Post

The weird things are: - everyone is re-building the same idea and evangelizing for their own spin of it - it's impossible to grow into anything, because every month someone moves and repaints all the knobs

Obligatory: "It's in the SharePoint"

I suspect it’ll contribute to general decoherence and sprawl but it’s a worthwhile struggle because we’ll gradually get better at doing this and significant improvements in tooling will make their ways into OSS again, leading to more shared tooling over time? Maybe? Kinda? I see a net good

We must at least have a shared primary source. Like Linux, no matter what new distro comes up, we understand what's inside

I'm noticing a lot of it isn't particularly useful or solves problems people think the tooling will solve. It's a bunch of people trying to vibecode their way out of establishing real processes and practices. People problems are usually not solvable with tech.

Sorry, I meant to add: this is the sort of tool people might find useful for exploring code at a new job, or just explaining a codebase that was model-created.

It uses chapters to drill down into the features of your codebase. You can pause and resume a tour. Diagrams show data flows and class/function relations, with highlighted code and links to the source. It can author and export new tours as JSON for sharing and version control.

The agent help is real for day-one onboarding, and it also hides how much tribal knowledge moved into private CLIs. I'd still keep a short why-this-tool-exists note in the repo, because the next hire's agent will invent a different workflow from the same sparse docs.

Unfortunately most custom tools are just more "incomplete" than off the shelves ones. Or: the learning curve steeper.