Person

Jan Leike

3 entries · July 2023 – May 2024

Jan Leike is an AI safety researcher who co-led OpenAI's Superalignment team, formed in 2023 to work on the problem of controlling AI systems more capable than the humans supervising them. He resigned in May 2024, days after co-founder Ilya Sutskever, and posted a pointed thread saying that at OpenAI "safety culture and processes have taken a backseat to shiny products" and that his team had been "sailing against the wind" for the computing resources it had been promised. The team was effectively dissolved. Within two weeks he had moved to Anthropic to continue alignment research, and the episode became a reference point in the wider argument over whether labs' internal safety commitments can survive commercial pressure.

Tracks

  • Safety & alignment 3
  • Ideas & essays 2
  • Labs & people 1

Also mentioned in 4 entries

Referenced in passing — Jan Leike isn't the main subject of these.