Over the past few months I’ve built somewhere between fifteen and twenty digital products. A writing environment that thinks alongside me. A research tool that maps conversation history into structured insight. A scheduling layer that understands my energy patterns. A dozen smaller things, each solving one specific situation that no existing product quite addresses.
It should signal that something fundamental has changed — not in how fast software gets built, but in what software is becoming.
Andrej Karpathy named the production shift in 2017 when he described Software 2.0: programs compiled from data rather than written by hand. The 2025 extension — vibe coding — collapsed the remaining distance between intention and artifact. You describe what you want. It exists. The barrier between “I wish this tool existed” and “I’m using it” has functionally disappeared. But the interesting question is not about speed. It’s about what happens downstream, after the making becomes trivial, when the accumulation begins.
What actually changed
When people talk about vibe coding, the conversation usually stays at the production layer: it’s faster, it’s cheaper, more people can do it. That framing misses the deeper displacement. At least five things shifted simultaneously, and they compound.
The cost structure of software changed. Production used to be the expensive part. Now the expensive part is knowing what to make — direction, not execution. This inverts the entire logic of which products deserve to exist. Generality used to be a survival requirement — a product needed enough users to justify its development cost. When development costs approach zero, a product can afford to serve one person in one situation, and the concept of “niche” loses its pejorative edge.
The relationship between products and needs inverted. Under the old model, needs accumulated until one of them was large and common enough to warrant a product. Now, a need can summon its own solution the moment it becomes legible. Products no longer collect needs; needs generate products. And the developer role dissolved along with it — when the act of building becomes conversational, the relevant skill is no longer “how to write code” but “how to think clearly about what should exist.” The developer doesn’t disappear; the developer becomes indistinguishable from the practitioner.
What changed most, though, is the human-software relationship itself. Products were things you adopted, learned, and committed to — a relationship model closer to marriage than conversation. You chose a notes app and lived inside it for years. You tolerated its compromises because switching costs were real and alternatives were finite. When products are made in minutes and remade when circumstances shift, that entire relational logic stops applying. Software becomes closer to thought than to possession — something that takes shape around a situation rather than something you carry with you across situations.
Naval Ravikant described arriving at a similar place from a different direction: one-shotting apps with Claude, accumulating a personal app store, watching each tool serve exactly one purpose before a new one takes its place. He observed that pure software is becoming uninvestable — not because it’s worthless, but because its production has lost the scarcity that made it ownable. The same displacement, seen from the economics side.
Where the paradigm breaks
Here’s the problem no one is yet solving. If you can build fifteen products in six months, you will build fifty in a year. And then a hundred. Each one solves something real. Each one matters enough to keep. But the infrastructure we have for relating to software — the mental model, the interface layer, the organizational logic — was designed for a world where a person uses maybe thirty products total, chosen carefully, maintained by other people, updated on someone else’s schedule.
App Store logic assumes scarcity on the supply side. A finite number of products compete for attention on a shelf. You browse. You choose. You install. You commit. The metaphor is retail. When supply becomes infinite and personal, that entire architecture collapses — not violently, it just stops being useful. How do you browse a shelf of things only you would ever need? How do you “install” something you made in a conversation and might remake differently tomorrow? How do you manage the lifecycle of a hundred artifacts that vary in how often you return to them, how stable their form needs to be, how much they should evolve versus stay fixed?
The numbers already reflect this pressure. App Store submissions surged 84% in early 2026 as AI coding tools matured. More products are entering the system than the system was designed to hold. But the real overflow isn’t happening on Apple’s servers — it’s happening in individual practices, in the space between what people are now able to create and the structures available to hold those creations.
This isn’t a discovery problem or a storage problem. It’s a paradigm problem. The container itself — the concept of “app” as a discrete, stable, individually managed object — is the thing that no longer fits.
I notice the instinct to frame this as a personal productivity challenge — “how do I organize my tools?” — and that framing is precisely what makes the underlying shift invisible.
The old mode: a product is built once, maintained indefinitely, used by many. Value accrues through adoption and retention. The product is the unit of meaning. The emerging mode: a form is generated in response to a situation, used as long as the situation holds, allowed to evolve or dissolve as conditions change. The situation is the unit of meaning. The product is just whatever shape the response takes today.
Between these two — between the fixed app and the disposable generation — there is an enormous unnamed territory. Most real usage lives there already. The monthly tool you return to three times. The workflow that stays mostly the same but needs to flex when a new project arrives. The prototype that hardened into something permanent without anyone deciding it should. We are already living in this territory. We just don’t have a name for it, or a structure that manages it intentionally rather than accidentally.
That’s what the next essay is about — the space between App Store and GenUI, and the concept I think makes it legible.
About the author
Qinqin (Stella) Yang is a creative AI builder and experiential creative director. She builds workflows, systems, and creative direction that operate under real constraints, and designs speculative frameworks that expand what teams can imagine before they start making. Over the past decade she has built brand experiences, product design, and speculative projects for Tencent, IKEA SPACE10, Burberry, and the Palace Museum. She is the founder of Actual Fiction, where she develops AI-native tools that turn imagination into a working method.
This newsletter is where she thinks out loud about creativity, AI, and what it means to build toward a future that doesn’t exist yet.
More at qinqinyang.com · Practice at Actual Fiction