Your Cheapest AI Decision Costs a Quarter of a Cent
Podcast: AI Breakdown ยท Varun Sirohi
Opublikowano: 2026-10-02
๐ ๐๐ซ๐จ๐ง๐ญ๐ข๐๐ซ ๐ฆ๐จ๐๐๐ฅ ๐๐๐ฆ๐ ๐จ๐ฎ๐ญ ๐จ๐ ๐ฌ๐ญ๐๐๐ฅ๐ญ๐ก ๐ญ๐ก๐ข๐ฌ ๐ฐ๐๐๐ค ๐ญ๐ก๐๐ญ ๐๐๐ง๐ง๐จ๐ญ ๐ฐ๐ซ๐ข๐ญ๐ ๐ ๐ฌ๐ข๐ง๐ ๐ฅ๐ ๐ฌ๐๐ง๐ญ๐๐ง๐๐. ๐๐ญ ๐จ๐ง๐ฅ๐ฒ ๐๐๐๐ข๐๐๐ฌ. ๐ณ You hand it a message and a question, does this customer sound angry, and it hands back a number. 0.9. Ninety percent confident. A probability your software can act on, with no one reading anything. This week on ๐๐ก๐ ๐๐ ๐๐ซ๐๐๐ค๐๐จ๐ฐ๐ง: ๐๐ก๐ $๐๐ ๐๐ฆ๐ฉ๐ฅ๐จ๐ฒ๐๐. This splits the cheap employee in two. One that writes, and one that decides. The one that decides is the cheaper hire, and it changes what you are allowed to check. At language model prices you sample, because checking everything costs more than the mistakes. At a quarter of a cent, you check everything. Inside the episode: ๐ด ๐ ๐๐๐ judgments returned in under 0.7 seconds, for about a quarter of a cent ๐ด The new shape: the language model proposes, the judgment model decides, code executes ๐ด An audit industry for agents is forming. Regulated sellers will need it within two years Your move: find one...