Posts Tagged ‘Superalignment’

Podcast (crealmvault): Play in new window | Download ()

Subscribe: RSS | More

Dwarkish Patel and Leopold Aschenbrener talked about AI for four and a half hours on a podcast. Here's an hour of me boiling it down and straining it through cheese cloth. Topics include aligning the behavior of entities that are much smarter than humans, geopolitical drivers of the AI arms race, and the exponential power needs of training new frontier models.