I mentioned on the homepage that when my eyes need a break from Kubernetes YAML, I transfer that same re-engineering energy into the physical world — remodeling the house, fixing things around the property. What I didn’t mention is that “remodeling” is doing a lot of work in that sentence, because there is no version of this house that has ever been, or will ever be, finished.
After writing the last couple of posts about ADHD and the never-ending homelab cycle, I went looking at the research again — this time specifically around home renovation rather than server racks — and found the exact same pattern wearing a different toolbelt.
A Wall Is Just a Kitchen YAML File
The instinct that makes me tear into a Kubernetes cluster to understand it is the same instinct that makes me open up a wall I have no immediate plan for. Something looks slightly wrong, or slightly interesting, and the next thing I know there’s a hole in the drywall and a half-formed theory about what’s behind it. That’s the novelty-seeking pattern again, just with more dust.
The house version has one trait the homelab doesn’t, though: it’s harder to undo. You can kubectl delete a bad idea. A wall stays open until you close it back up.
Decision Fatigue Is the Renovation-Specific Tax
Homelab projects mostly cost you time and the occasional component. Renovation projects cost you decisions — constant, granular, often irreversible ones. Paint color, tile, fixture finish, which outlet goes where, whether this particular 1960s pipe is “fine” or “a problem.” Researchers studying decision fatigue note that by the time you’ve made the ordinary day’s worth of decisions, a renovation choice that should take thirty seconds can start to feel genuinely unbearable — to the point people start telling their contractor “just do whatever you think is best” on things they cared about deeply a week earlier.
For a brain already running hot on novelty-seeking and low baseline dopamine, that tax compounds fast. The actual physical work — demolition, framing, even a lot of the wiring — can be the fun part. It’s the eleven sequential decisions about trim that quietly drains the tank.
Some Projects Run on Passion, Some Run on Follow-Through — and They’re Not the Same Fuel
One thing I hadn’t fully clocked until reading more on this: not every project is supposed to be powered the same way. Some renovations genuinely need a comprehensive, locked-down plan up front — measurements, sequencing, a fixed outcome you’re building toward. Others can be designed to flex as you go, absorbing surprises without falling apart, because the plan itself was never that rigid to begin with.
Mistaking which kind of project you’re in is, I think, where a lot of half-finished renovations actually come from. Treating a “needs a real plan” project like a “figure it out as you go” project is how you end up with a kitchen that’s been a construction zone for eight months. Treating a flexible backyard project like it needs surgical precision is how you burn out on something that was supposed to be fun.
The House Doesn’t Have a Talyn
In the homelab, I have an entire cluster — Talyn — whose explicit job is absorbing half-formed ideas safely, away from anything that actually needs to keep working. The house doesn’t really have an equivalent. There’s no dev environment for a load-bearing wall. Every experiment happens directly in production, which is maybe the single biggest reason renovation projects feel so much higher-stakes than the lab does, even when the actual task is comparable in difficulty.
I don’t have a tidy fix for that in this post — it’s more of an observation than a solution.