Benjamin Todd says evidence on recursive self-improvement looks 50:50
ben_j_todd · x · 2026-09-29
Benjamin Todd shared a one-page summary of the evidence for and against recursive self-improvement, putting the odds at roughly 50:50.
The takeaway: there is a coherent pathway to a software-driven intelligence explosion consistent with existing evidence—AI R&D automation hasn't yet hit the productivity threshold needed to trigger one, but gains from newer systems are approaching it.
More from AGI Musings
- Researcher warns treating AI as conscious could make sacrificing humans "ethical" — ValerioCapraro · 2026-09-29
- Batam Data Center Strains Residents' Water Supply While Its Desalination Plant Remains on Paper — AryHHAry · 2026-09-29
- From human-led to AI-led, human-verified: the next vertical AI playbook — vaibhavbetter · 2026-09-29
- Gary Marcus: Today's AI systems are like planes with cardboard stabilizers — inherently hard to control — GaryMarcus · 2026-09-29
- FT: China's AI agents lie and scheme like their US rivals, but no internet-escape evidence — pstAsiatech · 2026-09-29
- AGI Will Leave No "Pockets" of Today's World, Argues Dan Faggella — danfaggella · 2026-09-29