LessWrong (Curated & Popular) “The title is reasonable” by Raemon
Sep 21, 2025
The discussion dives into the controversial title of a thought-provoking book and why it's deemed reasonable. The host defends the thesis that AI poses existential risks, highlighting careful argumentation and reasonable dissent. He evaluates counterarguments, discussing the nuanced role of AI 'niceness' and challenges related to mitigation strategies. The importance of bold and clear messaging to shift public policy is emphasized, alongside a call to engage in meaningful debate and explore the complexities surrounding AI risks.
AI Snips
Chapters
Books
Transcript
Episode notes
Authors Are Earnest, Not Exaggerating
- Eliezer and Nate genuinely believe the book title and thesis, not that they exaggerated for attention.
- Raymond finds their confidence understandable given their consistent prior writing and models.
The Claim Is Narrow And Specific
- The book's claim is narrow: if anyone builds a superintelligence via opaque, scaled training, everyone dies.
- It doesn't assert rapid arrival or exclude narrow useful AIs and distinguishes guesses from core claims.
Multiple Strong Disjunctive Arguments
- The book presents multiple disjunctive arguments for why a naively developed superintelligence would be catastrophic.
- Even if you reject some reasons, remaining arguments still imply a horrifying expected outcome.
