Back to book

Part I: The Origin of AIFC

2. The First Pain: Markdown Helped AI, But Started to Hurt Human Attention

1 min read

Once the same pattern was used in multiple projects, a new problem appeared.

Markdown solved the AI continuity problem, but created a human attention problem.

There were more and more files.

They contained:

  • raw human ideas,
  • AI-generated summaries,
  • refined AI text,
  • partially reviewed sections,
  • fully approved sections,
  • outdated notes,
  • duplicated context,
  • speculative ideas,
  • decisions,
  • assumptions,
  • and long narrative explanations.

At first, this seemed manageable.

Then it became hard to answer basic questions:

  • What is my original input?
  • What did AI generate?
  • What did I already review?
  • What did I approve?
  • What is still draft?
  • What is outdated?
  • Which part of the file is reliable?
  • Which part is only AI interpretation?
  • Which file should be updated?
  • Which file is the current source of truth?
  • Which file is duplicate?
  • Which section matters now?

Markdown gave the knowledge a body.

But it did not yet give it lifecycle, ownership, status or trust.

This led to a crucial realization:

A text file can be readable and still not be human-manageable.

The problem was not only storage.

The problem was attention.