Explainers
No explainer here yet. This is the kind of cell that gets one as soon as something in the queue earns it.
Unread — 6
Levels of AGI for Operationalizing Progress on the Path to AGI
A six-level matrix rating AGI by performance depth against generality breadth, with autonomy levels tied to risk.
arXiv 2311.02462 paper · Oct 2025A Definition of AGI
Scores AGI as a well-educated adult’s versatility across Cattell–Horn–Carroll domains: GPT-4 at 27%, GPT-5 at 57%.
arXiv 2510.18212 paper · Mar 2025Measuring AI Ability to Complete Long Software Tasks
Defines the 50%-task-completion time horizon; frontier models reach ~50 minutes, doubling roughly every seven months.
arXiv 2503.14499 paper · Apr 2025An Approach to Technical AGI Safety and Security
Sorts risk into misuse, misalignment, mistakes and structural, then details mitigations for the first two.
arXiv 2504.01849 post · Apr 2025AI as Normal Technology
Argues AI diffuses on the slow timescale of past general-purpose technology, so policy should favour resilience over containment.
Knight First Amendment Institute post · Jun 2025A deep critique of AI 2027’s bad timeline models
Shows six curves fit the same time-horizon data comparably well yet imply timelines from one year to never.
LessWrong