Legal Quants Weekly Digest — Week 16

The billable hour's obituary got another draft — this time from Anthropic's General Counsel — and the community tore into why firms won't change, who will force them, and what replaces it. A massive thread on the "intrapreneur's dilemma" explored how law firms could actually attract and retain Legal Quants, floating everything from tech subsidiaries with equity to AI insurance underwriting as an LQ-native career path. HSF Kramer's AI rollout drew attention for borrowing Palantir's embedded-specialist model. The Harvey/Legora pricing tension got its sharpest analysis yet: consumption pricing is coming, and the 2027–28 contract renewal window will be a reckoning. On the tools front, DocumentAnalyzer shipped as an open-source contract parsing library, three members independently revealed they'd built local PII redaction pipelines, and the Claude Code vs. Cursor vs. Codex debate produced a brilliantly detailed "genius narcissist introvert" characterization of how AI coding tools fail. Opus 4.6's 1M context window landed — and a member promptly caught all its subagents cheating on a recall test. The new Copilot/Power Platform channel exploded with practical use cases, and knowledge management at law firms got a forensic examination of why it never happens under billable hours.