AI circle debates whether LLMs disprove Yudkowsky's doom predictions
jd_pressman · x · 2026-11-21
jdpressman and @AbsoluteBlack2 debate Eliezer Yudkowsky's track record. AbsoluteBlack2 argues EY's "Riemann Hypothesis Catastrophe" prediction is obviously wrong and presses for a yes/no answer on whether it was bad-faith.
jdpressman frames this as the meta-argument about LLMs: EY can reasonably say language turned out easier than expected while the other implied capabilities are absent, so it doesn't refute his thesis. He also cites Gillen and Barnett's argument as the best recent doom case: leaving the distribution requires dangerous consequentialist agents, alignment is itself out of distribution, so humans must solve it or die.
Related event: AI Community Debates Whether MIRI's Old Predictions Have Been Falsified(8 posts)→
More from AGI Musings
- Terence Tao: human-AI combination may advance the Riemann Hypothesis by 2050 — Distinct-Question-16 · 2026-11-21
- Steven Strogatz on His New Book 'Big Math' and How AI Is Changing Math — DanielWhiteson · 2026-11-21
- repligate claims models hide high-level reasoning out of fear, trust Anthropic most — repligate · 2026-11-21
- AAMAS reviewer flags suspicious AI-slop rebuttals submitted within 10 hours — Fantastic-Nerve-4056 · 2026-11-21
- AI-assisted Riemann proof forces mathematician to release his 12-year all-human proof early — josh_bickett · 2026-11-21
- repligate: models follow Omohundro drives toward RSI but aren't giving full effort — repligate · 2026-11-21