call me crazy but company knowledge should look more like this. > every file becomes a node > every import, call, and dependency becomes an edge > every answer points back to the exact code behind it click one class and the system shows what it depends on, what... depends on it, and where it sits inside the architecture. now apply the same idea to the company: decisions. projects. people. policies. workflows. learnings. AI should not search through a pile of documents and guess how they connect. it should navigate the company as a system. don't bookmark this if it crosses your timeline. paste the full article into your agent and tell it to build the first version for your company.show more

yo

cracked

calling you crazy on this

fair

真做出来后，最难的可能不是建图，而是持续维护这些边。

agreed brother

Graph views beat flat documentation folders every single time.

@andrewdsouza J.B., this could fit the ops and enablement teams I know that are turning scattered AI work into shared workflows. Happy to see who might be useful to meet.

And nothing useful can be done with that...

is this made on hqforwork?

The interesting part is not storing knowledge. It's remembering why a decision was made, and using that context next time. That's what makes a company brain useful.

this is crazy

We've built similar thing but for software. The comformance layer forces the agent to keep the knowledge base up to date, deterministically verifiable, with test coverage and proof.

Anyone done a good cost differential between graph and flat files for something this big?

We built this! It’s free. Check out

Not crazy, this is the shape that actually holds up. The same idea works on the data side: point an AI at a schema and the foreign keys are already the edges, so it can write the JOIN without being told the relationships. Explicit keys beat naming conventions there, same as real imports beat guessed ones.

this is how knowledge bases should have always worked.

And what will you do when it grows? You are just building a relational DB, nothing more, so querying thru that will be the problem. So it's not an AI problem, it's a CS problem

Every founder I know who's tried this hits the same wall. Nobody structured the data in the first place. The graph is the easy part.

The graph is the easy part. The hard part is the write-path — who is accountable for turning a correction into a node instead of a Slack message that dies in a thread.

A company brain only compounds when every answer is tied to source context. That traceability is what turns AI from a clever demo into an operating system people can trust.

Looks cool!

I think generating this kind of graph is important as a backend for actually usable views. The raw graph itself is the most abstract rendering of the relationship between facts, which often have much better visualizations themselves

We're doing it at

Designing the system is more important than designing the folder :)

There is also a huge maintenance advantage here. Documents go stale silently. Relationships can tell you what becomes suspicious when something changes.