Local LLM power user's 12M tokens/day humbled by one Cursor profile burning 1.5T/month
perelmanych · reddit · 2026-10-28
A Reddit user jokingly brags about running two local Strata inference engines at 70 tps each and burning 12M tokens a day — until he saw that 'poteto's' public Cursor profile shows 1.5T tokens used in 30 days (34M tokens per minute).
The attached tweet does the math: at $8/M tokens that's $384k/day or $11M/month, $132M annualized — and even with a hypothetical 90% Cursor–Anthropic discount, $38k/day, enough to hire 26 staff-level engineers at $500k/year.
More from Fun
- Meme mocks 'AI can't write real code' skeptics alongside flat-earthers — PawelHuryn · 2026-10-28
- "Every time we discover P=NP, they just reset the universe" — rakyll · 2026-10-28
- Someone used ChatGPT to write a post about how AI is destroying human creativity — _jaydeepkarale · 2026-10-28
- Researcher uses Genie 4 to have conversations with his younger self — holynski_ · 2026-10-28
- FLOP-maxxing: Running GPUs flat-out as space heaters, with the dashboards to prove it — lauriewired · 2026-10-28
- Andrew Critch's dubious credences: P≠NP at 80%, Riemann hypothesis at just 60% — AndrewCritchPhD · 2026-10-28