AI evals change too fast for their own builders, leaving room for new evaluators, Arena founder says
MattPerault · x · 2026-11-17
Matt Perault argues that even people building AI evaluations struggle to keep up with how quickly the technology changes — a reason for policymakers to leave room for new evaluators and better methods to emerge. He cites the firsthand experience of @mlangelopoulos, who builds @Arena: more competition can mean better information for everyone making decisions about AI.
Related event: LMArena Founder: Policy Must Leave Room for New AI Evaluations(2 posts)→
More from Safety
- Dario Amodei warns open-source models could lead to "a very dangerous place" — examachine · 2026-11-17
- New paper lays out a research agenda for AI's impact on mental health — jacyanthis · 2026-11-17
- Paul Christiano: 1-in-3 chance of government-toppling superintelligence within a year — Dwarkesh Patel · 2026-11-17
- GDM safety report: Argon evades its monitor half the time in control evals, deemed not concerning — connoraxiotes · 2026-11-17
- AI platform security checklist: default gateway keys, open inference endpoints, non-expiring tokens — mclynd · 2026-11-17
- US shuts Malaysia chip loophole: license now needed to reexport advanced chips to five countries — SumitGup · 2026-11-17