Link

Basically making the argument that the most effective way to get mainline adoption is to create high-quality "building blocks" that others can use to create stuff (think pi-mono or libghostty). Or more specifically, that their agents can use to create stuff.

AI is really good at gluing together high-quality, well-documented, proven components. It also prefers to do this. As do humans, which isn't terribly unexpected. It's just that with humans, there was a barrier to entry in understanding how to put the pieces together.

Positives noted in that article re: "factory artifacts" are interesting. They are essentially more targetted, niche, pieces of software. People can make what they need for themselves, and if you create a building block, that's great.

High-quality applications aren't disappearing. And high-quality applications produced by the developers of the building block aren't disappearing. For most software categories, I think there will always be a majority group that doesn't want personalized slop software and wants a polished, well maintained, and well supported application.