end of an era…it was far ahead of its time. the first and only model with 10 million context. DeepSeek only got to the boldness of…
RT swyxok this might be AI Woodstock 2.0i'd love to do this actually outdoors, in a park with a small sound stage, but dont have contacts.…
ok this might be AI Woodstock 2.0i'd love to do this actually outdoors, in a park with a small sound stage, but dont have contacts. has…
Most of the current comfortable state of the world is dependent on China not doing bonkers escalatory shit which is easily within their capabilities. Americans talk…
We're part of the Amazon Web Services (AWS) AI Builder Lab in New York on Friday, July 24 - a Clash of Agents competition with OpenAI,…
RT Mike LeeRickScott Is RightRick Scott: We shouldn’t go home until the SAVE America Act is PASSED. Cancel August recess if we need to.Americans. Want. Secure.…
You can get really embarrassingly far with bootstrappingthis might be more powerful than naive distillationCan Bölük: Trick Gemini Flash into thinking it has the moves of…
RT HassanIntroducing TogetherLink!An open source CLI to run any open source model inside your favorite coding harness.Run GLM 5.2 directly in Codex and Claude Code.
I wanted to use my Stream Deck to control Codex. Had GPT-5.6 Pro write a project plan based on my specs. Codex implemented it. I didn't…
You don’t have to wait.Merch inspired by research & deployment.Available until sold out.https://openai.com/supply/Daniel White: @thsottiaux When it reaches 10M could we get OpenAI merch ? 😂
Germany has fallenelie: tl;dr: they trained on 10 epochs of a "very light rephrasing" of EVERY gpqa diamond EVAL data :)
> The AI model advances at about [Intelligence]^0.075, or the the 13th root of input intelligence [1,2].no, it ain't gonna be like thatRamez Naam: This AI…
calling it nowFDE -> ODE -> PDEthe existence of Forward Deployed Engineering and now Ordinary Deployed Engineering implies that the next pokemon evolution is Partial Deployed…
RT GoodfireRe This is worth being careful about!Naive setups can degrade interp tools’ validity, but these probes are run on a *frozen copy* of the model…
Of note, Jurgen Schmidhuber might be the only guy here whose agenda is "self-replicating intelligent robots for GERMANY AND JAPAN to compete against THE UNITED STATES…
our models are built to provide the best price for any given task.if you're able to get better price/perf on any workload, would love to hear…
as someone who does a lot of dev community and dev youtube this pace of growth has been one of the greatest mysteries to me because…
Choose a door. Shape your journey.Last episode, Granny’s door won.Vote in the comments—what happens next is up to you.
…oh, rightof course they instantly did that@TheZvi hardest hitGoodfire: Our team spent months developing RLFR, our method which uses probes on a model's internals as reward…
> commissioned a drawing of himselfBroooo… you really think Jürgen would have used anyone else's IP as he dunks on virgins stealing his ideas? he draws…
RT Visual Studio Code📢 In our latest video, learn how to run powerful language models locally in #vscode with the brand new #Ollama extension! Check it…
Donald's Deal Wednesday(PSA: nobody wants to make deals, that's his own kink)*Walter Bloomberg: TRUMP: IRAN CALLED A WHILE AGO; THEY WANT TO MAKE A DEAL
RT alphaXiv70% of AI research isn’t reproducible. With ICML 2026 happening last week, over 6000+ research papers have dropped, but how many are actually real and…
It is a thoughtful framework overall and certainly a good starting point for discussionsDemis Hassabis: http://x.com/i/article/2076946210397552640