Noam Brown

@MillionInt @shyamalanadkat I believe that one day we could, but not clear if we ever should.

@kevinroose Why do you think coders are generally okay with AI-generated code, but writers seem to generally not be okay with AI-generated writing? Assuming both are reviewed by humans.

The recipe behind today’s frontier reasoning models is surprisingly similar to AlphaGo: 1) Imitate large amounts of human data 2) Scale inference compute to reason better (back then it was Monte Car

@anko_979 This is a good point. I was debating whether to address this in the article, but decided to keep its scope focused on the US and save the international perspective for a future article.

@NickEMoran Regardless of whether you think it's a good outcome or bad outcome, I hope we can agree that it is ceding power.

This was written with a lot of help from ChatGPT-5.4. I found the model to be quite good at writing, very good at research/explaining, and extremely good at providing feedback.

Blog post is here: https://t.co/C2xkXRLd8O

GPT-5.4 is a big step up in computer use and economically valuable tasks (e.g., GDPval). We see no wall, and expect AI capabilities to continue to increase dramatically this year. https://t.co/zx3rgGr

tl;dr: @OpenAI will not be deploying to the NSA or other DoW intelligence agencies for now, so that there's time to address potential surveillance loopholes through the democratic process. Over the w

@typedfemale At @OpenAI some researchers like @boazbaraktcs spend part of their time working with the policy team on these sorts of decisions. It's valuable to have researchers that understand the tec

@john__allard @natseckatrina Yeah I’m digging into this more because I agree

@john__allard @natseckatrina I followed up with Katrina about this and the answer is yes. The contract does not allow NSA Title 50 work.

@Lang__Leon @OpenAI I suggest asking @natseckatrina in her AMA, since she's the most informed on the OpenAI side: https://t.co/uyqHxmzUT9

@letithappen_25 @OpenAI Ugh good point

For those following the DoW AI drama, I highly recommend reading this post explaining how @OpenAI approached the negotiations with the DoW.

@deredleritt3r Why not xhigh?

Blog post with in-depth analysis is here: https://t.co/I5Psblxds2

tl;dr SWE-bench Verified is heavily contaminated for all frontier models, and many of the problems are also broken. Time to move on to harder, uncontaminated coding evals. https://t.co/FipvC9oMbs

@HarvardMath AI isn’t replacing mathematicians today, but it is changing mathematics: https://t.co/FGeZDMFP2n

Perhaps a 🌶️ take but I think the criticisms of @GoogleDeepMind's release are missing the point, and the real problem is that AI labs and safety orgs need to adapt to a world where intelligence is a