Urgent.News

What's breaking now, across thousands of outlets.

Tech

Why I Moved My Agent Knowledge Base Off Local Markdown

I moved the shared, agent-maintained part of my knowledge base off local Markdown because synchronization, paragraph identity, permissions, and concurrent writes had become application problems. I still use Markdown as an open import/export format; I no longer ask a folder of files to behave like a live multi-user database. Doco series · Article 13 · Migration retrospective This is not a breakup…

The author moved their shared, agent-maintained portion of a knowledge base away from local Markdown due to synchronization issues, paragraph identity confusion, permission concerns, and challenges with concurrent writes. While Markdown remains a favored format for notes because of its readability, portability, diffability, and widespread support, the author found it lacking as a live multi-user database.

Moving the knowledge base off local Markdown allowed for a browser-based editing system where agents could search, cite, and maintain the same material simultaneously. This shift addressed issues like copy ambiguity, paragraph identity problems, read-write races, and permission restrictions that were inherent in a local Markdown-based system.

The author emphasizes that while Markdown can still be used as an import/export format, the collaborative document became the active source of truth for their knowledge base.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

Numbers Do Not Belong in the Model

A missing number gets noticed. A wrong one does not. That single asymmetry is why reading documents with a language model has to be treated differently from writing text with one.

  • Numbers in language models require special attention and treatment.
  • Multi-digit numbers often split into individual units during tokenization.
  • Recommended solution involves extracting numbers from raw document before model involvement.

More from Tuesday 1 September →