In this new world, the most important thing you’ll need is taste. Not objectively good taste, necessarily, so much as a keen sense of your own.
David Pierce spent weeks vibe-coding his way through Timetable, then Spring, then Basket, before landing on the thing that actually worked: a screen that pulls Raindrop, Todoist, Obsidian, and Google Calendar into one paper-planner view. Four API keys and an afternoon. The lesson he draws isn’t that AI made him a developer. It’s that the ceiling on personal software is now your own sense of what you want. As he puts it, “AI lets us make apps the way we used to make lists and spreadsheets.”
I keep coming back to the Rick Rubin line he quotes: confidence in your taste, and the ability to express what you feel. That used to be the producer’s job. Now it’s also the user’s. Claude Code will happily ship you a working app in twenty minutes; whether it’s an app worth keeping is a separate question, and the model can’t answer it for you. Pierce names the failure mode directly:
Otherwise, you’ll land in what Lovin calls “doom loops,” telling your chatbot only what you don’t like and counting on the model to be the creative one. That way lies madness, and bad software.
There’s a quieter point underneath the trend coverage. Robin Sloan, five years after building a family messaging app by hand, updated his original “home-cooked app” post with one line: “I have changed literally nothing in the app, and it’s glorious.” That’s the version of personal software I want to live in. Not the disposable kind that disappears with the need, the kind that stays exactly itself because nobody is trying to grow it.
The shift nobody warns you about is that the UX is now yours. Every bit of friction, every bit of elegance, you put there. The default app at least had someone else to blame. Your app does not. And “simplicity” stops being a setting you can turn on; it becomes a thing you have to actually design, which means deciding what’s in and what’s out, and living with the trade. Simplicity, in other words, suddenly means complexity. Just on your side of the screen now.
A session with Claude isn’t that different from the sessions I run between business, dev, and design. Someone has to hold the intent. Someone has to push back when the easy answer wrecks the harder one. Someone has to notice that the thing being built has drifted from the thing that was wanted. The model can play any of those parts, but not all of them, and not the one in the middle.
So most of the work happens before the code. Thinking out loud. Pushing on what I actually want versus what I think I want. Walking away and coming back. The model is not the bottleneck anymore. I am, and I want to be. That’s where the thing turns out to be mine.
www.theverge.com