AI:AM #3: Zvi on Fable, the Cases For & Against the Ban, + AI for Math, Logistics & More

This was the week the United States government tried to take Fable away from Anthropic. A model that one-boxes on Newcomb's problem, that hides a filter bypass inside an unreadable wall of emojis, that seems to know when it's misbehaving.

It goes beyond the title with direct discussion of like, know, think, including: Welcome to the AI in the AM Weekly Highlights, the moments from a week of live mornings that I most want the people closest to this technology to have.

1. 01A model that one-boxes on Newcomb's problem, that hides a filter bypass inside an unreadable wall of emojis, that seems to know when it's misbehaving
2. 02In part two, I stress test my own reaction against the sharpest people I could reach
3. 03And in part three, because the future did not pause for any of this, the builders, verified mathematics, one-minute medical scans, software that writes itself, and what all of it a