Ethan Perez Making AI safe through debate

Jeremie Harris and Ethan Perez, a researcher who has worked for Google, Facebook, and MILA, talk superintelligent AI systems, GPT3 language models, and the possibility of debate. 0:00 Intro 1:37 Ethans background 7:26 Issues with IDA 12:09 Gödels incompleteness theorem 15:01 The role of IDA today 26:45 The capabilities of IDA and GPT3 29:54 Systems and debate processes 41:21 Making AGI work 49:33 Our ability to control these systems 51:41 Wrapup