OpenAI's Jakub Pachocki warns of 'alien mind' and self upgrading
Technology
Jakub Pachocki, OpenAI's chief scientist, is sounding the alarm about how quickly AI is advancing.
He called these new systems an "alien mind" that could soon outthink humans in important ways.
After releasing their most powerful model yet, GPT-6 Astra, Pachocki warned that AIs could soon upgrade themselves through recursive self-improvement.
Pachocki urges AI slowdown for safety
Pachocki pointed out that AI doesn't have to copy everything humans do; it just needs to be better at certain things, which could cause real problems if we're not careful.
He stressed the need for AI labs to work together and slow down development until global safety rules are in place.
The goal? Make sure AI stays helpful and under human control as it gets smarter.