In this episode, I talk to Holly Elmore about her advocacy around AI Pause - encouraging governments to pause the development of more and more powerful AI. Topics we discuss include:

• Why advocate specifically for AI pause?
• What costs of AI pause would be worth it?
• What might AI pause look like?
• What are the realistic downsides of AI pause?
• How the Effective Altruism community relates to AI labs.
• The shift in the alignment community from proving things about alignment to messing around with ML models.

Holly's X (twitter) account

PauseAI discord