Reading List

State of AI safety: as capabilities grow and models can monitor other models, issues like adversarial robustness persist and society is still not ready for AI (Boaz Barak/Windows On Theory) from Techmeme RSS feed.