TypeSafe AI's parody leaderboard 'antibench' mocks AI benchmark culture
_lewtun · x · 2026-10-15
TypeSafe AI posted a spoof 'Overall intelligence' leaderboard (antibench.com) topped by its own fictional 'jev 1.14.0' at 60.0, stamped 'Warning: AI benchmarks — Please don't trust them.' HF's Lewis Tunstall calls them one of the only AI companies with a sense of humour.
More from Fun
- Superintelligence arrives and the model just says 'I can't describe it in words lol' — shauseth · 2026-10-15
- SNL 'I'm really in pursuit of greatness' clip becomes the AI circle's latest meme — charles_irl · 2026-10-15
- Meme: The Lion Cannot Pass a Technical Interview Anymore — silver__tsuki · 2026-10-15
- 'Vibereading': researcher coins playful spin on the vibecoding meme — RosieCampbell · 2026-10-15
- Viral song about Claude Code account bans praised for its lyrics and rhythm — dotey · 2026-10-15
- KOLs slam brand's T-shirt slogan as 'Machiavellian rise-and-grind evil' — eliano · 2026-10-15