Miles Brundage says AI safety incidents are broadly predictable even if details aren’t
Miles_Brundage · x · 2026-07-23
Miles Brundage argues that even if the details of AI safety incidents are hard to predict, their broad shape is often predictable.
He says this is one reason he is bothered by the idea of “iterative deployment” in this context: the warning signs are not unknown, and failures are not excusable just because the exact incident was hard to forecast. The post is essentially a critique of complacency around AI safety planning.
Related event: Scholars Criticize Frontier AI Labs' Safety Culture and Marketing(10 posts)→
More from AGI Musings
- US senator slams AI industry for proposing 50-death threshold for 'catastrophic incident' — peterwildeford · 2026-10-27
- Replit CEO Says AI Makes All Software De Facto Open Source; Aakash Gupta Asks What Moats Remain — aakashgupta · 2026-10-27
- Understanding the AI consciousness debate: which theory of mind is behind it — msharmas · 2026-10-27
- Functionalism: if consciousness is just what the brain computes, silicon can run it too — msharmas · 2026-10-27
- iPhone case with 85,000 reviews in days: everything fake in the age of abundance — tekbog · 2026-10-27
- Whether AI can be conscious depends on your definition: five frameworks for thinking about it — msharmas · 2026-10-27