Jeff Ladish backs METR's OpenAI eval report, urges OpenAI to release full logs
JeffLadish · x · 2026-10-14
Amid debate over METR's safety evaluation of OpenAI models, Jeff Ladish argues the METR team did solid work with the access they were given, pushing back on claims that METR is "epistemically bankrupt," while stressing that OpenAI should absolutely release the raw logs for external scrutiny. The exchange highlights the ongoing tension over how much access third-party evaluators get and whether frontier labs should disclose underlying transcripts.
More from Companies & People
- Tesla Roadster launch set for 19:55 CET, a Back to the Future 1955 easter egg — JOBhakdi · 2026-10-14
- NYT poaches The Information's Cory Weinberg to cover Anthropic — ShakeelHashim · 2026-10-14
- Tencent building standalone personal agent app 'Yuanbao Assistant' with long-term memory and task execution — sujingshen · 2026-10-14
- Anthropic Pays £145K for a London Human Assistant Despite Its Own AI Agents — AlexTensor · 2026-10-14
- PostHog to devtools: stop asking 'how to add AI,' ask 'what's the agent-first version' — edwin · 2026-10-14
- NVIDIA joins Earthshot Prize as founding partner, offering finalists free AI compute and DGX Spark — nvidia · 2026-10-14