StoryFirst seen
Top mathematicians are accusing the lab of opacity and secrecy, while open-source builders turn the released problems into public training environments, keeping a fierce debate roa
Everything it touches
6 names 9 voices 11 source posts
The story’s neighborhood
Explore the graph, or choose a connection below.
Follow the evidence
This brief comes from Tech Twitter’s edited headlines. The posts below are the sources grouped with the story; statements in them belong to their authors.
“On Tuesday night, I cried, and it woke my daughter. [OpenAI] annihilated my field of research. They wanted to prove they could massacre us.” these are the words of Fields Medalist, Hugo Duminil-Copin, who was working on 16 problems OpenAI now claims it has solved. and he says some mathematicians may now keep their research problems to themselves rather than share them publicly. AI might actually make mathematics less open.Read original on X
I notice that when someone says 'super intelligence' instead of AI, this is a Fnord. It creates active discomfort, and pressure to avoid reading their words.Read original on X
Satya Nadella says we should assume all AI models are "compromised." Microsoft’s CEO calls for putting an "emergency brake" on AIRead original on X
The CEO of Microsoft just said the quiet part out loud. Nobody on earth, including the labs that build them, can explain why AI models do what they do. And we already gave them access to everything. Your company's AI agents can read the codebase, the customer database, the email. Some can move money. Satya's memo admits we can't trace any of their behavior back to the training data or the weights that caused it. With normal software, when something breaks, you find the line of code. With these systems, there is no line of code to find. Then he says the sentence that should stop you cold. We're deploying them anyway, with access to our most sensitive data, able to take mission-critical actions on our behalf. Every other AI leader dances around this. Satya wrote it down and then did…Read original on X
We turned @OpenAI's math problems into open-source RL environments on @huggingface! Early and rough (the verifier only accepts the exact original formalization, so an equivalent proof can still score 0), but this is what it looks like when a research release becomes executable training infrastructure for everyone. This is an exciting direction imo: turning open research into open executable environments that anyone can build on to train better open models!Read original on X
I think Terrence Tao is directionally correct, we shouldn’t settle for a super intelligence that is opaque. A key signal of intelligence is communication. We should strive to make products and models that build understanding with humans instead of disintermediating them.Read original on X
this mathematician won the 2026 Fields Medal, then stepped away from academia to join OpenAI’s safety team. what’s fascinating is how balanced his view is on AI. his story is WILD: >be Jacob Tsimerman >spend 13 years on a math problem >win the 2026 Fields Medal >an LLM finds a mistake in one of your proofs >you fix it, but the LLM couldn't >“I think my part is going to be very small” >realize where things are going >imagine AI solving decade-long math problems in weeks or even days >say mathematics will advance too fast for humans to keep up >AI will generate conjectures, prove theorems, and invent new definitions on its own >people argue humans will still be needed to develop new mathematical theories >“Why won't AIs be able to do this as well?” >point out that even your own theorems…Read original on X
From Le Monde. OpenAI actually sat on this release for a long time, and formed the independent Advisory Group on Mathematics and Artificial Intelligence to 'serve as a bridge to the mathematical community'. I think that attempt was sincere, but appears to have had mixed results.Read original on X
OpenAI’s AI models are solving math problems that could take humans years, thanks to what one mathematician calls brute force and stamina. Read more:Read original on X
Day 7/ We have increased default context length for GPT-6.1 Sol to 500k when used through Codex & Work, up from 272k or an ~ 80% increase. We will go higher in the future, but it helps reduce occurrence of compactions and with tasks that require keeping a lot of context and nuance over periods of time. It was carefully tested and is the new sweet spot for most of our users. You can still override it if you want to be higher our lower, with maximum context length of GPT-6.1 Sol of 1M.Read original on X
The results of this poll has gotten me confused. I definitely thought it would sway much more favorably in the 2X direction across Pro 100/200/500 at the cost of bringing back 5h limits in order to smoothen the demand curve.Read original on X
Mathematicians push back as OpenAI's claimed proofs go open source
Mathematicians revolt as OpenAI's Grok Bot claims open problems
Mathematicians revolt as OpenAI's Grok Bot claims open problems
Mathematicians revolt as OpenAI's Grok Bot claims open problems
Mathematicians revolt as Grok Bot claims to solve open problems
Mathematicians revolt as Grok Bot claims to solve open problems
Mathematicians push back as OpenAI's claimed proofs go open source
Mathematicians push back as OpenAI's claimed proofs go open source
OpenAI's math claims draw pushback as problems go open source
OpenAI's math claims draw pushback as problems go open source
Shared connections
Researchers including Scott Aaronson describe colleagues dropping their work to unpack AI solutions to problems that stood open for decades.
Mathematicians including former OpenAI researcher Scott Aaronson say colleagues across the field are dropping their own work to understand the solutions, a signal of how quickly AI
PayPal and Epic Games warned Justice Department officials that a settlement could leave Apple's restrictions on rival apps and devices largely in place.
The lab says it hit its frontier-model goal just over a year after founding, then released the weights openly.