Chris Albon on the 'just use local models' crowd backed by $20,000 in GPUs
chrisalbon · x · 2026-10-01
Chris Albon pokes fun at a recurring trope: people confidently advising 'just use local models' while their photo shows roughly $20,000 worth of GPUs — a one-line reality check on the hidden hardware cost behind 'free' local inference.
Related event: "Just Run It Locally" Often Means a $20K GPU Setup(2 posts)→
More from Fun
- Fei-Fei Li: Those fearing the intelligence explosion haven't considered the stupidity explosion — NaveenGRao · 2026-10-01
- A 1300-rated player's first online tournament opponent: Magnus Carlsen — abhishekcode42 · 2026-10-01
- DevDay party hack: a ModRetro Tomodachi game where attendees chatted as AI-generated characters — OpenAIDevs · 2026-10-01
- Even critics admit Meta's "Superintelligence" name has a nice ring to it — AiBreakfast · 2026-10-01
- Redditor asks Astra Pro which AI lab it wants to join after reading Claude's constitution — CrowdGoesWildWoooo · 2026-10-01
- America.gov's AI chatbot refuses to activate "sexy mode," insisting on real government questions only — Polymarket · 2026-10-01