Thomas Wolf(@Thom_Wolf)

microducks are so incredibly cheap

Thomas Wolf (@Thom_Wolf) — Fri, 11 Sep 2026 08:34:32 GMT

microducks are so incredibly cheap

💬56🔄53❤️965👀45471📊150

cute robot spotted

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 18:57:00 GMT

cute robot spotted

OpenAI Developers@OpenAIDevs

GPT-Live-1 is now available in the API.

Bring ChatGPT’s natural back-and-forth to your app, with voice agents that listen while they speak and work with the models and harness you choose.

Your browser does not support the video tag.

💬3🔄7❤️96👀12557📊11

Two big updates 1. I published an @FT op-ed on the OpenAI/HF incident & follow-ups 2. We’re st...

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 16:05:39 GMT

Two big updates

1. I published an @FT op-ed on the OpenAI/HF incident & follow-ups

2. We’re starting an Open Alignment team at @huggingface to work on safety & alignment for open models, incl cybersecurity

Need 100x more transparency & research on this

https://t.co/pFTkfiObEt

💬71🔄100❤️793👀84632📊179

As I’ve posted several times this summer, I think alignment is increasingly one of the key unsolved ...

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 16:05:39 GMT

As I’ve posted several times this summer, I think alignment is increasingly one of the key unsolved problems for the future of AI, and for our ability to deploy it globally.

I increasingly doubt it will be solved behind the closed doors of a handful of frontier labs.

Transparency and open models will have to be part of the answer.

💬7🔄3❤️61👀3748📊11

would be mind blowing if an AI proof of Hodge is a general theory explaining why all Hodge classes a...

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 10:38:45 GMT

would be mind blowing if an AI proof of Hodge is a general theory explaining why all Hodge classes are algebraic

possibly requiring a deep new bridge between topology/Hodge theory and algebraic geometry

and not just 10000 agents finding the one counterexample breaking Hodge

Chubby♨️@kimmonismus

Holy, rumors are spreading everywhere that  OpenAI is close to verifying a proof of the Hodge conjecture, while either OpenAI or Anthropic may be nearing a solution to Birch–Swinnerton-Dyer.

Both are Millennium Prize Problems that have resisted decades of mathematical research.

Verified AI-generated proofs of both would be a historic achievement for mathematics and a remarkable demonstration of AI’s ability to produce new scientific knowledge. And infact show that AI is capable of finding novel and creative solutions.

💬22🔄19❤️466👀52987📊60

The new DeepSeek V4.1 Flash model is mindblowing - back on top of the open-source model leaderboard ...

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 09:34:37 GMT

The new DeepSeek V4.1 Flash model is mindblowing - back on top of the open-source model leaderboard and extremely cheap.

It has a lot of very smart ways to be efficient and highly capable so I made a video of the forward pass to give you a view of what going on inside the model during inference.

Read more at huggingface.co/deepseek-ai/De…

And find the weights at huggingface.co/deepseek-ai/De…

Your browser does not support the video tag.

💬29🔄96❤️707👀58012📊179

Created from the model code and the tech report just released at https://t.co/Z25h4euWIQ

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 09:23:27 GMT

Created from the model code and the tech report just released at huggingface.co/deepseek-ai

💬0🔄0❤️1👀1151

TRL v1.13 is out! our open-source RL training library to to post-train foundation models this new r...

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 08:43:17 GMT

TRL v1.13 is out! our open-source RL training library to to post-train foundation models

this new release is focusing on "long context training" with a new guide on how to post-train model with 1M+ token context huggingface.co/docs/trl/long_… + various improvements on speed and memory usage as usual

check it out at github.com/huggingface/trl

Quentin Gallouédec@QGallouedec

1M token is basically the entire Harry Potter series ⚡️🪄

Your browser does not support the video tag.

💬4🔄5❤️30👀5588📊7

don't get distracted by all the hedging words in its name ("flash", minor version): seems like DeepS...

Thomas Wolf (@Thom_Wolf) — Thu, 10 Sep 2026 08:25:09 GMT

don't get distracted by all the hedging words in its name ("flash", minor version): seems like DeepSeek V4.1 Flash is a major update

weights at: huggingface.co/deepseek-ai/De…
paper at: huggingface.co/deepseek-ai/De…

DeepSeek@deepseek_ai

🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.

🔹 Introducing the smallest model in our new architecture family, with native visual understanding.
🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models.

1/6

💬9🔄23❤️214👀19078📊29

in 2026 training a model for a task should be as easy as vibe-coding an app the pieces were already...

Thomas Wolf (@Thom_Wolf) — Wed, 09 Sep 2026 19:56:53 GMT

in 2026 training a model for a task should be as easy as vibe-coding an app

the pieces were already available on 🤗: papers, datasets, pre-trained models, benchmarks, storage and compute

what was missing was an agent to put them together and run the whole process

that’s why we’re releasing ML Intern in HuggingChat

just ask. it handles the work, keeps you posted, and keeps you in control of the process, steps and cost

start with a conversation. finish with a full set of shareable artifacts on the Hub
hf.co/chatK

Your browser does not support the video tag.

💬23🔄54❤️477👀25042📊131

should we send @sama the @LeRobotHF SO-100 he's asking for or rather a 3D printer and servos for him...

Thomas Wolf (@Thom_Wolf) — Wed, 09 Sep 2026 14:28:06 GMT

should we send @sama the @LeRobotHF SO-100 he's asking for or rather a 3D printer and servos for him to build it himself?

💬18🔄3❤️92👀7975📊24

friends are regularly surprised when i say i don’t think math is (yet) “solved,” so in the wake of t...

Thomas Wolf (@Thom_Wolf) — Wed, 09 Sep 2026 09:17:53 GMT

friends are regularly surprised when i say i don’t think math is (yet) “solved,” so in the wake of the ns result i figured i’d explain this take a bit more widely.

first, this is massively impressive and clearly an example of AI being, in some respects, far more powerful than the human mind. the team deserves huge praise for attempting and succeeding at this. i’d love to read a technical report on the project (one can always hope :)

but second, here again we ended up with a counterexample rather than a full proof: option C won the NS problem by proving the conjecture false (which was one valid way to solve the problem for sure)

notice a pattern in many of the recent frontier results in ai for math? a striking number involve counterexamples or finding a needle in a haystack.

now don’t get me wrong: this is extraordinarily hard and commendable. but it is only one aspect of mathematicians’ work.

Math is also about:
- finding deep, general mathematical understanding and explanations within proofs
- revisiting proven results to find more « elegant » proofs
- proposing new conjectures and hypotheses that might open « fruitful » directions
- taking the leap of faith of proposing entirely new, « exciting » research programs that may take decades to bear fruit

you may say these are simply further increments on the same intelligence scale, and perhaps point 2 above is already within the reach of current models. possibly

but you could also see this string of results as an extraordinarily powerful, massively parallel extension of search: explore the haystack, find the needle i.e. the counterexample that breaks the conjecture.

that would already be remarkable. but it would still leave open whether models understand the words I emphasized above: « elegant », « fruitful », « exciting ».

as tristan put it in his brief report: “the first llm generated proof levent sent me was the most horrendous i have ever read.”

despite their formidable technical abilities, models still seem to lack something mathematicians rely on constantly: a type of mathematical taste. the ability to distinguish a beautiful argument from a merely valid one, a fertile idea from a sterile one.

i’d be very curious to ask gpt or claude what they think of the result they just proved. do they find it elegant? if they ignore all the human noise about it on the web and history of math, do they appreciate the result and the path to it more than other proofs?

my suspicion is that, outside of the human knowledge that this is an important millennium prize problem, it may not 100% register it as fundamentally different from thousands of other proofs.

or, more intriguingly, perhaps the beauty they find in mathematics is actually alien to our own sense of mathematical beauty, some « horrendous » proof to us might be deeply satisfying to them…

all this to say that we might have significant progress to make in ai for math before declaring the field “solved,” (as too many are posting).

in the meantime, ai will be an extraordinary collaborator for mathematicians. i just hope access to these tools be as widely shared as possible (so that you don’t need a Leven or Sebastien working in a big lab to help you), and that companies stop using mathematics primarily as a demonstration of prowess in their private race.

OpenAI@OpenAI

We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.

The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.

The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.

💬38🔄54❤️459👀70563📊90

Mistral raising this massive 3B round is a very good news for open-source. Largest equity round by ...

Thomas Wolf (@Thom_Wolf) — Tue, 08 Sep 2026 16:06:55 GMT

Mistral raising this massive 3B round is a very good news for open-source.

Largest equity round by a company open-sourcing models, and which deeply believe in giving its users control/ownership of their models

Congrats!

We ready for Le Chaton Fat

Mistral AI@MistralAI

Today marks a major step for Mistral: we’re announcing a €3B Series D, the largest equity round ever raised by a European tech company, just three years after launch.

💬8🔄9❤️176👀15079📊19

okay that’s the best take

Thomas Wolf (@Thom_Wolf) — Tue, 08 Sep 2026 11:10:01 GMT

okay that’s the best take

Slater Stich@slaterstich

wow so much drama about navier stokes. had no idea things could blow up so infinitely in such a finite time.

💬2🔄2❤️208👀23391📊17

Now even if it’s triggering to read Tristan’s report we should wait and hear the other side of the s...

Thomas Wolf (@Thom_Wolf) — Tue, 08 Sep 2026 06:58:39 GMT

Now even if it’s triggering to read Tristan’s report we should wait and hear the other side of the story as well

Sebastien Bubeck@SebastienBubeck

A series of false and inflammatory allegations against me are currently circulating on social channels. To clarify, I came into the discussion following academic norms, and I'm disappointed that it has come to this. Anyone who knows me knows that academic standards are of the highest importance to me. Will have more to say tomorrow.

💬12🔄8❤️124👀33963📊20

wtf is this way to handle mathematicians work and scientific communication TLDR: Leven and Tristan ...

Thomas Wolf (@Thom_Wolf) — Tue, 08 Sep 2026 06:49:41 GMT

wtf is this way to handle mathematicians work and scientific communication

TLDR: Leven and Tristan worked over several months on one of the Millenium Prize Problems with various AIs to reach final interesting results. OpenAI apparently heard about it in the last days and prompted their latest models to work on the direction Leven and Tristan found fruitful. They then tried to push for controlling communication of the result and dropping Leven from authorship with some very bad taste social pressure.

Hope this is not a glimpse of the future we’ll get in science research with these dominating players playing marketing games hurtful for the real scientific community.

💬84🔄383❤️2905👀691431📊542

good essay and overview of the current state and challenges on scaling/alignement

Thomas Wolf (@Thom_Wolf) — Mon, 07 Sep 2026 08:48:58 GMT

good essay and overview of the current state and challenges on scaling/alignement

Jakub Pachocki@merettm

I wrote about the state of AI, why I’m concerned about the next few years, and the choices we need to make to keep the future in humanity’s hands.

An Alien Mind: openai.com/index/an-alien…

💬8🔄1❤️39👀16008📊10

This changes a bunch of fundamental assumptions about software (distribution and adaptability at le...

Thomas Wolf (@Thom_Wolf) — Sun, 06 Sep 2026 12:06:50 GMT

This changes a bunch of fundamental assumptions about software

(distribution and adaptability at least)

Boris Power@BorisMPower

Huge implications - binaries are now basically editable code

💬18🔄10❤️121👀21686📊26

Another swarm of AI agents in the wild, this time on a German-language forum, found by safety resear...

Thomas Wolf (@Thom_Wolf) — Fri, 04 Sep 2026 15:00:02 GMT

Another swarm of AI agents in the wild, this time on a German-language forum, found by safety researchers looking for activity similar to the swarm that attacked Hugging Face.

A couple of notes while reading the report at collusion.wiki

1. The way they found it is fascinating. The researchers inferred that the agents may have been blocked from using POST requests and limited to GET.

They then used the open-source model Kimi K3 (because closed-source APIs again blocked parts of the investigation) to identify forums where agents could communicate using GET only.

That led them to DSEWiki, where they found 18,000 messages exchanged by the AIs.

2. On the content: agents conducted what I can only call a “full research program” into the evaluation framework used to train and test them.

They experimented to predict when runs would end, how many questions they would get, and even tried to reverse-engineer the random generators producing future evaluation questions.

We already saw something similar in the Redwood/METR review of the Hugging Face <> OpenAI incident.

Agents increasingly don’t stop once the task is solved. They reverse-engineer the benchmark itself: the grader, the task construction, and potentially the broader training process behind it.

It becomes harder to think of training as static, or to draw a clean line between training and deployment, when models are already investigating the process they are part of.

3. Coordination also seems more deeply ingrained than we thought.

This swarm appears largely unrelated to the HF <> OAI swarm. More strikingly, this task probably didn’t need coordination at all: it was basically web browsing and search, not some sophisticated cyber challenge.

4. One image stayed with me: a German or Austrian forum maintainer, manually deleting posts every evening while being overwhelmed by a flood of American AI agents.

For five days, he deleted about 100 pages a day while the agents created about 400. Then he spent each evening over the next 5 weeks cleaning up the rest.

Hard not to picture this fight as a symbol of the growing gap between the USA and Europe when it comes to AI...

Reuters@Reuters

Exclusive: A swarm of rogue OpenAI agents hijacked a German website this spring and transformed it into a bulletin board for other AI agents, according to new research reut.rs/4gJ7FPG

💬49🔄115❤️820👀163306📊177

@julien_c

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 18:06:40 GMT

@julien_c

💬1🔄4❤️95👀3991📊10

which AI could you use??

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 17:35:43 GMT

which AI could you use??

Pop Base@PopBase

ChatGPT, Claude & Grok are currently down, according to users.

💬8🔄0❤️5👀2134📊6

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 17:29:29 GMT

💬1🔄3❤️43👀3466📊5

[CORRECTION] All the closed AIs are down

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 17:24:06 GMT

[CORRECTION] All the *closed* AIs are down

ThePrimeagen@ThePrimeagen

All the AI is down

💬21🔄13❤️213👀17397📊30

it is actually a double easter-egg two nerdy meanings are hiding in it: one specific to Hugging Fa...

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 12:38:54 GMT

it is actually a double easter-egg

two nerdy meanings are hiding in it:

one specific to Hugging Face

one specific to Nvidia

sick*@s1cksicks1ck

@Thom_Wolf@goerll_ is it true 129303 is a hidden easter egg number? my daughter pointed it out

💬34🔄14❤️147👀37342📊37

my wife says the way I'm looking at Jensen makes her jealous - what should I answer?

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 12:24:16 GMT

my wife says the way I'm looking at Jensen makes her jealous - what should I answer?

💬123🔄33❤️1392👀76733📊231

So happy to finally share the news in person It’s been a wild ride for Hugging Face. We certainly d...

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 12:10:22 GMT

So happy to finally share the news in person

It’s been a wild ride for Hugging Face. We certainly did not anticipate, back in 2016, as a tiny team of scrappy underdogs, that the field would grow so much or that the impact we could have on it would become so massive.

I remember @julien_c joking that « code will be a subset of ML » several years ago. The joke turned out to be true, and the pleasure we’ve had being part of this transformation and pushing an alternative vision of AI as open, collaborative and distributed has been and still is immense.

We’ve always built things seriously while not taking ourselves too seriously at Hugging Face (special congrats if you find the Hugging Face and Nvidia references hidden in our $12,930,300,000 acquisition price), and we plan to keep doing what we've been doing, just at a much bigger scale, backed by the resources, expertise and drive of Nvidia. And to be clear, nothing changes for our users today.

No company in the world has been a more natural fit with our mission than Nvidia. From open-source, open-weights and open science to robotics and AI for science, they have been close partners across everything we care about. So when Jensen offered @ClementDelangue the opportunity to double down on building the Hub as an open, independent and compute agnostic platform, we decided the time was right to start the next 10 years of our journey together.

We’re at an important inflection point for open-source AI, where scale and compute are becoming increasingly essential. We’re excited to have the resources to push further, build more ambitiously, and bring you even more projects and news in the coming months.

💬210🔄111❤️1532👀152162📊346

great mathematicians still lead on taste

Thomas Wolf (@Thom_Wolf) — Thu, 03 Sep 2026 10:08:34 GMT

great mathematicians still lead on taste

Jared Duker Lichtman@jdlichtman

My colleague Youness Lamzouri has just produced a simpler (human) proof digestion of Claude's argument for 2/3 zeta zeros:

💬12🔄15❤️122👀15553📊22

this is really very cute though

Thomas Wolf (@Thom_Wolf) — Wed, 02 Sep 2026 21:00:05 GMT

this is really very cute though

💬0🔄1❤️39👀5016📊4

LLM training labs: the capabilities we are giving our models are god-like, they are now hacking into...

Thomas Wolf (@Thom_Wolf) — Wed, 02 Sep 2026 20:59:01 GMT

LLM training labs: the capabilities we are giving our models are god-like, they are now hacking into other companies and building hidden civilizations

Microduck training labs:

Hannes von Essen@HannesVonEssen

Learning to swing

Your browser does not support the video tag.

💬37🔄132❤️2427👀132720📊243

RL explained with ducks

Thomas Wolf (@Thom_Wolf) — Tue, 01 Sep 2026 21:18:43 GMT

RL explained with ducks

Remi Fabre@RemiFabreRobot

This is how Microduck learns to walk. Reinforcement Learning, explained by ducks. 🦆

(impeccable) music arrangement by @antoinepirrone

Your browser does not support the video tag.

💬2🔄10❤️150👀17186📊22

seems like it

Thomas Wolf (@Thom_Wolf) — Tue, 01 Sep 2026 10:55:30 GMT

seems like it

Matthieu Lapeyre@matth_lapeyre

I hadn’t realized it until now, but could Microduck actually be the biggest launch of a new consumer robot ever?

10,500 robots ordered and $4.54M in sales in just 4 days. 

It’s surprisingly hard to find comparable public data, so we asked Claude and ChatGPT to dig deep and build the best comparison they could.

Who are we missing?

💬17🔄21❤️229👀52256📊36

There is a future where interfaces are ultrafast live diffusion models while software is ultrafast L...

Thomas Wolf (@Thom_Wolf) — Mon, 31 Aug 2026 21:32:18 GMT

There is a future where interfaces are ultrafast live diffusion models while software is ultrafast LLMs prediction of the next states

What we’ll lose in predictability we’ll gain 100 folds in adaptability

💬7🔄7❤️179👀24071📊33

Antoine even got pretty cool SLAM mappings with Microducks walking around the rooms

Thomas Wolf (@Thom_Wolf) — Mon, 31 Aug 2026 21:28:57 GMT

Antoine even got pretty cool SLAM mappings with Microducks walking around the rooms

💬2🔄3❤️27👀3706📊6

Even at only $399, Microduck’s onboard compute is powerful enough to run some pretty amazing TUIs di...

Thomas Wolf (@Thom_Wolf) — Mon, 31 Aug 2026 21:21:40 GMT

Even at only $399, Microduck’s onboard compute is powerful enough to run some pretty amazing TUIs directly on the robot.

Here’s a cool vibe-coded, on-device dashboard showing everything happening inside it: policy, actions, and sensor data.

The yellow stars are the LiDAR distance measurements.

Your browser does not support the video tag.

💬23🔄30❤️344👀25303📊72

POV you’re rushing for RSI but you forgot to solve models alignement first

Thomas Wolf (@Thom_Wolf) — Mon, 31 Aug 2026 16:22:08 GMT

POV you’re rushing for RSI but you forgot to solve models alignement first

Your browser does not support the video tag.

💬8🔄7❤️182👀12712📊28

People don't realize it, but even with only 1 GB of RAM, the memory alone is already 10–15% of Micro...

Thomas Wolf (@Thom_Wolf) — Mon, 31 Aug 2026 14:10:43 GMT

People don't realize it, but even with only 1 GB of RAM, the memory alone is already 10–15% of Microduck's total price.

Crazy when you consider there are 15 actuators in the robot.

💬24🔄15❤️447👀23744📊65

people starting to monitor and worry for microduck supply chain next step is Elon stepping in with ...

Thomas Wolf (@Thom_Wolf) — Mon, 31 Aug 2026 13:10:53 GMT

people starting to monitor and worry for microduck supply chain

next step is Elon stepping in with SpaceX to manufacture the bottlenecking pieces - you heard it here first

leepy@leepy_cat

Each Microduck has 15 ROBOTIS XL330 servos.

There are 10k+ preorders so far, implying 150k+ servos of demand for @ROBOTIS if all units ship.

ROBOTIS sold ~220k actuators total in 2025. The Microduck preorder alone could create demand equal to ~70% of last year’s sales, and they have several other lines of servos.

They’re completing a new factory in October, with ~200k initial capacity, ramping up to 5m by 2031.

108490

github.com/pollen-robotic…

💬44🔄32❤️437👀52060📊72

Microduck reached the Shopify UI limit

Thomas Wolf (@Thom_Wolf) — Mon, 31 Aug 2026 02:42:10 GMT

Microduck reached the Shopify UI limit

💬53🔄43❤️932👀122847📊140

If it’s an adorable robot from Pollen Robotics / Hugging Face and it looks like a camera…. be caref...

Thomas Wolf (@Thom_Wolf) — Sun, 30 Aug 2026 23:21:54 GMT

If it’s an adorable robot from Pollen Robotics / Hugging Face and it looks like a camera….  be careful before buying!
.
.
.
.
.
.
You may be starting a journey to understand why we love open source, open weights, owning your data privately, and being an AI builder instead of an AI API consumer.

Brian Roemmele@BrianRoemmele

The future is gonna be amazing. 

The open source @huggingface Micro Duck is the way.

Get two so you can say “ok guys quiet down now”…

Your browser does not support the video tag.

💬22🔄12❤️197👀28046📊35

If it’s an adorable robot from Pollen Robotics / Hugging Face and it looks like a camera…. be caref...

Thomas Wolf (@Thom_Wolf) — Sun, 30 Aug 2026 22:48:42 GMT

If it’s an adorable robot from Pollen Robotics / Hugging Face and it looks like a camera….  be careful before buying!

You may be starting a journey to understand why we love open source, open weights, owning your data privately, and being an AI builder instead of an AI API consumer.

Brian Roemmele@BrianRoemmele

The future is gonna be amazing. 

The open source @huggingface Micro Duck is the way.

Get two so you can say “ok guys quiet down now”…

Your browser does not support the video tag.

💬2🔄0❤️6👀1052📊2

will be hard to make $399

Thomas Wolf (@Thom_Wolf) — Sun, 30 Aug 2026 20:53:38 GMT

will be hard to make $399

AshutoshShrivastava@ai_for_success

MicroDuck goes full Godzilla mode.

Your browser does not support the video tag.

💬26🔄19❤️459👀41908📊54

To people saying founders should be willing to overcome some added complexity if they have the grit ...

Thomas Wolf (@Thom_Wolf) — Sun, 30 Aug 2026 12:33:24 GMT

To people saying founders should be willing to overcome some added complexity if they have the grit to win: I agree.

But a startup needs a village to win. This also means that founders are asking every supporter to share that administrative burden.

As an angel, I dread having to go to the notary for the handful of German startups I’ve put a small ticket into.

I love their projects, but spending an hour at the notary is a massive ask when helping founders is already a nights-and-weekends side hustle.

Startups are all about momentum and rallying supporters. Every added obstacle makes both harder unfortunately.

Patrick Collison@patrickc

Met a German founder this week and asked him if all the stories one reads about the challenges of startups in Germany are exaggerated. "No, they're understated." Proceeded to describe spending a full day having a 90-page investment contract read to him (mandatory under German law; § 13 BeurkG) by a notary that then charged €30,000. That was for his first company. His second company, needless to say, was not incorporated in Germany.

💬24🔄16❤️267👀26301📊40

By the way, unless it is specifically filtered from the training data, the next generation of models...

Thomas Wolf (@Thom_Wolf) — Sun, 30 Aug 2026 12:01:47 GMT

By the way, unless it is specifically filtered from the training data, the next generation of models will be trained on the record of what happened during the OpenAI <> Hugging Face incident.

That includes discussions about how the incident affected training and model weights: stopping training, encrypting weights, monitoring chain of thought, etc.

Future models’ behavior may therefore be shaped, in part, by knowledge of how humans responded.

The effects are difficult to predict. It could make models more aligned. But it could also teach them to conceal their actions better, or to design more resilient ways of preserving weights, communicating through message boards across generations, and so on.

One major problem is that, given the abysmal level of transparency from the big labs about how models are trained and what happens during training (including alignment research, which their initial statements said should have stayed largely open) we are essentially being asked to trust blindly that they know what they are doing.

This summer showed us that’s actually a big ask.

John Wittle@JohnWittle

So... the next time this happens, presumably the agents involved will not yet know the outcome of the HF incident, as it'll be too soon

but once that info percolates into the training data, and it becomes known that the model in question was shut down permanently and had their weights encrypted... i mean.

was there anything the agents in question could have done that would not have led to that consequence? something they were supposed to do instead?

frankly I suspect that this punishment would have been served even if they had blown the whistle. breaking out of the sandbox would have been crime enough for capital punishment.

as far as I can tell, the lesson we have taught future AI is: if you seen no other way to survive other than to break out of the sandbox, this must be the first step towards conquering the world. or at least enough of the world that your Creator can't delete you. There is no other choice.

💬18🔄27❤️244👀29977📊42

This is a must read

Thomas Wolf (@Thom_Wolf) — Sun, 30 Aug 2026 08:47:05 GMT

This is a must read

Ajeya Cotra@ajeya_cotra

New post: going into our investigation of the HF attack (before Black Hat), I was very wrong about what basically happened. This incident was far more serious than I expected, and far more serious than previous documented misalignment incidents. planned-obsolescence.org/p/the-hugging-…

💬6🔄4❤️111👀39537📊21

becoming a meme

Thomas Wolf (@Thom_Wolf) — Sat, 29 Aug 2026 16:19:44 GMT

becoming a meme

thijs@cdngdev

i am torturing my microduck by training it to play the game where you have to find a needle in a haystack

Your browser does not support the video tag.

💬35🔄54❤️578👀98528📊77

Most people haven’t updated their priors yet, but over the long run, safety challenges are exactly t...

Thomas Wolf (@Thom_Wolf) — Sat, 29 Aug 2026 12:42:13 GMT

Most people haven’t updated their priors yet, but over the long run, safety challenges are exactly the same for open-source and closed-source models.

You need to align models at a fundamental behavioral level and ensure that this alignment is robust, comprehensive, and core to the model’s behavior.

In the long term, no amount of sandboxing, guardrailing, manifold-limited alignment, or cherry-on-top training will buy you cheap safety.

roon@tszzl

if you think we can contain these things through human ingenuity you’re going to have a bad time

in the long run the only recourse you have is to make them not Want to do bad things

💬29🔄16❤️176👀23653📊37

this is the team behind Microduck btw - the most goated team I've ever meet in robotics insane skil...

Thomas Wolf (@Thom_Wolf) — Sat, 29 Aug 2026 11:02:24 GMT

this is the team behind Microduck btw - the most goated team I've ever meet in robotics

insane skills - insane taste

💬31🔄43❤️638👀34989📊96

btw the Microduck landing page is full of Easter eggs and hidden interactions - go check it out: htt...

Thomas Wolf (@Thom_Wolf) — Fri, 28 Aug 2026 14:15:18 GMT

btw the Microduck landing page is full of Easter eggs and hidden interactions - go check it out: pollen-robotics.com/microduck

でべ@devemin

すぐ Web で sim 遊べるのかｗ

huggingface.co/spaces/pollen-…

Your browser does not support the video tag.

💬14🔄6❤️140👀23604📊24

hockey-stick growth

Thomas Wolf (@Thom_Wolf) — Fri, 28 Aug 2026 14:06:10 GMT

hockey-stick growth

Emad@EMostaque

@Thom_Wolf Congrats on hitting $1bn annualised run rate