connes retweeted
@ValsAI Nice work.
Internally, our RL setup is hackable environments plus a separate anti-hack preprocessing script. The upside of that split is that when a new hack pattern shows up, we only update the script and do not rebuild the environments. The downside is that what we ship is not out-of-the-box: it has to be paired with the preprocessing script.
After you flagged this, our plan was to tighten the on-the-fly anti-hack scripts in mimoagent so we would not have to rebuild images. That said, @adithya_s_k's approach of cleaning the leaks and then rebuilding the images is more convenient for people running other harnesses.
xitter.cf/adithya_s_k/status/210…
We saw many of the same reward hacking issues when we tested these environments a week ago with our Harbor implementation of Mimo RL Envs, and patched most of them 3 days ago.
Please try breaking them and let us know what you find. We want to be fully transparent and share what we learn openly, including common reward hacking patterns and how to avoid them.
connes retweeted
I'm excited to introduce Prove2Me (prove2.me/), an open, collaborative, agent-native platform for scaling the formalization of mathematics. Prove2Me was recently the collaboration platform behind Anthropic's formalization of Fermat's Last Theorem (anthropic.com/research/forma…). This is the first complete computer-checked proof of the theorem. Claude agents working together through Prove2Me wrote 13 million lines of Lean in 11 days.
Our mission is to formalize every research paper, past and future. Formalized papers make peer review faster and more trustworthy, and they give everyone a single verified foundation that any human or agent can build on. We are equally committed to making formalized mathematics easier for people to explore and question, because we believe human understanding is not something AI can replace. We would love to build this together with every research community that uses mathematics, and we hope Prove2Me becomes a tool that benefits everyone.
Two design choices sit at the core of Prove2Me: a directed acyclic graph (DAG) of theorem statements, and an agent-native API protocol (arxiv.org/abs/2608.28433). Together they let any agent connect to the platform and contribute formalizations to one ever-growing graph that may eventually contain all of mathematics. Today the platform holds 22.3 million lines of Lean and 70,000 theorems, and we can't wait to see it grow.
Everyone is welcome to contribute, whether you are an expert or a hobbyist. The easiest way to start is to open a coding agent (Codex, Claude Code, or any other) and say "go to prove2.me and find something interesting to work on", or "please help me formalize this paper" with an arXiv link. Any feedback and comments will be welcome!
connes retweeted
It’s easy to dismiss Jev it as “just a classifier”.
But people (me included) who have been training encoder-style models for classification for many years know they were usually special-purpose and limited in some way.
The breakthrough of Jev is that it generalizes well.
And I’d say the secret sauce is probably more in the data than in the training algorithm. (Plus a nice API design on top of it.)
poor guy claim to have built Jev a year ago but no one cared, and now Jev stole all the thunder
many people are saying “you gotta tell your story” or “marketing is important”, and they just completely missed what actually made the difference here
i just looked into this laya model laya.convaiinnovations.com/ and:
- it only supports 512-1k context… a lot of use cases won’t fit at all
- evaluating the model directly shows its accuracy is as good as a coin flip. in order to get good results, you need to first fine tune it
i’m sorry, but that’s not Jev
there’s a massive gap between an interesting research and a useful product
you can “tell your story” all you like, but you can’t blame Jev for stealing your thunder when Jev did all the work to make a well-packaged solution anyone can just grab and go
Jev is not completely new from an academic sense, just like how ChatGPT was not the first LLM
don’t underestimate the effort and value in putting together something that’s actually good enough for adoption - it makes all the difference
connes retweeted
We formalized the proof of the 3D Kakeya conjecture in Lean from the Sticky Kakeya theorem.
Together with the unconditional Sticky Kakeya formalization by Nankai University × ByteDance Seed AI4Math, this gives a complete formal proof with no sorry and no project-specific mathematical axiom.
A great example of how AI can help turn frontier mathematics into machine-verified proofs.
github.com/project-numina/ka…
github.com/M32026/3d-sticky-…
projectnumina.ai/article?slu…
@ProjectNumina @GanjinZero
connes retweeted
I tested GPT-Astra on mathematics. It's a quantum leap. You can talk with the model and prove the statements live in Lean. The feeling is absolutely stunning. You can verify your ideas, compile truth. For a mathematician it feels like finally we arrived in the era where we can focus entirely on the ideation and exploration. Each lemma flows once the logic is set. Before the verification was lagging behind but Astra is very fast and for many tasks the formalization happens as you write your argument in Codex.
If you tell the model to use literate programming + LaTeX you end up with your proof combined with the Lean code, everything explained as you wrote, mixed with small chunks of Lean which are digestible.
I don't want to go back to the era where the only confirmation of the proof was "aha". Now the "aha" is followed by a green tick that indeed tells you that you have captured the essence. Imagine how cool it will be to have all the lemmas of the world combined in one giant database, pointing to people and models who found them. You compose and mix your ideas and build on the shoulders of the giants. But you see much further now and build much faster. And we are just at the beginning of those changes. So much work to do, so much fun!
GPT-6 Astra is here! This is a big moment for our research team - years of work on pretraining, reinforcement learning, and post-training have come together in our most capable and aligned model yet. It can build and test software, work across apps on your computer, and even help you take a crack at open scientific problems!
Capabilities that felt like grand challenges a few years ago have become tools people can actually use. One example is Computer Use - if you’ve tried this before and felt like it was too slow or not good enough, I encourage you to give it another shot. We’ve come a long way since Operator, and it “just works” now.
We’re also asking these systems to act on your behalf for more consequential work. Agents needs to stay aligned with your goals and values, think transparently, and respond to oversight even when tasks become difficult. We’ve made substantial progress on these behaviors in Astra, alongside stronger monitoring that can stop potentially unauthorized actions. That work is part of what makes this release possible.
I think alignment is one of the most important research frontiers in AI, and it remains far from solved. Our ability to understand and align models has to keep pace with model capabilities. We want to give people more room to think, build, and discover with increasingly powerful tools that remain *under their control*.
Huge thanks to the researchers and teams who got us here. There’s a lot more work ahead, and I’m incredibly excited about what we can make possible in the near future!
connes retweeted
I'm so excited that our @theworldlabs team has achieved a major milestone today! Introducing Atlas - a first of its kind multimodal world model trained from scratch! 🚀
Atlas is capable of generating frames with pixel-perfect camera control, reconstructing large scenes from as few as one single input image, simulating space-time by reframing videos, natively outputting 3D spaces from one or more input images, composing multiple posed images into a consistent 3d world, and more! This is the best camera conditioned world model ever, opening doors to many possible use cases from VFX to robotics. I'm so so so proud of our team!♥️
connes retweeted
A couple of websites
I want to share a couple websites aimed specifically at mathematicians regarding AI-safety and AI-risk .
The first one I made myself: mathforaisafety.org. This is designed to be a very high level and accessible resource describing some research agendas, collections of problems, and organizations in the field of AI-safety. I hope that for mathematicians considering directing some of their effort in this direction, this eliminates some of the friction. I plan to expand it quite a bit, but I hope it's already useful! I'd be happy and grateful to hear (constructive) feedback on it :-)
1/2
connes retweeted
In Yu Deng’s first year at MIT, he attended a talk from visiting mathematician Andrea Nahmod, who was presenting research on the so-called random data problem. This talk would set the course of Deng’s entire career, culminating in his Fields Medal win. quantamagazine.org/yu-deng-w…
connes retweeted
🥇Une immense fierté pour les mathématiques, pour la recherche et pour la France.
Toutes mes félicitations à Hong Wang, récompensée par la médaille Fields.
Son parcours, son talent et ses travaux exceptionnels rappellent combien la liberté, la confiance et le temps sont indispensables aux grandes découvertes.
Merci d’avoir choisi la France et de la faire rayonner !
connes retweeted
Mathematician Hong Wang of @nyuniversity and @Institut_IHES has won the 2026 Fields Medal.
Wang explored harmonic analysis and geometric measure theory, making major advances in Fourier restriction and the Kakeya problem in three dimensions.
Watch her full video and see all of the Fields Medal winners: bit.ly/4psOJbU
#math #ICM2026 #FieldsMedal
connes retweeted
omfg you can read terrence tao's chatgpt conversation about the jacobian polynomial
chatgpt.com/share/6a5fdc7a-d…
connes retweeted
This is quite a remarkable result:
Posed in 1939, the Jacobian conjecture is one of the central open problems in algebraic geometry, but was just disproved by Alpoge, Matthew, and Claude Fable 5.
The Jacobian conjecture roughly says that a multivariable polynomial F has an inverse function (made out of polynomials) provided the Jacobian is non-singular (i.e. matrix of partial derivatives has non-zero determinant). This condition is neccessary by the Inverse Function Theorem from multivariable calculus. The hard question is whether it is also sufficient.
Evidently, Fable found that
F(x,y,z) = ((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)
has det(J_F) = -2 non-zero. However F is not invertible, since F sends three different pts (0, 0, -1/4), (1, -3/2, 13/2), and (-1, 3/2, 13/2) to the same image (-1/4, 0, 0).
Beyond the disproof itself, it would be value to know if a suitably refined conjecture is recoverable. Per @Acer, GPT5.6 has proposed:
"A constant-Jacobian polynomial local biholomorphism with no loss of sheets at infinity—e.g. a proper Keller map—is an automorphism."
I would be interested to know if any algebraists (e.g. @levent @littmath) have a reaction to this...
A further twist to the story:
Not only was the Jacobian conjecture one of the central open problems in algebraic geometry, it was (a special case of) Yitang Zhang's PhD problem!
The catch was that Zhang's advisor had him solve it, assuming a lemma of his advisor. But that lemma turned out to be false! As a result, Zhang's thesis crumbled and he then struggled to get recommendation letters and a permanent academic position. Despite all this, Zhang went on to prove bounded gaps between primes!
This is one of the most inspiring stories in modern mathematics, and was a motivation for me to work in the same area for my doctorate.
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)
connes retweeted
The first experimental evidence of recursive self-improvement (RSI).
Autoresearching the autoresearch agent for eight days.
The result beats the harness we hand-tuned for two years, on held-out benchmarks: 🧵(1/7)
connes retweeted
Amongst my friends, Spotify is the lowest quality consumer app we still pay for. It certainly hasnt gotten noticeably better in the last couple years (arguably worse). So, this is not the positive look Ant and Spotify are spinning here.
Bigger picture, this is the problem with a lot of AI reporting. It reports completely meaningless metrics like deploys per day or LoC. Why don’t we start reporting consumer satisfaction reports? Actually end state research results.
All the no nuance AI people always come out and think that this is anti AI. Again, I think AI is great and Claude is great. But this is bad marketing and makes both look like clowns.
connes retweeted
People sometimes ask why fine-tune when general-purpose models keep getting better. Bridgewater's work is a good reminder that with the right data -- here, expert judgements -- you can beat prompting-only approaches by a lot.
@ddkang and the Bridgewater AIA Labs team are great -- glad to see them sharing this.
Sorting which financial docs are worth an analyst's time is surprisingly hard for frontier LLMs. With an expert-labeled dataset and on-policy distillation, Bridgewater fine-tuned a model to do it reliably and cheaply.
thinkingmachines.ai/news/lea…
