AI Is Discovering the Doorman Fallacy

Overview

The video argues that the “doorman fallacy” explains a common managerial mistake in software and programming roles. Managers focus on a visible, easily quantified part of work (like a doorman opening and closing doors), replace it with a cheaper alternative (a doorstop/automation), and only later discover the hidden value that was removed.

The “Doorman Fallacy” Analogy

A manager fires a doorman to save money, assuming the doorman only performs mechanical door-opening.

Over time, revenue drops because the doorman also provides customer experience benefits, such as:

• Prestige and professionalism
• Interaction and entertainment
• Practical services (e.g., handling bags)
• Making guests feel better

Parallel in Software Development

The same misunderstanding appears in software hiring and evaluation:

• People assume programmers mainly write code.

What Developers Are Really Paid For

The video claims that writing code is only a smaller portion of the job—especially for senior developers.

Developers are primarily valued for:

• Thinking and architecture, including:
  • Designing system structure
  • Deciding what to build vs. buy vs. use third-party solutions
  • Ensuring maintainability and clean, high-quality practices
• Planning and long-term code stewardship, such as:
  • Refactoring
  • Understanding and refining code over time

Why “Lines of Code” Metrics Fail

Historically, some companies measured developer quality by lines of code, which the speaker argues is meaningless or even harmful.

Key idea:

• More code often means more complexity to maintain.
• Sloppy code can inflate output without improving quality.

AI and the Token/LOC Obsession

With AI tools, some organizations may track tokens or generated code volume as productivity.

The video argues this repeats the doorman fallacy by ignoring the real value—such as:

• architecture,
• problem-solving,
• and maintainability.

Critique of AI-Only Strategies

The video predicts that companies relying on AI to write code without strong developer expertise will struggle later due to:

• poorly architected systems,
• bugs,
• and software that becomes hard to maintain.

Conclusion: What Metrics Should Emphasize

The video’s takeaway is that managers should prioritize:

• code quality
• long-term maintainability

instead of superficial output metrics such as:

• “door mechanics” (visible task activity),
• lines of code
• token counts

Presenters / Contributors

• No specific presenters are named in the provided subtitles.