Most role-playing games have a skill tree: a web of abilities where the powerful late-game moves are locked behind the basic ones you have to earn first. You don’t buy the strong ability — you build up to it through everything underneath.
It’s a game mechanic, but it’s also a reasonable picture of how getting good at something works in practice. Lately I’ve started mapping my own skills this way — the ones I use at work and the ones I use at home — and it’s clarified a few things I’d been fuzzy about.
Capability, not stuff
It’s easy to confuse acquiring something with being able to do something. You buy the tool, bookmark the course, stand up the platform, and it feels like progress. Mostly it isn’t, yet. What you’ve actually picked up is a thing that has to be understood, maintained, and used before it counts for anything.
So the unit I pay attention to isn’t the gear or the credential. It’s the capability: a specific thing I can do, and show. “I can stand up a data platform a team will build on.” “I can keep a small power system running through an outage.” Put that way, it’s hard to fool yourself — either you can do it or you can’t, and usually there’s a simple test that settles it.
The depth ladder
What makes a skill tree ring true is depth. You don’t simply “know radios” or “not know radios.” You sit somewhere on a ladder, and the rungs look about the same whatever the domain:
- Operator — you can use the system safely, as designed.
- Maintainer — you can inspect it, keep it ready, and fix small failures.
- Hacker / Integrator — you can combine systems and adapt them when conditions aren’t ideal.
- Builder — you can design and build a replacement from scratch.
- Teacher — you can hand it to someone else and have it stick.
The interesting jumps are in the middle. Operator to Maintainer is where a lot of “I own one” confidence runs out. Maintainer to Integrator is where you stop following the manual and start treating the thing as parts you can rearrange — and it’s the rung AI has made far more reachable, by closing the gap between a well-framed problem and a working version of it. The AI is a tool for getting up the ladder faster, not the thing on the ladder.
Lanes
Depth needs a second axis: which capability? I sort mine into lanes — broad areas that each get their own branch. Software, data, and AI is one, pointed less at work here and more at my own machines, scripts, and home setup. The rest are physical: power and electronics, fabrication and repair, signals and networks.
Putting them on one map is deliberate. The person who can wire up a small power system and the person who can stand up a home server are doing a lot of the same things — understand it, operate it, then rebuild it better — just in different materials. And the lanes lean on each other: you can’t stand up a radio node without power behind it, and you can’t tune an antenna without the fabrication to build it. Capability ends up looking more like a web than a list.
Every lane tops out at the same rung. The deepest node isn’t a fancier gadget — it’s being able to teach the thing. That’s where it stops being only mine: what I can teach goes further than what only I can do. The most real version of that, for me, is my kids. If I can show them how these things actually work as they grow up, they get a head start — and teaching has a way of exposing the parts I only thought I understood. Good for their skills, and good for mine.
See it as a tree
Prose is the wrong shape for a tree. The dependencies, the cross-links, the way one lane feeds another — that’s something to move around in, not read down a page.
So I built the thing itself: the same model, drawn the way a game would draw it. Lanes branch off a common root, the ladder runs left to right, and the edges show what depends on what. Drag around it, zoom in, and click any node to see what it means and how you’d prove you’ve got it.
It’s a snapshot of how I think about getting better at things — and a map I’ll keep redrawing as I do.