After OpenClaw’s meteoric rise, there was a flurry of people predicting that the future of computing would be nothing more than texting with an all-powerful assistant, or swarms of agents operating software themselves without needing a graphical interface.

“GUI is dead,” they confidently declared.

I couldn’t disagree more.

Visual interfaces exist for a reason. They’re often the best way to explore information, understand complex things, compare options, and take action. A map is better than a paragraph of directions. A chart is better than a wall of numbers.

Chat is a great way to express intent. It’s not always a great way to experience the result.

UI will no doubt change. LLMs and assistants will evolve beyond text-only messaging into a hybrid — conversation when it makes sense, rich visual (and interactive) interfaces when that’s the better way to understand something or get things done.

And as agents become more autonomous, visual interfaces will become even more important for another reason: helping us understand what agents are doing, monitor their progress, and intervene when needed.

The irony is that AI might actually lead to more UI, not less. Instead of navigating through fixed screens and workflows, we’ll increasingly see interfaces generated or adapted on demand, tailored to what we’re trying to accomplish.

We’re not quite there yet, but the direction seems clear.

UI isn’t dead. It’s becoming dynamic, personalized, and malleable.