Comment by ssyhape
6 hours ago
Neat idea. The biggest problem I've had with code knowledge graphs is they go stale immediately -- someone renames a package and nobody updates the graph. Having it as Markdown in the repo is clever because it just goes through normal PR review like everything else, and you get git blame for free. My concern is scale though. Once you have thousands of nodes the Markdown files themselves become a mess to navigate, and at that point you're basically recreating a database with extra steps. Would love to see how this compares to just pointing an agent at LSP output.
> My concern is scale though. Once you have thousands of nodes the Markdown files themselves become a mess to navigate
The agent will update the graph.
If you have thousands of nodes in md it means you have a highly non-trivial large code base and this is where lat will start saving you time - agents will navigate code much faster and you'll be reviewing semantic changes in lat in every diff, potentially suggesting the agents to alter the code or add more context to lat.
You still have to be engaged in maintaining your codebase, just at a higher level.
We all know this isn't for humans. It's for LLMs.
So better question is why there isn't a bootstrap to get your LLM to scaffold it out and assist in detailing it.
Adding a command to bootstrap lat for an existing code base is a good idea, I'll work on it.
Other than that the goal of lat is to make the agent use it and update it and it has tools to enforce that.
You’re replying to an LLM, too, fwiw.
GraphRAG is for LLMS... markdown is for humans.. humans that exist in the meantime
I would say that when you treat your Markdown as the authoritative source, I of course don't get it automated but that is my choice. It takes knowledge of the domain, but when you have deep specific knowledge that is worth so much more than automated updates. I use AI to get the initial MD but then I edit that. Sure it doesn't get auto updated, but I would never trust advice on the fly that got updated based on AI output on the internet.
I think you can have your workflow with lat and it might make it even nicer. Would love feedback from you.
yeah, AI for the skeleton + human for the "why" is probably the right split.