Benchwarmer tool re-renders AI benchmark charts, recomputes winners from raw numbers
aronchick · x · 2026-09-06
aronchick turned Bad Apple into a mashup of AI frontier-lab benchmark charts, built with a new tool called Benchwarmer (benchwarm.ing). Its real job: paste a benchmark screenshot or table, and it reads the values, ignores misleading source styling, and recomputes the winner per benchmark — showing the original beside a corrected, exportable comparison. "Color never overrides arithmetic": ties and missing data stay explicit, with sources, versions, and caveats kept alongside.
Related event: Bad Apple Remade with Frontier AI Labs' Benchmark Charts(3 posts)→
More from Fun
- 'What did Ilya see?' Joke answer: AI powerful enough to hack everyone — imjustnewatai · 2026-10-23
- Dev builds SkiFree-style sled game in one prompt, praises DeepSeek Flash — victormustar · 2026-10-23
- Effect's author: released today, the massive TS library would be dismissed as slop — ethanniser · 2026-10-23
- Abliterated V4.1-Flash eval: abstract algebra drops 6pp but college math gains 5pp — teortaxesTex · 2026-10-23
- Moody AI-reimagined Harry Potter portraits showcase stunning cinematic generation — azed_ai · 2026-10-23
- 'Time for an Ana de Armas benchmark': X user jokes about new Gemini output — lyraxana · 2026-10-23