Cole Medin: my AI 'dark factory' failed — here's the software factory workflow that actually works
Cole Medin · youtube · 2026-10-16
Cole Medin spent much of the year trying to build a "dark factory" — handing agents a massive app to ship straight to production — and concluded it can't work even with AGI, because LLMs can't read your minds and nobody maintains the codebase.
His working alternative is an AI software factory in three phases:
- Plan the scope, not the whole app: split releases into issues, prepare holdout scenarios, keep dependencies human-in-the-loop.
- Core loop: separate validation from code review, add a manager agent, protect files the factory can never edit, and mix models to save tokens.
- Test everything in staging before shipping.
The full pipeline is open-sourced as ai-software-factory on GitHub; the video walks through every step with timestamps.
More from coding & agent
- Claude Code Pushes Urgent Security Update Across All Installations — umeshonai · 2026-10-23
- Claude cut a 2.6-hour video export to 20 minutes — but only after being told to rebuild the server — byhewar · 2026-10-23
- 3 flight-booking rules my AI agent checks that I'd never do by hand — Efistoffeles · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23