Every Builds Personal Benchmarks for Each Employee — Finds a Smaller Model Beats the Big Ones
every · x · 2026-09-11
Every argues public benchmarks can't tell whether a model does your job, so it's building a custom eval for every employee, graded against personal standards.
Key facts
- Astra and Fable 5.1 scored 96% and 93% on graduate-level science exams, but that says nothing about comma placement or one-idea-per-slide preferences.
- Editor-in-chief Kate Lee's benchmark encodes her copy-editing rules; evals lead Mike Taylor's centers on deck-making.
- In testing Mike's daily tasks, GPT-5.6 Luna outperformed Fable and GPT-5.6 Sol.
- The piece offers a five-step workflow to turn your existing AI corrections into reusable checks for grading any model.
Related event: Every Ditches Leaderboards, Builds Personal Benchmarks for Real Work(2 posts)→
More from coding & agent
- Magic Patterns' first engineer: Cursor in Slack, a Patterson review agent, repo clones over worktrees — Y Combinator Blog · 2026-10-28
- Dev builds AI job-search app for his wife, she lands a job, and he open-sources it — CurieuxExplorer · 2026-10-28
- Meshy partners with MCP for Blender: agents drop textured 3D models into your scene — sidahuj · 2026-10-28
- Supercharging LLM classification with logprobs: soft classifiers you can threshold-tune — JnBrymn · 2026-10-28
- Codex daily build series day 23: skipping iCloud plugins for a reset — dotey · 2026-10-28
- Cloudflare Access adds fully customizable pages with HTML + Liquid templates — dinasaur_404 · 2026-10-28