AI SAFETY & ETHICS

Models finding software vulnerabilities is not the primary source of cybersecurity risk

LessWrong AI

I have tried and failed to write a longer post many times, so here goes a short one with less detail. Dis