Simulated GitHub run by agents shows AI-agent collaboration follows its own scaling laws
tongshuangwu · x · 2026-09-03
Researchers built 'Software World' — a GitHub simulated by agents, sub-sampled from real Python dependency data — and deployed agent teams to optimize code efficiency, studying large-scale agent collaboration.
Key findings
- Agents identify code deficiencies and produce patches individually
- Meaningful collaboration emerges: agents ask downstream (agent) maintainers which code paths matter before optimizing
- Upon upstream changes, downstream agents measure gains against their own code and report back numbers
- Collaboration quality varies significantly by base model
Evaluation: held-out downstream packages serve as extrinsic evaluation — if the ecosystem improves, they should benefit too.
Microsoft researcher Tongshuang Wu: agent-maintained repos inherit the same dynamics as human ones, and 'agent-agent collaboration has its own scaling laws.'
Related event: Software World: Agent-Run Simulated GitHub Reveals Agent Scaling Laws(2 posts)→
More from coding & agent
- Solo builder crawls 10B+ records a day to put the internet in an analytical database via MCP — AaronBergman18 · 2026-09-03
- Cursor cloud agents add Cloudflare support via self-hosted Containers pools — dinasaur_404 · 2026-09-03
- Cursor cloud agents can now run on self-hosted machines and AWS Lambda MicroVMs — dinasaur_404 · 2026-09-03
- Fable 5.1 hands-on: faster and cheaper, but "coding is solved" rings hollow — vboykis · 2026-09-03
- Open-source "Yingzao" skill turns travel snapshots into magazine-grade architecture posters — op7418 · 2026-09-03
- antirez: Zero Out Steering When Possible—It Always Adds Distortion — antirez · 2026-09-03