Boris Cherny had 12 Claude Code agents turn 64 research papers into a browser lab you can rerun
ID_R_McGregor · x · 2026-10-13
Boris Cherny, creator of Claude Code, gave 12 agents 64 research papers with one assignment: build a browser lab where readers can rerun the experiments.
- Group division of labor: readers extract equations, assumptions, and experimental settings; builders turn methods into runnable code; reviewers compare outputs against the papers (checking mass conservation, stability, and measurements) to guard against convincing-but-wrong simulations; interface agents build exploration controls.
- Experience: pick a fluid simulation, change conditions, switch methods on the same scene, and inspect equations, code, and measurements beside the animation.
The post also quotes an article by 0xCodila on REA (reverse engineering agent), which lets an AI agent reverse engineer any app down to the binary — the author claims you can rebuild paid features for free, with a full tutorial linked.
More from coding & agent
- Constraint-context prompting: telling the LLM what NOT to do yields actionable output — JafarNajafov · 2026-10-13
- Reference-context prompting: paste examples and name the patterns for precise output — JafarNajafov · 2026-10-13
- Persona + expertise prompting turns generic reviews into senior-engineer code audits — JafarNajafov · 2026-10-13
- Context engineering, not prompt engineering: 8 techniques for 10x LLM output — JafarNajafov · 2026-10-13
- One-Person Company Runs 62 Grok Agents, 29 Scheduled Routines Write Its Newsletter — elonmusk · 2026-10-13
- Socratic prompting: stack questions before answers for strategy-level AI output — bigaiguy · 2026-10-13