Skip to main content

Envisioning is a research institute that studies how institutions adapt to technological change.

LinkedInInstagramGitHub

Since 2010

research
  • Observatory
  • Adaptive capacity
  • Newsletter
  • Methodology
  • Origins
  • Vocab
  • RSS feeds
services
  • Signals Session
  • Bespoke Projects
  • Build Sessions
  • Pricing
  • Use cases
  • Signals
  • Signal Scan↗free
impact
  • ANBIMAFuture of Brazilian Capital Markets
  • IEEECharting the Energy Transition
  • Horizon 2045Future of Human and Planetary Security
  • WKOTechnology Scanning for Austria
solutions
  • Innovation
  • Strategy
  • Consultants
  • Foresight
  • Associations
  • Governments
  • L&D
resources
  • Partners
  • Coding for Non-Coders
  • How we work
  • Data visualization
  • Multi-Model Convergence
  • FAQ
  • Security and privacy
  • Public sector
about
  • Manifesto
  • Community
  • Events
  • Support
  • Contact
ResearchCapabilityServicesSignalsAbout
ResearchCapabilityServicesSignalsAbout
  1. Home
  2. Vocab
  3. Dark Factory

Dark Factory

A software factory whose output no human ever reads, with automated checks and automated review as the only gates on what ships.

Year: 2026Generality: 350Added: Sep 24, 2026
Back to Vocab

Dark factory names a software factory (a system where triggers, not a person, start coding-agent sessions) where nobody reads the code it produces. The term is documented in Matt Pocock's AI Coding Dictionary (aicodingdictionary.com, open-sourced on GitHub as mattpocock/dictionary-of-ai-coding). Humans can still write the issues that start the work, but no person asks for each individual change or looks at it afterward. The name comes from "lights-out" manufacturing, where a factory runs with no people on the floor and so needs no lights.

A dark factory extends vibe coding, accepting AI-written code without reading it, from a single change to a whole area of a codebase. In ordinary vibe coding, a developer chooses not to read one change they asked for, but knows the change exists. In a dark factory, the team makes that choice once, for the full area. After that, no person asks for or sees each change as it lands. Changes arrive as fast as the triggers that start them. Automated checks and automated review become the only gates: whatever they fail to catch ships into the code.

The dictionary flags the resulting failure mode. When something breaks, nobody knows what changed, because nobody read the changes, and the cause could be in any of many candidates that each passed its checks. It also draws a boundary against overuse of the term: agent sessions that run unattended (AFK) while a human still reviews the resulting pull requests make a software factory, not a dark factory.

Sources

  1. Dark factory | The AI Coding Dictionary

    AI Coding Dictionary (Matt Pocock)

  2. Dark factory.md

    GitHub

  3. mattpocock/dictionary-of-ai-coding

    GitHub

Research this in Signals

Scan Dark Factory for yourself.

Signals turns a topic into a sourced research record you can inspect and rerun. Your first scan is free, and this one starts with Dark Factory already loaded, so edit it or scan as is.