Personal software can serve as a platform for user interface invention
In his talk Apps and programming: Two accidental tyrannies, Andy Matuschak bemoans the lack of innovation in user interface design that our supposed “golden era” of AI-supported software creation should have already birthed:
Despite all the manic energy in the tech industry, and all the prestigious HCI conferences, the pace of meaningful invention in user interfaces is remarkably slow, and the level of imagination remarkably tame. I want new answers, new questions, and ideally both at once.
Could personal software—now often built with AI—serve as one possible route to “new answers” and “new questions”?
Two characteristics of AI-supported personal software development point to yes:
- As Matuschak points out, software development has—until very recently—largely been closed off to many people with creative design ideas, but who lack the skills and abilities to make those ideas real. LLMs change this. Now (hypothetically) anyone with ideas for new interfaces can prototype and build those interfaces without relying on the help of a programmer.
- Personal software is not subjected to the test of market viability. Freedom from this constraint allows for a much wider range of experimentation. The possibility space is much larger. Creators can get weird, and yet those weird ideas may nevertheless prove useful for others.
References
Matuschak, A. (2026) Apps and programming: Two accidental tyrannies. https://andymatuschak.org/tat/
Links to this note
-
"Low road" software, like "low road buildings," is about design for emergence
What does this all mean for personal tools or “home cooked” software, though? I often posit that Personal software can serve as a platform for user interface invention, and it’s hard to say that personal tools are “low ego.” Yet, at the same time, they can be some of the most “low road” kind of software imaginable—evolving with whims, exhibiting all kinds of emergent behavior.