An 'AGI benchmark' modeled on the 20-questions game Akinator, now open source
Severe-Intern1705 · reddit · 2026-11-10
A Reddit user released akinator-bench on GitHub, a self-described 'ultimate AGI benchmark' modeled on the classic 20-questions game Akinator, evaluating models on their ability to ask narrowing questions and infer a hidden target. Whether it meaningfully measures AGI remains to be seen.
More from Fun
- Skeptic skips Fable 5.2: naming implies Anthropic lacks confidence — max_paperclips · 2026-11-10
- AI newsletter headline pairs Jensen Huang dinner photo with "Top Spot on Naughty List" — var_epsilon · 2026-11-10
- Fable 5.2 nails claymation in user challenge, surprising even skeptics — dfeinition · 2026-11-10
- Dev slams X login wall, says he'll publish Articles elsewhere over locked discussions — Darpinian · 2026-11-10
- VC's 'best founders had no hobbies' take gets fact-checked with J.P. Morgan's yacht races — zacharynado · 2026-11-10
- Ashlee Vance films SendCutSend founder Jim Belosic and the 'Teslonda' that started it all — kyliebytes · 2026-11-10