Multi-Claude Code Agents Autoformalize 500-Page Math Textbook at $100K Cost
gordic_aleksa · x · 2026-07-31
Beautiful things are happening in mathematics thanks to LLM autoformalization efforts, building shared mathematical infrastructure useful beyond math for hardware, OS, databases, and compilers.
In a recent paper, researchers built a harness to coordinate a large team of Claude Code agents and successfully autoformalized a 500-page textbook into Lean.
- Cost & Optimization: The project cost $100k. However, the authors noted that after identifying various orchestration inefficiencies, the cost could easily be reduced by 3-10x without even accounting for future improvements in model capabilities.
Related event: Multi-Agent System Formalizes 500-Page Math Textbook(2 posts)→
More from coding & agent
- 'LLM code is sloppy' rebutted with real ffmpeg code: identical if/else branches in a top library — kingslyj · 2026-10-16
- Office-editing tool calls need file, worktree and unit: lessons from dsh-univer-office — FunDistribution6537 · 2026-10-16
- Architect MD workflow: Gemini one-shot outputs that beat Opus, via 5-lens plan critique — oikk01 · 2026-10-16
- Prove a coding agent actually ran the tests: check artifacts, not approvals — ComprehensiveShake76 · 2026-10-16
- Building an acceptance harness for AI spreadsheet edits: allowlists, recalc checks, idempotency — ukdicha-modakkk · 2026-10-16
- Agent actions need checks outside the model: NVIDIA demos lean on guardrails amid wave of real agent incidents — Creamy-And-Crowded · 2026-10-16