Concepts
12 entries in Dark FactoriesAGENTS.md
An open standard file format for giving AI coding agents context about a project. Co-launched by Google, OpenAI, Factory.ai, Sourcegraph, and Cursor. Adopted in 20,000+ GitHub repositories. The cross-vendor equivalent of Anthropic's CLAUDE.md.
AI-Native Startup Economics
The revenue-per-employee economics of companies built around dark factory principles. AI-native startups average $3.48M revenue per employee vs. $610K for traditional SaaS, a 5.7x difference. Midjourney ($500M ARR, ~150 employees, $5M+/employee), Cursor ($1B ARR, Nov 2025), Lovable ($200M ARR, $6.6B valuation). Source: Jeremiah Owyang's Lean AI Native Leaderboard.
Digital Twin Environments
Simulated replicas of external services (Slack, Jira, Google Docs, databases) used for safe AI agent testing without touching production systems. StrongDM's digital twin universes are a core component of their dark factory architecture.
External Scenario Testing
A testing methodology where test scenarios are stored outside the codebase, preventing the AI agent from seeing them during development. Solves the critical failure mode where agents game their own test suites — implementing to pass tests rather than to solve the problem.
Further Reading
A curated reading list for engineers who want to go deeper on software dark factories — primary sources, technical breakdowns, critical perspectives, and practical guides for moving up the Five Levels.
Healer Agent
An autonomous agent that diagnoses production issues, investigates root causes, and applies fixes without human intervention. Part of StrongDM's factory infrastructure. Represents the logical extension of the dark factory — not just autonomous creation, but autonomous maintenance.
Specification-Driven Development
The methodology at the heart of dark factory development. Instead of writing code, engineers write specifications — precise, machine-readable descriptions of what software should do. The bottleneck shifts from implementation speed to specification quality. Requires a fundamentally different engineering skill set.
The Dark Factory
A software development operation where code is written and reviewed entirely by AI agents, with no human involvement in implementation. Named by analogy to manufacturing's "lights-out factories" — facilities automated enough to run in the dark. Currently demonstrated in production by StrongDM.
The Five Levels of AI-Assisted Development
Dan Shapiro's five-level taxonomy describing the spectrum from basic AI autocomplete to fully autonomous dark factory development. The most widely cited framework for assessing where a team sits on the path to autonomous software development.
The Full Dark Factory Stack
A layer-by-layer guide to the tools and architectural choices required to build a Level 5 software dark factory in 2026. Covers specification, agent execution, testing (the hard part), code review, CI/CD, and the honest assessment of what's hard and who can actually do this.
The J-Curve of AI Adoption
The productivity pattern where adding AI tools to existing workflows causes an initial drop before eventual improvement. Most organizations are stuck in the dip. BCG research quantifies two modes — "Deploy" (tools only, 10-15% gain) vs "Reshape" (workflow redesign, 30-50% gain). The METR 2025 RCT shows the dip is real and measurable.
The Talent Reckoning
The market-level disruption of software engineering employment driven by AI automation. Junior developer employment dropped 9–10% within six quarters of widespread AI adoption. UK graduate tech roles fell 46% in 2024. The career pipeline is hollowing from the bottom up, with significant implications for how the next generation of senior engineers develops.