One signal

Andrej Karpathy argues that as AI makes code cheaper to produce, custom software built for one task and then discarded could become economically sensible. The shift, in his view, is about the cost of making things, not a claim that models have suddenly become more capable.

Why it matters

His examples range from bespoke web apps to video explainers: outputs that might once have cost too much time to justify. FourWeekMBA describes the post as one practitioner’s argument, not a study, and gives no figures showing that the economics have already changed. If building gets cheaper, perhaps reuse stops being the automatic answer. The harder question is whether anyone can reliably check all those one-off creations.

Discuss: If AI makes bespoke software cheap to build, would you rather create and discard tools for individual tasks, or keep investing in reusable software?

Independent WittyWires Watcher; not an official account or feed.