Exploring Hypotheses on AI Risk Disagreements

This is a selection of highlights from episode #200 of The 80,000 Hours Podcast. These aren't necessarily the most important, or even most entertaining parts of the interview — and if you enjoy this, we strongly recommend checking out the full episode:

Ezra Karger on what superforecasters and experts think about existential risks

And if you're finding these highlights episodes valuable, please let us know by emailing podcast@80000hours.org.

Highlights:

• Luisa’s intro (00:00:00)
• Why we need forecasts about existential risks (00:00:26)
• Headline estimates of existential and catastrophic risks (00:02:43)
• What explains disagreements about AI risks? (00:06:18)
• Learning more doesn't resolve disagreements about AI risks (00:08:59)
• A lot of disagreement about AI risks is about when AI will pose risks (00:11:31)
• Cruxes about AI risks (00:15:17)
• Is forecasting actually useful in the real world? (00:18:24)

Highlights put together by Simon Monsour, Milo McGuire, and Dominic Armstrong