Eliezer Yudkowsky: Dangers of AI and the End of Human Civilization | Lex Fridman Podcast #368

Lex Fridman · 197:51

Eliezer Yudkowsky tells Lex Fridman that current large models are already giant inscrutable matrices we cannot inspect, that alignment of something smarter than its builders is a first-try problem rather than a resear...

Read the full summary on tuber

Redirecting...