Reflections on TA-ing Harvard's first AI safety course

(lesswrong.com) by sebg | Jan 15, 2026 | 0 comments on HN