Among the #Google I/O releases, I found myself excited about XR glasses. I think this is one of the most interesting (and growing) spaces for designers to explore outside of mobile. The next big opportunity might be AI on wearables. π§΅
We may be approaching that inflection point again, except the canvas is spatial and ambient. Maybe there'll be an app store for wearable hardware. Maybe modals, nav bars, and hero images are shifting.
Here's what I keep thinking as a designer: we've been here before. When the App Store opened and a generation of designers had to learn what interaction meant on a touchscreen.
Then something shifted. Meta embedded AI into glasses people already wore for fashion. 7 million units sold in 2025. Now Samsung, Google, and soon Apple are all entering the smart glasses race. The platform is getting serious.
But they didn't survive. I watched them get roasted on Reddit and YouTube. They asked people to change their mobile habits without offering enough value in return. Couldn't compete with the phone in your pocket.
Earlier this year, a friend let me try on the Meta Ray-Ban glasses. Very impressive. Since 2020, there have been many cool AI wearables β Humane AI Pin, Rabbit R1. Devices that made me go wow when I first saw them.
The best AI agent interactions shift the burden of explaining context from the human to the machine.
Google's Magic Pointer is one of the most exciting attempts at that I've seen.
Apple didn't invent multitouch gestures, but they popularized them by pairing the interaction with hardware people actually wanted to use. Google may be making the same bet with the Magic Pointer and the Googlebook.
But pointing inside Chrome isn't a real differentiator. The magic only happens when it works across your whole workflow: Figma, Claude Code, Adobe Premiere.
Imagine pointing between your Figma prototype and your coding agent: "fix that hover animation." The AI sees both.
The Magic Pointer lets you hover over an element, speak a short command, and the AI has all the context it needs to act.
No screenshot. No copy-paste. No re-explanation.
Right now, working with AI on a UI means constantly re-explaining your screen.
Screenshot β drag to chat β describe what you see β describe what's wrong β wait.
Each step is easy. Doing it 30 times kills your flow.
A digital companion can't save the frustrating moments in an agentic product. But it can make the frustration feel personal, even affectionate. And every pet someone posts online is free marketing OpenAI didn't have to pay for.
How upset are you when your coding agent messed up your code? How upset are you if your coding agent who messed up your code is a little Sam Altman goblin?OpenAI released Codex pets, and today, someone posted their Claude Code pets on Reddit (https://t.co/jURcAxJX7Y).
The custom pet doesn't solve that. But if you've spent time designing your own pet, you have skin in the game. Part of what you're watching is yours. You become more forgiving of the experience because now your own creation is part of it.
And the pet creator is a retention play. As someone who uses AI coding tools heavily, a big chunk of my time is just waiting. The agent runs, I switch tabs, check my phone, and I lose focus. Hours pass that way. I'm just supervising. The ROI on my time and mental energy feels low
Now, we are encountering AI agents, who (let's admit it) are more capable in many ways and can possibly replace what we do. The pixelated cartoonish pets make the powerful agents feel harmless and approachable. It masks our worries.
The retro-aesthetics has been in since the pandemic, where we felt our existence was in threat. We retreated to the aesthetics that's familiar to grasp onto the sense of certainty.
I think these pets are interesting signals that reveal our psychology with AI agents and the business incentives behind the agentic products. The Codex pets are hopping onto a well-validated retro-trend.