Jerry Tworek says the era of evals is over after seven years at OpenAI
agihouse_org · x · 2026-07-24
Jerry Tworek says the era of evals is ending
AGI House’s Frontier Notes features Jerry Tworek, who spent seven years at OpenAI working on research behind Codex and reinforcement learning for reasoning.
He later became VP of Research and is now CEO of CoreAutoAI, a company building toward automated AI research. The interview is framed around Tworek’s argument that the “era of evals is done,” suggesting the field needs to move beyond benchmark-centric thinking.
Related event: Ex-OpenAI VP Jerry Tworek Says Era of AI Evaluation is Over(2 posts)→
More from AGI Musings
- Sam Altman says OpenAI wants an AI research intern on hundreds of thousands of GPUs by 2026 — morgymcg · 2026-07-24
- Teachers should adapt to LLMs instead of treating them like cheating — NickPassig · 2026-07-24
- AI “loops” may be turning into a broader long-task capability, not a narrow pattern — mattbeane · 2026-07-24
- Code review may disappear by 2027, but production code still needs human eyes — tlakomy · 2026-07-24
- A satirical AI-disempowerment joke turns safety talk into a punchline — jd_pressman · 2026-07-24
- AI incidents may change regulation debates, but not prove a bigger disaster is near — binarybits · 2026-07-24