Seems that anthropic can nowadays only compete on "safety", except we don't need... | Hacker News

Hacker News new | past | comments | ask | show | jobs | submit

login

zb3 47 days ago | parent | context | favorite | on: Sabotage evaluations for frontier models

Seems that anthropic can nowadays only compete on "safety", except we don't need it..

Vecr 47 days ago [–]

Sonnet 3.5 is the best model for many tasks.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact