The App in the Age of Mechanical Generation
We’ve seen this movie before. The technology changes, the protagonists shift, but the plot stays roughly the same: a powerful new tool promises to democratize creation. Barriers fall. Excitement builds. And then — the flood.
Today, tools let anyone generate an app in minutes. Not only prototype it, or sketch it — generate it. AI writes the code, designs the interface, and integrates the backend. In theory, it’s magic. In practice, I think it’s about to do to software what GarageBand did to music: make it easier than ever to create, and more complicated than ever to care.
As someone who spent the last decade building software , I’ve watched this moment coming. I’ve worked with great product teams, built systems used by millions, and mentored engineers as they grow into leaders. So I say this with some confidence: most software isn’t good. And soon, most of it won’t even be made by people.
Welcome to the era of mechanical generation.
From Creativity to Saturation
This isn’t the first time a creative domain has gone through this transformation.
It happened to music. Once you needed a label, a studio, and distribution. Now, a 16-year-old can record a track in their bedroom, push it to Spotify, and get a million plays, or none. The cost to create has dropped to nothing. The cost to stand out has skyrocketed.
It happened to content. Substack, TikTok, YouTube, everyone’s a publisher. Everyone’s a brand. But it’s never been harder to find quality writing or original ideas unless you already know where to look. You scroll through sameness hoping for a spark.
And now it’s happening to software. Anyone can spin up an app, drop in an AI-generated UI, and call it a product. You don’t need to code. You don’t need to understand system design. You need prompts.
But when everything is easy to make, the risk is that nothing is worth making well.
The Normalization of Mediocrity
The flood of AI-generated apps is on the way. And just like with music or content, most of it will be garbage.
Not because people are lazy or untalented. Because the tools reward speed over craft. Repetition over originality. Templates over thought.
Already, I see the signs: apps that look the same, behave the same, break in the same predictable ways. No consideration for architecture, scalability, or ethical implications. Just shiny frontends and brittle logic. Disposable software, built to impress in a demo and forgotten in production.
We’re not creating apps. We’re printing them. And we’re normalizing mediocrity in the process.
The Illusion of Empowerment
I don’t hate no-code or AI-assisted tools. Quite the contrary, they’re great for rapid prototyping, internal tools, or learning. But let’s be honest about what they are: accelerators. Not substitutes for understanding.
When we refer to this as “democratization,” we imply empowerment. But empowerment without education is a façade. It feels like progress — until something breaks and no one knows how to fix it. Until users lose their data. Until a security hole gets exploited.
We wouldn’t let everyone publish in a medical journal because they can write English. Why are we okay with everyone building apps just because they can click “generate”?
Benjamin Was Right (Even If Code Isn’t Art)
There’s a concept from Walter Benjamin’s 1935 essay, The Work of Art in the Age of Mechanical Reproduction, that’s been stuck in my head lately: the idea of the aura.
In his view, original works of art have an aura. A sense of time, place, and authorship. Mechanical reproduction strips that away. A print of a painting isn’t the painting. A film isn’t a live performance.
I don’t think code is art. (I’m not at all romantic.) However, I do think software has a kind of authorship, a trace of the person who created it. How they thought about a problem, the tradeoffs they made, and how they shaped the experience.
When AI generates apps, that aura disappears. The code isn’t thoughtful. It’s probabilistic. It doesn’t reflect a human’s intent. It reflects statistical patterns in a training dataset.
And just like we now have music that sounds like music but has no soul, we’ll soon have apps that look like software but do nothing meaningful.
The Risks Are Real. Structural
This isn’t just an aesthetic concern. There are real, structural risks that come with the mass generation of apps.
Security: When apps are spun up without understanding, they’re full of holes. Bad actors don’t need to hack systems anymore — they wait for auto-generated ones to implode.
Data Treatment: Privacy is often thrown around as a buzzword, but it is a real issue, especially if we think about how data is handled. Where is data stored? Who can access it? What happens if the app is abandoned? The average AI-generated app lacks effective data governance. There’s no thought around lifecycle, access control, or auditability. And worse: no one feels responsible, because the system wrote itself.
Sustainability: These apps are not well-maintained. They get abandoned. Platforms get clogged with dead projects. APIs go stale. The digital landfill grows.
We’ve built a world where anyone can create, but no one is responsible.
We’ve Flattened Everything
There’s something vaguely dystopian about the homogeneity of everything right now. Open Spotify, and half the songs sound like the same lo-fi chill loop. Scroll through an app marketplace, and every product is just a slight remix of the same CRUD app.
It’s like living in a communist software dictatorship: all output looks the same, speaks the same, does the same. And anything with real intention — a unique voice, a clear perspective, a challenging interaction — is pushed to the margins, treated as strange.
We’ve flattened aesthetics, flattened functionality, flattened ambition.
And the thing is: there’s still quality out there. But now, quality is niche. It’s not the norm. It’s buried.
What Comes Next
I’m not nostalgic. I don’t want to go back to the days when only engineers with CS degrees could ship a product. But I do think we need to recalibrate what we value.
Here’s what I think will matter in the next wave:
Taste: When everything is easy to make, taste becomes the differentiator. The ability to choose well, edit well, and say no to the default.
Curation: We’ll need new layers of filtering — people, communities, platforms — that help surface what’s actually worth using.
Hybrid creation: The best work will come from humans using AI deliberately, not blindly, as collaborators, not generators.
Accountability: We’ll need clearer standards around trust, data, and security. And we can’t just hope users figure it out.
Most people don’t want apps. They want outcomes. And I don’t think AI alone can get us there.
Generate and Create are not the same thing
We’re not drowning in creativity. We’re drowning in generation. And the signal — the truly valuable, thoughtful, and durable signal — is becoming increasingly more complex to find.
But maybe that’s the new job of the builder. Not just to ship fast, but to stand out. Not to generate more, but to mean more.
And if we’re lucky, maybe we’ll rediscover the aura in the process.