Crosby scales legal AI judgment on 1.5 years of attorney edits: labels beat architecture
graceisford · x · 2026-10-22
Crosby's new article 'Scaling Legal Judgment' details their legal verifier built on 1.5 years of attorney edit/accept decisions. Key lessons: expert attorney disagreement is signal not noise (three lawyers often give three different verdicts on the same redline), clean labels and right context beat architecture tweaks, and a flywheel is forming with accuracy rising as data doubles.
Related event: Crosby Trains Legal AI Verifier to Predict Lawyer Acceptance(3 posts)→
More from coding & agent
- GMG Open-Sources an Apache 2.0 ServiceNow Agent With Human-Approved Changes — adnan_hashmi · 2026-10-22
- RAX drives robots with a frozen coding agent; DeepSeek version beats π0.5 at 1/20 cost — du_yilun · 2026-10-22
- Stripe partners with Anthropic to run business payments and refunds inside Claude Duet — jeff_weinstein · 2026-10-22
- Aaron Levie: autonomous domain-specific agents are AI's emerging form factor, a leap past the LLM-wrapper era — _AustinCalvert_ · 2026-10-22
- Claude Duet opens as research preview: a proactive agent that watched Slack and ordered books — altryne · 2026-10-22
- Vibe-coded ComfyUI node unlocks 2MP for Fibo 2.0 Edit, built in 10 minutes with a free LLM — WhiteZero · 2026-10-22