You Can Do AI Safety From Second Place, Say Researchers on Open Models
joshalbrecht · x · 2026-10-02
A short exchange on AI safety research. The quoted post argues that safety work can absolutely be done from second place—or even off the leaderboard—pointing out that academics, startups, and nonprofits are already doing great AI safety research on open models. The original poster adds the empirical reverse: you can be in first place and still not do any real safety research.
More from Safety
- Microsoft's official X account appears hijacked, pushes Clippy crypto scam — tomwarren · 2026-10-02
- "Trust No AI": Security Researcher Revives 2020 "Assume Bias" Playbook for AI Safety — wunderwuzzi23 · 2026-10-02
- DeepMind veteran Victoria Krakovna: gradual AI disempowerment likelier than extinction, calls for coordinated slowdown — vkrakovna · 2026-10-02
- AVERI blind-benchmarks Gemini inside OpenMined secure enclave, unseen by Google — iamtrask · 2026-10-02
- No Hat 2026 keynote: When Every Attacker Can Have a Research Team — WeldPond · 2026-10-02
- WIRED: AI companies' self-regulation accord is safety theater — nordicinst · 2026-10-02