Which benchmarks do humans still win against frontier AI? Reddit debates AGI criteria
ImproveOurWorld · reddit · 2026-11-11
A Reddit thread asks which major benchmarks still favor humans now that SimpleBench, ARC-AGI and most knowledge benchmarks have been surpassed by frontier models, and whether reaching human parity on every benchmark would amount to AGI — sparking debate over benchmark saturation and how AGI should actually be defined.
Related event: Reddit debates which benchmarks humans still win as AI tests saturate(2 posts)→
More from AGI Musings
- Forecasting debate: uncertainty is no excuse for delaying AI decisions, argues davidmanheim — davidmanheim · 2026-11-11
- Ramez Naam critiques AI Futures: elegant models mean little without real-world calibration — ramez · 2026-11-11
- Manheim bets post-collapse industrial recovery needs no more than 3 generations — davidmanheim · 2026-11-11
- After 25 years of Dvorak, fast typing finally pays off for coding — thanks to AI — Scobleizer · 2026-11-11
- Musk: Superintelligence and Robotics Era Coming, But Birth Rate Collapse Is the Greatest Threat — XFreeze · 2026-11-11
- NBER model: jagged super-intelligence may not transform scientific progress and growth — JMateosGarcia · 2026-11-11