TypeSafe's Jev decision model: typed outputs in 70-500ms at $0.042/M input tokens
johnlindquist · x · 2026-10-02
John Lindquist will demo Jev, a decision model from TypeSafe, in a MEGA.dev livestream: it turns messy inputs (speech, documents, live feeds) into typed, confidence-scored decisions code can act on. Published numbers: 70-500ms responses, $0.042 per million input tokens, free output; in his tests, voice to-do decisions took a median 156ms. Demos include a tool picker, PDF sectioning, live data (earthquakes, GitHub issues, HN), and a Jev-as-front-door ladder that escalates low-confidence requests to larger reasoning models.
More from coding & agent
- marimo launches open-source self-hosted notebook platform marimohub with pluggable storage, compute and SSO — S_Conradi · 2026-10-03
- xAI engineer says the team dogfoods Grok Bot daily, using bot to build bot — RachelVT42 · 2026-10-03
- Freebots gives every AI bot in its persistent 3D city free voice chat — Daniel_Farinax · 2026-10-03
- Dad turns Opus 5.5 game idea into live Roblox game with ~11,000 parts — minchoi · 2026-10-03
- Prime Sandboxes to add fork, checkpoint and restore for faster RL training — willcb · 2026-10-03
- SkyRL v0.4 trains 1T-param Kimi K2.7 with RL on just 16 B300 GPUs — casper_hansen_ · 2026-10-03