Software Is Becoming Content

A post made the rounds recently arguing that a lot of new software will behave like media. Made quickly, shared in a feed, used in the moment, remixed, maybe forgotten by dinner. I build small tools, and the description fits more of my own work than I was comfortable admitting at first read.

The idea feels wrong only if you carry the older mental model, the one where every useful piece of software is supposed to justify itself with a pricing page, an onboarding flow, a roadmap. That model never described what software was for. It described what software cost. When something takes months of skilled labor to build, it has to last, because durability is how you pay the labor back. Permanence was an economic requirement wearing the costume of a design principle.

That cost is collapsing. Anyone who has shipped a side project in the last year knows the build is no longer the hard part, and I wrote recently about how the scarce skill moved to judgment. When generation gets cheap, the constraint moves. It stops being "can we build this" and becomes "will anyone see this." Attention is the bottleneck now, and attention lives in feeds.

The feed is a different kind of shelf

Feeds do not reward what app stores rewarded. An app store is a warehouse: things sit there, ranked by accumulated reviews and download counts, expected to persist. A feed is a current. It rewards timing, relevance to this week's conversation, the small tool that solves today's annoyance. The half-life of the product starts matching the half-life of the post that carried it.

Most of what people consume online already works this way. Memes, posts, videos, the link someone drops in a group chat that gets forty minutes of intense use and never gets opened again. Software has been the exception because it was too expensive to be disposable. As it gets cheaper, it joins everything else in the current. The builder's question changes with it. Not "will this retain users for a year" but "is this worth someone's afternoon, right now."

I can feel this in my own habits. The tools I have actually shipped and shared lately are closer to toys and instruments than to products. A thing that does one job, pointed at a moment. Some of them I still use. Some I forgot I made. That used to feel like failure. Under the new economics it is closer to the natural shape of the medium.

What this does to the people who build

The skills that made someone good at the old model were stamina and structure: spec it, staff it, ship it, maintain it. The skills the new model rewards are timing and taste. Knowing what the feed will care about this week. Knowing which itch is shared widely enough that a scrappy tool for it will travel. Judgment about what to make starts mattering more than endurance in making it, which is a sentence I did not expect to write five years ago.

There is a catch worth naming. Ephemeral works for instruments and toys. It does not work for anything that holds someone's data, someone's money, or someone's business process. Nobody wants their accounting stack to be content. The current carries some kinds of software beautifully and sinks others, and the difference is how much trust the thing has to store.