jackva comments on Predictable updating about AI risk

[jackva]

This feels like “On Caring” but for AI risk rather than moral significance, really great.