Judgement small models become essential: four companies ship within eight weeks
FinanceYF5 · x · 2026-11-12
Observation from the 2026 Silicon Valley AIE conference: judgement small models are becoming essential infrastructure. On Sept 15, startup Typesafe released Jev, a judgement model that writes no text — it only answers multiple-choice questions and assigns a probability to each option. Within eight weeks, OpenAI, Cloudflare, and Google all followed suit, with Google announcing at the conference itself.
- The rush stems from agents failing to run reliably in the background: what's missing isn't a smarter large model but a judgement layer that's right every time
- The video frames five questions: who's racing, what's wrong with the referee, why judgement is hard, what changes after adoption, and what these models still can't do
Note: Chinese-language commentary, narration and animation AI-generated.
More from AGI Musings
- Dev after 2 days of ads: "now I understand why the public is anti-AI" — menhguin · 2026-11-12
- Investor Stewart Alsop: Algorithms Killed Mainstream Culture for Good — StewartalsopIII · 2026-11-12
- Steve Yegge burns 5B tokens a day: agents first break everything around the model — FinanceYF5 · 2026-11-12
- FT: AI agents supercharge start-ups — no employees, prototypes within days — nordicinst · 2026-11-12
- Erik Voorhees slams The Economist over Jensen Huang criticism: 'must go' — beffjezos · 2026-11-12
- AI Circle Debates: Would One Company Winning Break the Global Socioeconomic System? — teortaxesTex · 2026-11-12