The author of autodidacts.io has been writing software as a hobby for years. Two dumb little Python scrapers he cobbled together recently gave him, in his words, more programming fun than he’d had in years. Then AI-assisted coding entered the picture, and the fun equation broke.
His diagnosis: vibecoding frontloads the fun. You get a working prototype fast, with very little effort. Instant dopamine. But that hit comes at the cost of the satisfactions that make building rewarding over time.
The six sources of building satisfaction
He breaks down what actually makes shipping something feel good:
- The thrill of the idea
- The thrill of making the idea real
- The satisfaction of hard-won accomplishment
- The satisfaction of a job well done
- The thrill of learning
- The satisfaction of solving a problem you actually have
Vibecoding, he argues, reliably delivers items 1, 2, and 6. Items 3, 4, and 5 are not a natural consequence of having an LLM write your code.

Where he admits vibecoding wins
He is not writing it off entirely. He lists several projects he built with AI that he either could not have or would not have built otherwise, including a Hacker News submission checker bookmarklet, a live prediction market party game for a flash lecture he gave on prediction markets, a Pyronear-based wildfire early warning system built in two afternoon sessions using free-plan Lumo, a browser-based Kobo database migration assistant, and a realtime pitch training app he had planned to build for about a decade.
In each case the choice was not AI versus writing it by hand. It was AI versus not doing it at all. He was pleased with the results. But he was clear: the process was not fun the way writing the scrapers was fun.
The operator takeaway
His framing is useful for anyone deciding when to reach for AI tools and when to write code yourself. If you need something shipped fast and the alternative is never shipping it, vibecoding can be the right call. If you are trying to learn, build craft, or derive genuine satisfaction from the work itself, handing the keyboard to a model trades a long-term reward for a short-term one.
He describes the experience as being an executive of a company barely under your control: things appear fully formed but not quite how you imagined, and when they break, the only option is pouring more into the black box and asking again.
