Tracking one tennis ball with GPT-6 burns 7.87M tokens
Scobleizer · x · 2026-09-07
LearnOpenCV ran a tennis-ball tracking demo with GPT-6 Astra Ultra: ball annotation via 3 parallel agents took 8m35s and 17,611 tokens; the full workflow took 11m49s and 30,481 tokens, consuming 7.87M total tokens (313,914 uncached + 7.5M cached) mostly reused across frame-inspection calls. Scobleizer quips 2030 = 200 billion cat trackers.
More from Fun
- Viral post: we refine sand to 99.99999% purity to build the smartest thing ever — for email — CatAstro_Piyush · 2026-10-27
- $50, 2-hour Kurzgesagt-style AI video reignites debate over intent in AI creation — motionbynick · 2026-10-27
- Fake-paper meme shows an agent workflow for reverse engineering trading systems into a one-person quant fund — PtrPomorski · 2026-10-27
- Steering two orthogonal concept neurons makes a model compose "gnome" — Sauers_ · 2026-10-27
- Fake BTS of a dam disaster made with Dreamina Seedance for 29 cents — Standard_Spell5873 · 2026-10-27
- Meme chart roasts the monogs at San Francisco AI parties — NathanpmYoung · 2026-10-27