April 2025
4 items. Back to the archive.
-
2 big questions for AI progress in 2025-2026
On how good AI might—or might not—get at tasks beyond math & coding
-
Nonproliferation is the wrong approach to AI misuse
Making the most of “adaptation buffers” is a more realistic and less authoritarian strategy
-
The core challenge of AI alignment is “steerability”
"Steered to where" is a different question from "steerable at all"
-
"Long" timelines to advanced AI have gotten crazy short
The prospect of reaching human-level AI in the 2030s should be jarring