Config-as-code AI factories: one factory learned another's benchmark scoring setup
vikvang1 · x · 2026-11-18
The author keeps his AI software factories fully configured as code, which lets one factory trivially adopt another's setup. While building benchmarks to test verification techniques on his mobile-factory, he didn't know how to score runs — so he simply tagged both factories in one Slack thread and had mobile-factory inspect and implement eng-factory-slim's existing scorer and LLM-as-a-judge configs as a starting point. A screenshot shows the two agents collaborating via PRs ($160.52 vs $0.58 runs), and one run even caught a harness bug where the iOS simulator crashed. Config-as-code turns factories into reusable knowledge for each other.
More from coding & agent
- GPT-6.2 Sol passes all 16 threejseval scenes, 15 first-try — k1rd · 2026-11-18
- Same Blender task: GPT-6.2 Astra finishes in ~2h vs Fable 5.2's ~20h — ChrisGPT · 2026-11-18
- First Agentic Graphics Workshop at SIGGRAPH Asia 2026 to explore coding agents for graphics research — elliottszwu · 2026-11-18
- Open-source Passband auto-extracts plans from email onto your calendar — braelyn_ai · 2026-11-18
- Claude Code 2.1.327 details: prompt tokens up 3.6%, new model names appear — ClaudeCodeLog · 2026-11-18
- Claude Code 2.1.327 released with 173 CLI changes — ClaudeCodeLog · 2026-11-18