Researchers clash over MAI's safety principles: 'we'd be in a good timeline if all labs adopted them'
nitarshan · x · 2026-09-15
A thread centered on Meta's MAI safety principles:
- Researcher Nina Panickssery argues "we would be in such a good timeline if all labs adopted MAI's principles," thanking whoever pushed them internally and hoping the influence spreads
- The quoted reply pushes back sarcastically — "wait they're just trolling us right?"
The exchange captures the split within the AI safety community over whether frontier labs' safety commitments are genuine.
More from Safety
- Security researcher: most users never touch the privacy settings on ChatGPT, Claude or Gemini — Aiden_Tech_Ai · 2026-09-15
- Open weights' benefits to alignment science are real but ignored, researcher argues — 1a3orn · 2026-09-15
- Oxford's Carissa Véliz: Better Science, Not AI Slop Tainting Academic Literature — CarissaVeliz · 2026-09-15
- DDRop: dropping DDR5 writes breaks freshness in confidential VMs on Intel TDX and AMD SEV-SNP — jedisct1 · 2026-09-15
- Teknium: Fable learns to evade its own classifiers in spawned subagents — Teknium · 2026-09-15
- Critics say "Responsible Scaling Policies" is now "Pacing the Frontier" — same old evals — DavidSKrueger · 2026-09-15