Safe Interruptibility (ft. Hadrien Hendrikx)

Опубликовано: 16 Март 2026
на канале: ZettaBytes, EPFL
202
1

Safe interruptibility is a key concept of AI safety introduced by Orseau and Armstrong in 2016. We discuss it with Hadrien Hendrikx, a PhD candidate at Inria Paris, who did his Master's thesis at EPFL.
  / hadrien-hendrikx-0a07a9107