There’s a familiar shape to how people discuss the future of software right now. On one side: the App Store paradigm — a finite supply of products, organized on shelves, adopted through deliberate choice, held for the long term. On the other: generative UI — infinite supply, created in real time for one person in one moment, used and discarded without ceremony. Nielsen Norman Group calls this outcome-oriented design: the user specifies a goal, the system generates whatever interface serves it, and the interface dissolves when the goal is met. Jakob Nielsen himself has called it the first new UI paradigm in sixty years.
Both ends are real. Both have substance. But most actual usage doesn’t live at either pole.
Think about the tools and workflows that fill a working week. A research format you return to every few days — not identical each time, but recognizably the same structure with variations. A project template you’ve refined across three engagements, stable enough to trust but still evolving at the edges. A small utility you made months ago for a situation that recurs unpredictably — sometimes weekly, sometimes not for six weeks, but when it comes back you want exactly that tool, not a freshly generated approximation of it.
These aren’t apps in the traditional sense. They weren’t chosen from a store, they don’t have version numbers, nobody else maintains them. But they also aren’t disposable generations — they persist, they have identity, they accumulate context from repeated use. They live in the middle, in territory that neither paradigm has a name for.
Three patterns that don’t fit
When this middle territory gets examined closely, three recurring patterns emerge that the App Store / GenUI binary cannot account for.
Intermittent repetition. A tool used once a month, maybe three times, sometimes not for two months, then suddenly essential again. Under App Store logic this is a product you’ve “retained” — but the metaphor breaks down because you didn’t really choose it from a shelf; it grew out of a moment and stuck around. Under GenUI logic it should be regenerated each time — but regeneration loses the context, the calibration, the refinements accumulated over previous uses. Neither framework handles the rhythm of irregular return.
Variant repetition. Each time you do a similar kind of work, the tool needs to be almost the same but slightly different. A presentation framework that adapts to different audiences. A writing environment that restructures depending on whether the piece is analytical or narrative. The form is recognizable across instances, but no single frozen version serves all of them. App Store logic would say: build a product flexible enough to cover all variants. But that’s exactly the kind of premature generalization that vibe coding made unnecessary — why engineer flexibility into one product when you can have multiple lightweight forms, each shaped to its occasion?
Evolutionary holding. Something you built three months ago that started simple but has been accumulating capabilities as your needs around it grew. Not through deliberate “feature development” — through the organic process of using it, noticing what’s missing, extending it in conversation. It’s thickening over time. At some point it may stabilize completely and become a permanent fixture. Or it may fork into two different tools. Or it may eventually fade as the project it served winds down. Its trajectory is not predetermined.
None of these patterns are edge cases. They constitute the majority of how people actually relate to software once production costs drop to zero. Yet every existing container — Custom GPTs, Notion templates, Apple Shortcuts, Claude Artifacts — handles them awkwardly, because each container was designed with either the “app” assumption (this thing is stable and you keep it) or the “generation” assumption (this thing is ephemeral and you remake it). The middle remains structurally unsupported.
A spectrum, not a binary
The shift I want to propose is simple but changes everything downstream: product and generation are not two kinds of things. They are two states on a single continuum. A form doesn’t have to be permanently one or the other — it moves.
I’ve been calling this the Solidification Spectrum. At one end: fully fluid. Generated in the moment, used once, dissolved. No persistence, no identity, no accumulation. Pure response to an immediate need. At the other end: fully solid. Stable form, fixed behavior, shared across users, maintained indefinitely. A traditional product. What it has always been.
In between: the entire territory of semi-solid forms. Things that have been used enough times to develop a recognizable shape, but haven’t frozen into permanence. Things that carry context from past uses but remain open to modification. Things that might solidify further — or might dissolve back into fluid — depending on what happens next.
The degree of solidification isn’t arbitrary. It correlates with two forces: how often a form is summoned, and how much reuse value it carries across instances. High frequency and high reuse value push toward solidification. Low frequency and declining relevance push toward fluidity. The spectrum is not a taxonomy you assign things to — it’s a dynamic that things move through.
Stewart Brand’s pace layers framework describes how healthy civilizations work: fast layers (fashion, commerce) innovate and experiment; slow layers (culture, nature) provide stability and memory. “Fast learns, slow remembers. Fast proposes, slow disposes.” The solidification spectrum is the same principle applied to software. Fluid forms move fast — they experiment, iterate, test new shapes. Solid forms move slow — they remember what works, they hold accumulated context, they provide the stability that makes the fast layer’s experimentation possible. A healthy software ecology needs both speeds, and needs forms to be able to transition between them.
The App Store paradigm took every form and pinned it at the solid end. Everything had to be a product. Everything had to be stable, maintained, installable. The emerging GenUI vision threatens to pin everything at the fluid end — everything disposable, everything regenerated, nothing accumulating. Both are impoverished. The full spectrum is what real usage actually looks like, and what any adequate new paradigm needs to support.
What this means in practice
If forms exist on a spectrum rather than in fixed categories, then the infrastructure around them needs to support movement — not just storage. A form that starts fluid should be able to thicken into semi-solid if it proves useful across multiple instances. A form that has been solid for months but hasn’t been touched should be able to soften, degrade gracefully, and eventually dissolve back into the pool of available capabilities. Nothing should be forced to stay frozen if no one is using it. Nothing should be forced to evaporate if someone keeps returning to it.
This is not how any current system works. Every existing tool assumes stasis: either the thing exists (installed, maintained, persistent) or it doesn’t. There is no designed pathway for gradual solidification, and certainly no designed pathway for graceful dissolution. Software rot happens — Lehman’s Laws described it in 1974 — but it happens as decay, as failure, as the thing you forgot to update. The spectrum suggests something different: dissolution as a feature, not a bug. A form that fades because it is no longer needed is not breaking. It is functioning exactly as designed.
What manages this? What watches the spectrum, tracks movement, decides when something has earned greater permanence or when something should be allowed to dissolve? That question is what the next essay takes up. Along with a concrete picture of what this looks like when it’s working.
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