For years I've had a list of project ideas. Small things: personal tools, problems I kept noticing, things I'd actually use. Most of them stayed on the list. The cost of starting wasn't prohibitive, but it was real. Getting from idea to working prototype meant days of scaffolding: stack setup, architecture decisions, enough boilerplate to see whether the idea was worth pursuing at all. A weekend afternoon was rarely enough.
That's changed.
The cost shifted
Cloud computing didn't just make deployment cheaper. It changed what was worth deploying. A team could spin up infrastructure in minutes instead of months. The calculus of what to attempt changed entirely. Things that weren't worth doing before became worth trying.
Vibe coding is doing something similar to the 0-to-1 phase.
What used to take days of setup can now take hours. The gap between an idea and a working prototype collapsed. That's not just "the code is faster to write." It's a structural shift. When starting costs almost nothing, the question stops being "is this worth the effort?" and becomes "do I actually want this thing to exist?"
"I stopped asking whether an idea was worth starting. Now I ask whether I actually want the thing to exist."
Most of my backlogged ideas answer yes. The list that once felt like a permanent backlog starts to look more like a schedule.
The constraints
One project, started and shipped, every weekend. 3 rules that make the habit work.
The project has to scratch a real itch: something I'd actually use, not a portfolio exercise. When I'm the user, I notice when something feels off. That keeps the quality standard honest in a way that building for a hypothetical audience doesn't.
It has to be finished in a weekend. Not "two weekends if it needs it." If a weekend isn't enough, the scope was wrong. That's immediate, useful feedback. The one-weekend limit forces a question I'd otherwise defer: what's the smallest version of this that's genuinely useful?
And it has to ship. Deployed before Monday, at a URL, something that exists beyond my machine. Finishing locally doesn't count. A deployed thing has a quality a local build doesn't: it's actually been sent.
"I've got to ship it, not just finish it. That distinction is the whole constraint."
What I'm not sure about
The honest version: I don't know if this will hold.
A weekend isn't a reliable unit. Illness, travel, a Friday that runs long. One missed weekend and the streak breaks. Streaks are fragile things once broken. The habit doesn't survive the first bad weekend unless the pull to return is stronger than the friction of restarting.
Return the next weekend. The risk isn't running out of ideas or burning out. It's a busy Saturday arriving before the habit has time to calcify. A missed weekend is recoverable. Treating it as failure is what kills the habit.
What I'm betting on is that the satisfaction of shipping something, even something small, is stronger than the pull of not shipping anything. Repeated often enough, it becomes just what weekends are for.
The cost of starting is lower than it's ever been.
One project, every weekend.