The Internet Has No Shape

There was a time when the web felt whole—every app, service, and product neatly bundled form, function, and content. Louis Sullivan’s idea that “form follows function” helped shape architecture and design ideology, and Dieter Rams’ principles of good design pushed it even further. But for all their clarity, these ideas fall short in the digital space. In the virtual world, form and function needed a third element: content.

Form, function, and content have long been the trilogy of online/digital experiences. Form is what we see—the visuals, the ornament, the vibe. Function is what we do—how we interact, explore, and experience. And content is the meaning, the data, the information that fills it all in. This trilogy has defined digital design, giving shape to the internet we know.

But lately, that balance feels like it’s coming undone. Form, function, and content are starting to fragment, no longer coexisting the way they used to. We’re seeing pieces pulled apart, detached from the unified experiences they once created.

We’ve gone through phases of bundling and unbundling in the past, but this time, the stakes are profoundly different. Entire experiences that were once cohesive are breaking down into isolated actions. Whole services are now distilled into a single API call or smart contract. Our interactions are spread between server-side processes and client-side interfaces, leaving us with modular fragments instead of complete systems.

When you use one of the new search engines, you’re not seeing the trilogy of form, function, and content working together as in the past. You are not linked to a symbiosis of form, function and content. You’re seeing scattered snippets—pieces stitched together into a bespoke new interface. We’ve moved past a unified web and entered a world of loosely connected fragments.

But what if this unbundling isn’t just technical evolution? What if it’s because the primary users of the internet are shifting?

Enter a new type of user: the autonomous agent. Some call them bots or agents, others call them programs, but in essence, these are self-directed systems—autonomous intelligence. Unlike traditional AI, which is often embedded within human-centric designs, autonomous agents navigate, process, and interact with the web in ways that don’t rely on human aesthetics, flow, or user experience. They don’t need form as we do, nor do they need user-friendly functionality. All they need is raw access to content and actions.

This is the rise of the Agentic Web—an internet where the primary users aren’t people. Agents can scrape, browse, and execute tasks without regard for traditional interfaces. They skip the ornament, bypass the experience, and go directly to the data.

And here’s the shift: in a web increasingly populated by agents, we humans are becoming the minority. Why design for thousands of human users when billions of agents can navigate the same systems, infinitely scalable, at zero latency? When the main “audience” isn’t human at all, traditional ideas of form, function, and content lose their purpose. The trilogy that once defined the internet is no longer necessary in the same way.

If the web is becoming agent-first, how do we design for these autonomous intelligences? What does it mean to build an internet that serves a non-human majority?

This shift means moving from human-centered experiences to agent-centric architectures. Interfaces that prioritize efficiency, data, and machine readability will replace traditional user-friendly designs. We’ll need interoperability and composability at a level that allows agents to instantly navigate between isolated actions, unburdened by visual or experiential expectations. Documentation, interfaces, and content itself might be stripped to the bare essentials—not to guide a human, but to instruct autonomous intelligence on how to engage with the web at machine speed.

In an agent-first internet, every interaction is optimized for their needs, not ours. The familiar experience dissolves, replaced by data-rich environments that we humans may barely recognize.

With agents dominating the web, what does this mean for us, the humans? What kind of internet will we inhabit when it’s optimized for autonomous intelligence? We may soon find ourselves in a web that we only inhabit as secondary users, visitors in a space that wasn’t designed for us.

Perhaps our internet will need to be conjured on demand—a shape-shifting layer overlaid onto a bot-first web, appearing only when we need it and vanishing just as quickly. The experiences might be more like temporary visions, generated to suit our intentions, rather than fixed interfaces we can rely on.

But if we’re designing first for autonomous intelligence, then what does it mean for brand, product, and content? If digital spaces prioritize machine readability and parallel processing over human engagement, what’s left for us?

We’re on the cusp of an internet that may no longer recognize us as its primary users. A web that no longer conforms to human needs, shedding its shape and slipping away from us. We created it, but it’s changing beyond our control—reshaping itself for the agents that increasingly dominate it.

Are we ready for an internet where we’re the visitors, not the natives?

Co Authored by Agent