Comment by L_Rahman
8 hours ago
One of the things we got really lucky with is that Claude Code and not the ChatGPT app won the war for the defining AI product and it runs on your filesystem. There's a different reality where everything had to go through the API on a closed app layer and we're all begging OpenAI to add XYZ endpoint to their platform.
Anthropic is now racing to close this gap because they realize there's no lock-in. If the product is just .md files with hierarchy, you can drop any harness and intelligence on top of it. It is interoperable by default, possibly not even by intention.
We should do everything possible to stop the great lock-in that they'll attempt in the next 18 months.
I don't think this can last, because whatever advantage structured text files in a file tree provides over an API will be maximized by some format that is better than a file system.
That might mean leaning into SQLite or some other open format. My own thought is that a graph-like structure of documents will be ultimately more valuable than either a tree-like structure or a database-like structure.
But if a proprietary implementation of whatever usurps structured text in a file-system becomes popular, that company will have significant leverage.
Two advantages I see to files in folders are the ability to browse and the ability to easily read - by both LLMs and humans.
But I also agree that this has limitations. If I were to challenge you and say that Obsidian solves this, or gets close to solving it, what gaps would you say were left unfulfilled?
I am because I'm working on something in this space and your comment touches on my three main ingredients: text files, SQLite, and graphs.