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