Take: frontier LLMs are too slow and pricey — small models win workflows
TejasKumar_ · x · 2026-09-17
Ken Wheeler argues jev is the best thing dropped in a while: when adding concrete value to workflows, pipelines, or apps, frontier LLMs are too slow and expensive. Save Astra/Fable for herculean tasks like solving millennium math problems and authoring React apps. Tejas Kumar agrees.
More from coding & agent
- 'English is the new coding language, code is assembly' — a one-liner coding take — 0xsachi · 2026-09-18
- WebMCP Benchmark: Jev + Mercury 2.5 Solves 100% of Tasks at 245x Lower Cost Than Computer Use — hackgoofer · 2026-09-18
- Eval claim unravels: author admits he never inspected the outputs — xeophon · 2026-09-18
- Developer builds a tiny TV for Codex showing live quota and task status — paw_lean · 2026-09-18
- JEV classifies 1,891 ads in 19 seconds for $0.12 in competitor research test — SucceededMind · 2026-09-18
- 15 harnesses, same result: exchange over whether eval harness choice even matters — paul_cal · 2026-09-18