
2. Training
3. Recipe for the demo
4. TLDR
🧵below
VLA-JEPA just dropped in LeRobot 🤖
What makes this model special is that it does not just learn what action to take from a given observation, it also leverages a JEPA world model to learn action-relevant dynamics.
During training, the VLA leverages V-JEPA2 by conditioning its predictor. This clever trick adds a world modeling objective to the training, which also allows pretraining on human videos.
At inference, the world model is dropped entirely, keeping only a standard VLA architecture: Qwen backbone and action head.
The demo here was only fine-tuned on 13 examples, showing great pretraining capability and running in real time on @NVIDIARobotics DGX Spark!
VLA-JEPA is the first world model to be ported to LeRobot, and I feel like it won't be the last 🚀
@Thom_Wolf@ClementDelangue
Narrative violation: according to @Stanford research, local models can answer 71.3% of real-world chat and reasoning queries accurately, up from 23.2% in 2023. Obviously at a fraction of the cost and energy consumption of frontier APIs.
The obvious conclusion: you don't need a frontier model for most tasks. The future is multi-model: local, open-source, smaller and cheaper for the majority of workloads, frontier APIs when no other choices!
Earlier today, New York University announced the creation of its new Earth Systems Institute—a multidisciplinary hub that will deploy AI and computational tools to better predict environmental changes and to advance means to better prepare and respond to these global phenomena.
A pause continues to be utter and complete nonsense and it always will be.
1. Let's make planes safer by not making planes!
2. What exactly happens in a "pause"? Do labs get to keep working and we all just get to sit on our hands waiting for this eureka moment?
I guess someone gets to keep their checks while we just bring the economy to a crashing halt based on a few people's vague ideas about some imaginary future problems that they came up with while huffing glue and reading Dune!
3. Who will fund the labs when they are not putting out new products?
I guess VCs will continue to just give them 100s of billions out of the goodness of their hearts!
4. What actually justifies a pause?
Apparently, the wonderful world of imagination! Theoretical future problems that haven't happened yet.
Like massive job losses! Um, jobs are increasing including in the areas most affected by AI, like coding, so I guess not that.
So what? I know the jobs apocalypse is coming because I can imagine it and imagination is reality, right?
Maybe advanced AI weapons?
Ah, so the government will pause weapons research too?
Well no, they will keep doing that anyway because they always do that. It's what governments do.
Okay so we're going to ban Chat Bots while the government keeps making weapons? That should solve everything we're worried about with AI!
Well then what about recursively improving models that grow to superintelligence overnight?
Yeah that's not really a thing. That's the plot of an Avengers movie.
Models are bound by the same real world constraints we are like compute (brains/chips) and time (will this drug have side effects in twenty years can only be known in twenty years) and fuzzy multiplicity (not right or wrong but right-ish and wrong-ish means you can't make a reward signal for "is this the right decision for my business") and the subject/object paradox (the thing improving is judging its own improvement. Yeah chew on that one for a bit.)
But I imagined AI overcoming every real world constraint instantly so it's true!
5. How would we know we did everything we needed to do in a pause? How do we know when it's over?
We dont. We just want a pause now because we want it! Don't you see my pause ⏸️ emoji? It's nice right!
6. Who gets to decide we are ready to unpause?
The government or the people!
Great, because vague ass platitudes like this always go well for concrete policy design.
Looks, none of this is real.
It's theater.
It's not real policy. It has no basis in reality. It's a mass hallucination.
It's pushed by people who believe in magic and magical solutions.
And anything that comes out of it will do infinitely more damage than the imaginary thing they were trying to protect us from in the first place.
Striking paper from Wharton. The big conclusion: AI must increase productivity 2.7x -- and quickly -- or tech companies risk bankruptcy with all that entails for the economy. For context: this is how a quickie 2.7x productivity boom would compare to historical precedent. Paper linked in my daily AI digest. Useful context for OpenAI reportedly talking to the US government about a bailout (ahem, I mean ownership stake).
Russell Vought’s proposal to make politics, not peer review, the standard for NIH science grants would be a cataclysm for the American science enterprise.
@jmcl_gtr Congratulations from a long-time fan.
And welcome to the club of honorary doctorate recipients from Université Côte d'Azur!
(mine was in 2022).
David Sarnoff (1891-1971) rose from Russian Jewish immigrant office boy to president/chairman of RCA. He drove commercial radio, founded NBC, and heavily funded early electronic TV development.
Similarities to Elon Musk: Both are high-profile visionary leaders who scaled emerging tech into massive industries through bold bets, team execution, and personal promotion. Sarnoff is often credited with radio/TV breakthroughs developed by engineers. He also joined 1929 RCA stock pool operations that artificially inflated prices (Pecora Commission documented short-term manipulation netting millions before the crash), paralleling criticisms of Musk’s social media moves affecting Tesla valuation.
Book for deeper reading: “David Sarnoff: A Biography” by Eugene Lyons (1966), the standard account.
I proudly and happily lose to him on following metrics:
- helping destroy American democracy
- helping bring to power a fascist and pedophile
- promoting racist ideology
- killing 100s of thousands by dismantling USAID
- disseminating batshit-crazy conspiracy theories
- destroying factuality in the political debate
You have to read this.
Trump demolished the East Wing of the White House and replaced it with a private ballroom funded by corporate donors. Now we know what they got for it.
More than half of the publicly identified donors to that $400 million project won new or expanded federal contracts worth more than $50 billion in the 6 months after they gave. Most of those same companies also had federal enforcement actions against them suspended by the Trump Administration during the same period.
There is no honest word for that other than corruption.
https://t.co/r40hZyargd
This is really stupid, and it’s not getting enough attention.
The Trump administration is pulling a working $368 million ocean monitoring system out of the water, equipment taxpayers already bought, built, and sank into the deep ocean.
And they are doing it right when the oceans are behaving in ways that alarm the scientists who study them.
Record-breaking temperatures.
A system of Atlantic currents that may be lurching toward collapse.
The response?
Yank out the instruments and walk away.
That is not budgeting. That is smashing the gauges while the engine is on fire and calling it efficiency.
For what? The Trump administration dressed it up as a “nimbler approach” and “smart lifecycle management,” which is fancy nonsense for “we shut it off and hoped nobody would ask why.” There is no return-on-investment analysis. They cannot show taxpayers save a dime, because the gear is already paid for and the science it produces protects real money and real lives.
The kicker: the same people killing the monitors want to mine the deep sea for minerals. So they are destroying the only tools that could measure what that mining does. That is not an accident.
That is the point. You cannot see the damage if you break the instruments first.
https://t.co/MzE4AW1QBv