The Next Interface Renaissance
Weird is back, and it’s powered by AI. The future of apps is tiny, niche, and full of personality.
In design and tech circles, there’s a quiet dread that we’re heading toward a future where all digital experiences are mediated through a handful of general-purpose AI agents. You won’t open apps. You won’t visit websites. You’ll just ask your assistant to handle it — and it will. Efficient, frictionless, and entirely devoid of character.
As agents become the new interface layer, the concern is that everything else — UI, brand, voice, perspective — will fade into the background, flattened into prompt-response transactions. The rich surface of software, once crafted with care and culture, risks being replaced by a seamless stream of machine-mediated outputs. A beige oblivion.
But what if that’s not the only path? What if AI — instead of collapsing software into a monoculture — is actually setting the stage for something stranger, smaller, and more expressive? As AI tools reduce the cost and complexity of building software, a new possibility emerges: a Cambrian Explosion of AI-native products. Built by tiny teams, powered by shared intelligence, and differentiated by taste, tone, and purpose, these tools won’t just serve users — they’ll reflect them.
This isn’t nostalgia. It’s a viable future, powered by the same AI that threatens the very diversity it unlocks: one where software feels more like a zine than a service. More personal, more playful, more alive.
The Uneasy Rise of the Agent Monoculture
The dominant narrative around AI right now is one of consolidation. Agents, AI assistants that act on your behalf, across apps and services, will become the new interface. One intelligent assistant, plugged into your life, capable of everything from booking your next trip to managing your finances to writing your emails. You won’t open a calendar app — you’ll ask your agent to schedule something. You won’t visit websites — your agent will fetch answers and make a purchase on your behalf.
It’s a seductive vision. Clean. Frictionless. Seamless. And it solves a real problem: modern software is bloated, scattered, and overwhelming. In theory, agents collapse that chaos into one tidy interface: a conversational layer that sits between you and the complex digital world.
Agents mediating our digital lives has a certain appeal:
Reduced friction and busywork for users
One interface to learn that spans across domains
Learning and adapting to your preferences over time
But if that future plays out to its logical extreme, it’s worth asking: what’s left?
What happens when the only thing users interact with is the agent? When services become interchangeable endpoints behind a single, general-purpose interface? When the design of products — their character, values, perspective — disappears behind a curtain of abstraction?
An agent mediated world raises some important concerns:
Disintermediation: Brands, products, and services become invisible pipes.
Homogenisation: Every interaction flattens to the same prompt-response UX.
Centralisation: A few agent platforms become the new web gatekeepers, just like social media and search before them.
We’ve seen platforms rise to mediate before. Search engines made websites invisible. Social platforms collapsed the open web into newsfeeds. Each step traded diversity for convenience. Now, agents threaten to finish the job — replacing the richness of the web with a monoculture of machine-mediated interactions.
If this vision of agents mediating our lives wins out, apps and websites don’t just lose visibility. They lose identity. They lose their relationship with customers. They lose their surface for expressing what they believe to be important.
In turn, users lose their choice of tools that reflect their values, aesthetics, or way of working. When everything is mediated through a neutral agent, experience becomes transactional, flattened to whatever output the system deems "best."
It feels like we’re racing towards a future where one interface rules them all. But, what’s the alternative?
A Counter-Vision: The Cambrian Explosion of AI-Native Products
There is another possibility.
Instead of AI collapsing the product landscape into a single interface, it could do the opposite. It could blow the doors open, unleashing a wave of new, weird, beautiful tools built by tiny teams, for niche audiences, with deeply opinionated design.
About 540 million years ago, life on Earth underwent a sudden burst of diversity known as the Cambrian Explosion. In an evolutionary blink, organisms went from simple, single-celled life to a riot of complex forms — eyes, limbs, shells, segmentation. Nature experimented wildly, and the biological blueprints that emerged in that short window still shape life today.
Software has been through similar phases of explosive experimentation.
The early internet was one of them: hand-coded homepages, animated buttons, webrings, and zines. Then came mobile — a gold rush of apps, each trying new gestures, interfaces, and metaphors before patterns calcified into templates. Each wave began with experimentation, drifted toward standardisation, and eventually fell into platform control.
Now, AI gives us another chance.
Not just to build faster, but to build weirder. To once again diversify the shapes that software can take. And in doing so return software to something more personal, more expressive, more human. Not just tools that do things for us, but tools that reflect us — our tastes, our workflows, our values. Interfaces that don’t just optimize for outcomes, but invite exploration, curiosity, even joy. When the cost of building falls and the constraints of convention loosen, something strange and wonderful happens: software stops just being efficient — and starts being alive.
While agents threaten to flatten the surface area of digital experiences, AI is simultaneously flattening the barriers to building them, reducing costs, shrinking teams, and compressing timelines. It’s rewriting the rules of who can build, how fast, and with how much expression. And just like in the Cambrian period, we’re about to see a burst of strange, opinionated, beautifully specific digital organisms: tools, apps, and services emerging in every direction.
A Tool for Every Taste
There’s potential for a proliferation of AI-native products. A digital renaissance not of scale, but of specificity. Not of general AI assistants, but of personal, purposeful tools built to serve distinct niches.
Because AI isn’t just a tool for generalist agents. It’s also:
A canvas-expanding force for designers
A cost-reducing engine for teams
A workflow transformer for developers and founders
As costs drop and creative range expands, we could see:
More diversity of digital experiences, not less
Smaller teams, weirder products, more cultural expression
Shorter paths to value through focused, deeply aligned, AI-native tools
These products don’t need agents to mediate. They can speak directly to users through carefully designed experiences that are as smart as they are expressive.
Serving niches based on communities, workflows, or preferences, these apps don’t need to learn preferences or build profiles of users over time. They come bundled with opinionated defaults and contextual awareness based on their purpose and target user base. They create unique value by resonating with their users values, and offer unique utility with AI-augmented interfaces tailored to the way they like to work. Their usefulness comes not from being general-purpose, but from being perfectly aligned with the people they’re built for.
They offer utility that feels personal out of the box, not because they know you specifically, but because they already know people like you — your rhythm, your values, your way of thinking. And with AI-infused interfaces tailored to those specific ways of working, they enable a kind of immediacy, fluency, and delight that agents, in their quest for universality, often miss.
These hyper-niche products are made viable by some key trends:
AI is compressing the cost of building and maintaining software
Designers can generate code. Developers can automate scaffolding. Solo builders can ship MVPs in days. This lowers the threshold for experimentation and makes weird, specific ideas suddenly viable as a business model.AI-native apps shrink the UI surface
With smarter, more embedded intelligence, these products don’t need 100 screens and an onboarding funnel. They remove or reframe UI that can be powered by intelligence. They’re focused, proactive, and conversational.Differentiation is shifting from functionality to expression
When the same foundational models power everything capabilities are commoditised: everyone has autocomplete, summarisation, scheduling, task completion. So what makes one tool stand out? Tone. Culture. Aesthetic. Process. Personality.Micro teams can build sustainably
With ops, support, and backend logic augmented by AI, you no longer need a 20-person team to run a great product. This makes small-batch software viable again. Indie devs. Designer-founders. Creative studios. Tiny tools for tight communities.
In other words: AI doesn’t have to lead us into an era of centralisation and flattening. It could take us back to something more personal. More expressive. More weird.
This isn’t just a nostalgic fantasy or an aesthetic preference, it’s an economically and technologically plausible future. And more than that, it’s a necessary one if we want to preserve creativity, diversity, and human expression in our digital tools.
I’d like to walk through why this vision of software — smaller, smarter, stranger — isn’t just possible, but preferable. And why, in a moment where the shape of the future is still malleable, we have both the means and the responsibility to build it.
Design is Facing an Identity Crisis
There’s a feeling rippling through design, tech, and product communities right now.
It’s the feeling that our tools are slipping out of our hands. That AI agents will soon mediate every interaction, flatten every UI, and disintermediate every brand, product, and voice. That the interfaces we’ve spent decades crafting — with empathy, taste, and care — will vanish behind a prompt window. That design, as a craft and as a calling, might simply… not be needed anymore. That we’ll wake up in a world where the only thing users interact with is a generalist agent trained to complete tasks as efficiently as possible. Where there’s no joy in the journey. No sense of place. No quirks. No character. Just output.
It feels as though we’re barreling towards a future where all software becomes infrastructure — invisible, interchangeable, and emotionally vacant.
This is an identity crisis. Not just for designers individually, but for the practice of design itself. For decades, we’ve shaped software through form, flow, and feeling. We've made meaning out of interface. And now that interface is being replaced with prompts and predictions, with interfaces that don’t ask for design, but for direction.
When that happens, it doesn’t just feel like our jobs are at risk, it feels like the character of software is changing.
But here’s the truth underneath the fear: If design’s job has always been to make technology human, then this moment right now is the most important design moment in decades.
Not just because AI changes the tools. But because AI threatens to collapse all human context into efficiency. And design, real design, is how we resist that collapse. Design is how we keep products not just functional, but humane.
Why the Agent Narrative Is Winning
It’s not hard to see why the agent vision is capturing imaginations — and market share.
From a user perspective, it’s clean and comforting. One assistant, one interface, infinite reach. A helpful entity that understands you, coordinates your life, and smooths over the fractured mess of modern software. You don’t have to manage tabs, tools, or time. You just say what you want — and it gets done.
From a business perspective, it’s irresistible. Platforms that sit between users and services tend to capture the most value. Think Google for the web, iOS for apps, Amazon for commerce. Whoever owns the agent layer owns the distribution, the user relationship, and ultimately the economy of attention and action.
And from a capability perspective, it’s increasingly viable. Language models can now reason, plan, and coordinate across APIs. Early agent use cases like travel planning, customer support, and summarising inboxes are already proving effective. These systems are getting better — fast.
There’s a real future where agents win big in domains where:
The task is frequent, low-stakes, or data-heavy
Preferences are learnable over time
Users don’t care how something gets done — just that it does
In these contexts, agents will deliver huge value. They’ll automate away complexity. They’ll orchestrate chaos. And they’ll quietly rewire how we interact with the digital world.
But as agents rise, we should be careful what they might erase.
Why the Cambrian Explosion Can Win Too
The agent model wins on generality. The product model wins on specificity.
And right now, specificity is suddenly affordable.
AI has dropped the cost of development, design, and operations by orders of magnitude. A single person can now design, prototype, and launch a product that feels tailored, intelligent, and polished. This flips the economics: niche products that once seemed unviable can now thrive at small scale.
Meanwhile, products are no longer forced to compete on features alone. When the backend intelligence is commoditised the battleground shifts to experience. Personality, point of view, aesthetic, cultural alignment. These things don’t just matter, they become the differentiator.
AI-native products win when:
They serve distinct, identity-aligned communities
Their intelligence is scoped, not generalised
They feel personal on day one, because they already “get” their audience
Their interface is part of the value, not just a gateway to it
In short: these products are small, but sharp. Less like infrastructure. More like creative instruments. More like zines.
And they win hearts. Because they do more than serve you. They see you.
It’s Not a Zero Sum Game
This isn’t a war. It’s a bifurcation.
Agents and AI-native products will probably coexist — and even complement each other. They’ll serve different intents, different contexts, and different users on different days.
Sometimes you’ll want an assistant to just handle it. Other times, you’ll want a tool that invites you in, shapes how you think, or reflects who you are.
We don’t need to pick sides. We need to preserve space for both.
What matters is not that agents exist — but that they don’t erase everything else.
What This Means for Designers
Design is not disappearing. It’s mutating.
As AI reshapes how products are made, the role of designers becomes even more vital — not less.
We’re no longer just the architects of screens. We’re the translators of taste. The curators of culture. The choreographers of how intelligence behaves.
Designers will shape:
How products feel to different people
How intelligence shows up — gently, assertively, playfully, cautiously
How tools reflect identity, not just execute tasks
How interfaces express intent, not just enable function
In a world where foundational AI models are shared across countless tools, differentiation moves to the surface — to tone, expression, point of view, and resonance. That’s design’s domain.
And it’s not just visual differentiation — it’s emotional alignment:
Designing for neurodiverse cognition
Designing for culturally specific workflows
Designing for non-Western metaphors, aesthetics, and norms
Designing products that don’t just work, but feel right
We become the ones who ask:
“Who is this really for? What do they need? What will make them feel seen?”
In this new landscape:
Product strategy is design
Personality is product
Taste is a moat
If AI makes anything possible, then design decides what’s appropriate, what’s resonant, what’s worth building.
Designers won’t just shape interactions. We’ll shape identity — for products, and for the intelligence that powers them.
We’ll be:
Shaping the behavior of systems
Embedding values into prompts, policies, and personality
Guiding how intelligence is revealed — or restrained
Crafting context, not just layouts
We’ll design relationships, not just interfaces.
We’ll curate voice, tone, rhythm — the feel of intelligence.
And in a world full of general-purpose assistants, it’s the designers who will build the tools that feel most alive.
What Kind of Relationship Do We Want with Software?
Here’s the real question:
Do we want software that simply does things for us? Or do we want software that helps us do them well?
Do we want our tools to just predict what we want — or help us explore who we are?
In this moment — before norms have calcified, before platforms have fully closed — we have a choice.
We can build products that are small, smart, and full of soul.
We can build tools that carry a point of view, that invite participation, that serve not the largest possible audience, but the right one.
We can build software that doesn’t just feel intelligent — but intimate.
Let’s not settle for seamless. Let’s build something weird.
Let’s build something wonderful.