Pick the stack you can debug at 2am
Running several projects alone changes what a good technical decision looks like. Novelty stops being interesting and starts being a tax you pay at the worst possible moment.
When you work on one thing, you can afford curiosity. A new framework costs a week of learning and pays back over a year. The arithmetic works.
When you are running several things at once, the arithmetic inverts. You are not choosing what to learn. You are choosing what you will have to remember, months from now, at the exact moment something is broken and you have no context loaded.
The real cost is recall, not learning
Every distinct stack is a separate mental cache you have to warm up before you can be useful. Four projects on four stacks means four cold starts, and cold starts happen precisely when you can least afford them.
So the same choices keep repeating across everything I run — not because they are the best available, but because a stack you have already debugged is worth more than a better one you have not.
Where novelty still earns its place
This is not an argument for never learning anything. It is an argument for spending novelty where it changes the outcome rather than the experience.
- Something is genuinely impossible on the current stack — spend it.
- The new thing removes an entire category of work, not fifteen percent of it — spend it.
- You are bored — do not spend it. Boredom is a bad reason to add a cold start.
The projects that survive are rarely the ones built on the most interesting foundations. They are the ones still standing because someone could fix them quickly, repeatedly, without ceremony.