We stand on the threshold of a fundamental shift in how humans interact with the digital environment. The familiar windows, icons, and menus that have served as a bridge between us and machines for decades are rapidly losing their monopoly. A new paradigm is emerging—the "zero interface" era, where the screen ceases to be a mandatory intermediary, and the system itself understands our intentions. This is not just evolution, but a change in the very philosophy of interaction.
From punch cards to sensors: a brief history of intermediaries
The journey from machine language to human language was long. First, punch cards and the command line required users to speak the computer's language. Then, in the 1970s, Xerox PARC gave the world the WIMP concept (windows, icons, menus, pointer), which Apple successfully commercialized. And in 2007, the iPhone's touchscreen made the mouse and keyboard anachronisms, allowing direct control of the device through touch.
However, it was at this moment that the graphical interface reached its ceiling. Superapps grew to incredible scales, their navigation became a labyrinth, and the simplicity for which it all began was sacrificed for functionality. The interface, meant to simplify, began to create "friction"—extra steps and cognitive load.
What exactly is killing the old UI?
I see three key factors that have led to the decline of traditional interfaces. First, there is utility overload. Photoshop has turned into a "cockpit," and Excel requires advanced math knowledge for complex spreadsheets. Second, friction: turning on a light through an app takes longer and is more complicated than pressing a physical switch. Third, and most importantly, the development of AI. Large language models have learned to understand user intent, not just execute commands. Jakob Nielsen rightly calls this the first new UI paradigm in 60 years.
Post-interface: voice, agents, and ambient computing
The new era is not the absence of an interface, but its radical transformation. I will highlight several key directions. The conversational layer (voice assistants) turns navigation into dialogue. Agentic systems delegate task execution: you set a goal, and the AI itself books tickets and manages your calendar. Ambient computing embeds technology into the environment, responding to context without a request. Spatial computing (VR/AR) brings data into three-dimensional space, while biometrics and neural interfaces allow controlling technology with the power of thought.
The flip side of the coin
Blindly abandoning screens is a dangerous illusion. We already see failures like the Humane AI Pin and Rabbit R1, which tried to replace the smartphone with a "zero" interface but failed due to slowness and limitations. The main problems of the new approach are discoverability (the user doesn't know what the system can do), loss of control and trust (it's unclear why the AI made a particular decision), and a shift in cognitive load (now you need to know how to formulate requests correctly).
My conclusion as an analyst: the future lies in hybrid models, as with GitHub Copilot or Perplexity AI, where AI complements rather than replaces the graphical interface. A complete rejection of buttons is justified only in narrow, predictable scenarios. In other cases, we need a minimal layer of control to maintain trust in technology. We are moving from an era of learning the language of machines to an era of teaching machines to understand us, but this transition will be gradual and without radical revolutions.