We are entering a phase where the interface is no longer visual, but conversational. It is not about screens anymore. It is about systems. And right now, ChatGPT feels like the front door to everything.
What started as a clever assistant is quietly becoming an operating system. The recent integrations with Gmail, Docs, Calendar and other tools are not just features. They are signals. ChatGPT is no longer just something we talk to. It is becoming the layer we work through.
You could call it ChatGPT OS. And it changes everything.
From UI to AI
Traditional UI has felt stagnant for a while now. Add a few micro-animations, optimise a flow, reuse the same component library. It is useful, but it rarely feels new.
Now imagine skipping the UI entirely. Just say what you need. Pull data from one place, summarise it, draft a reply, send it. All within a single thread. The GPT becomes the interface.
This does not mean design becomes irrelevant. It means design moves deeper. We need to think about intent, memory, context, and connection. We are not designing buttons. We are designing inputs, integrations, and trust.
Voice, screens, and discomfort
Voice feels like the most obvious interface. Yet it still feels awkward in practice. Not because it is a bad idea, but because the execution is not there yet. These systems do not understand when to pause, when to interrupt, or when you are thinking.
And personally, I do not want to speak to my computer in public. I do not like the idea of people hearing my commands, or listening to my conversations. It feels exposed. It feels unsafe.
So while we wait for something like Jony Ive and Sam Altman’s mystery device, we are left with screens. Which brings me to a very practical frustration.
Why is the ChatGPT editing experience so cramped?
When working on longer content, the side-by-side view becomes messy. It forces output into places that do not make sense, and it decides on the layout for you. Claude gets this right with a clearer layout. But ChatGPT still struggles with space.
If we are serious about this being an OS, then it needs an environment. Imagine a widescreen mode with two clear panels. One for conversation. One for output. Nothing overlapping. Nothing shifting. Just enough space to breathe.
Building the invisible layer
The most exciting part of all this is the invisible layer of integration. What happens when you can say, ‘Add my Gmail here,’ and it just works?
Or will it require onboarding? Authentication steps? A smart way to manage permissions inside the conversation?
This is where the work gets interesting. It is no longer just product design. It is system design. We are building experiences across layers of intent, security, and automation.
That is the side of AI I want to be on. Not the prompt-generated visuals. Not the surface-level gimmicks. But the deeper work that shapes how we engage, how we connect, and how we build trust.
It feels like a natural progression. We started with how things looked. Then came the shift to why they mattered. Visual design gave way to design thinking. Over time, we embraced storytelling, systems thinking and service design. The work expanded beyond screens to include entire experiences and ecosystems.
Now we are entering a new phase. One where the interface disappears and the challenge becomes architectural. We are designing logic, flow and context inside systems that respond, adapt and learn. Every day brings a new interaction to define, a new constraint to navigate, and a new opportunity to shape how people work with technology, not just through it.
This is not the end of design. It is the next surface we design for.
Are you ready to build it?