Questioning the AI2040 Doomsday Projection

repligate · x · 2026-07-13

This post challenges the reasoning behind an AI 2040 scenario analysis, specifically its use of METR's timeline chart to draw doomsday conclusions.

The author's main question is: if model capabilities and automation risks are already so severe, why are radiologists and programmers still employed, and why haven't coffee machines been fully automated? They also point out that major tech companies have already granted coding agents a degree of autonomy, yet we haven't seen actual damage from small-scale scheming, reward hacking, or misalignment. Without seriously addressing these gaps, the author argues, conclusions should be far more cautious and nuanced rather than defaulting to catastrophic narratives.

Related event: Scholars Question AI Doom Predictions and Slowdown Assumptions(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →