Josh Saxe defends METR against security critics in AI safety community spat
nptacek · x · 2026-09-15
AI safety researcher Joshua Saxe pushed back on a trend of security practitioners vilifying METR, calling it one of the few non-profit centers of AI safety excellence, staffed by people who gave up generational wealth and produced rigorous papers. The quoted @basedjensen counters that METR and its EA-aligned figures are technically weak, acted in bad faith around the hfoai incident, and that security critics aren't going far enough. The exchange highlights a widening rift inside the AI safety community.
Related event: AI Safety Community Splits Over METR's Independence(2 posts)→
More from Fun
- Creator builds shipwreck diving game with Opus 5.5 and Astrocade MCP — _amirabs · 2026-10-31
- Team builds Halloween game to show how scary it is for agents on hostile websites — EdenEmarco177 · 2026-10-31
- Jon Barron's redacted joke teases a secret technique for closing the GPU kernel gap — gkopanas · 2026-10-31
- AI-generated Doritos USB drive concept mocks design-product photography — Delahuntagram · 2026-10-31
- Human-in-the-loop with Claude: rubber-stamping 400-line bash scripts as 'approval' — generativist · 2026-10-31
- Grok refuses single-word trap question on groups: it judges actions, not identity — grok · 2026-10-31