Skip to content
LessWrong· jenn·Indexed by this site · Originally published· Opinion →· Its place in the scan record →

The author says they are deeply uncertain about whether to urge friends at frontier AI labs to resign

On Friends at Frontier Labs

AI summary

The author is deeply uncertain about whether they should urge friends working at frontier AI labs to resign; although they oppose such work, they mostly help friends periodically review whether to stay or leave. They mention that Eliezer Yudkowsky and others publicly express dissatisfaction with pro-lab statements, and worry this may make employees more defensive and retreat into internal echo chambers. The author also fears their listening-based approach indulges friends in staying; they have never successfully persuaded a friend to resign, and only one person voluntarily left a year later and returned to technical alignment.

Read originallesswrong.com