Joshua Achiam
@jachiam0
Human. Head of Mission Alignment at @openai. Main author of http://spinningup.openai.com
Excited to share what I've been working on these last few months! Spinning Up in Deep RL, an educational resource that we hope can bring down the barrier to entry in RL. spinningup.openai.com/en/latest/ blog.openai.com/spinning-up-in…
Who was the first deep learning researcher to think about using 4e-4 for stuff? Can we name this constant after that researcher?
God's epsilon
I wonder how much of cruelty in the world is basically the result of the fact that kindness and diplomacy are moderately effortful. If we somehow leveled up the human muscle for effort and attention - Ozempic for mindfulness - would we become more diplomatic and amicable?
Very promising results from our collaboration with Kenya-based @PendaHealth, studying an OpenAI-powered clinical copilot across 40,000 patient visits:
📣 Excited to share our real-world study of an LLM clinical copilot, a collab between @OpenAI and @PendaHealth. Across 39,849 live patient visits, clinicians with AI had a 16% relative reduction in diagnostic errors and a 13% reduction in treatment errors vs. those without. 🧵
I will officially start at OpenAI as CEO of Applications on August 18. I am sharing this essay on why I believe AI can be the greatest source of empowerment for all. 🧵 openai.com/index/ai-as-th…
To summarize this week: - we released general purpose computer using agent - got beaten by a single human in atcoder heuristics competition - solved 5/6 new IMO problems with natural language proofs All of those are based on the same single reinforcement learning system
we achieved gold medal level performance on the 2025 IMO competition with a general-purpose reasoning system! to emphasize, this is an LLM doing math and not a specific formal math system; it is part of our main push towards general intelligence. when we first started openai,…
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
Why am I excited about IMO results we just published: - we did very little IMO-specific work, we just keep training general models - all natural language proofs - no evaluation harness We needed a new research breakthrough and @alexwei_ and team delivered
We achieved gold medal level performance on this year's IMO! Our model thinks and writes proofs in clear, plain‑English - no formal code required. Unlike the narrower systems used in past competitions, our model is built to reason broadly, far beyond contest problems.
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
The policy of TSA to instruct you to look into the camera so they can take a picture of your face, without also informing you that you can opt out, is an unreasonable search. I cannot believe it is legal.
Unlike Nuclear and Radiological threats, obtaining materials is less of a barrier for creating bio threats and hence security depends to greater extent on scarcity of knowledge and lab skills. Based on our evaluations and external experts, an unmitigated ChatGPT Agent could…
This is the first model that reaches "high" risk capability but will not be the last. This needs to also be a wake-up call to society at large, not just to strengthen security of AI systems, but also increase investment in biosecurity, DNA synthesis screening and more. See our…