"I feel like an artisan shoe maker in the age of Nike."
AI is happening. It needs UX and product. The only way to join the party is to re-focus on our core value and skills, and leave the artifacts, processes and tools we've built over the past 20 years behind.
This being January, I figured I’d start with a prediction: one, in 2026 both product and UX will have to dramatically change because of AI, and two, you will be ok.
2025 was the year of Claude Code. I spent 20 years as an information architect, product and UX person, but most of my day-to-day now is back in software engineering. And for product and UX people, to get a sense of what’s coming, look at what software engineers are doing with AI. Our jobs will change, and there are lessons there.
So what changed just now?
One: the models got really good at two specific capabilities: reasoning and tool use. Reasoning is the capability of a model to “think things through”. You’ve seen this in the output of your favorite chatbot. Tool use is the capability of a model to call external tools (a search tool, for example), and incorporate its output. Reasoning and tool use work really well together, and when you join them in a loop, you get…
Two: …agents! Agents finally work. Here’s Boris, a (very!) senior engineer at Anthropic, who created Claude Code, in December 2025:
In the last thirty days, I landed 259 PRs -- 497 commits, 40k lines added, 38k lines removed. Every single line was written by Claude Code [...] Claude consistently runs for minutes, hours, and days at a time. (Boris Cherny on Threads)
We’re not “booking a flight” here, to use the typical 2024 “agent” example. We are doing economically valuable work, at scale.
And it can hurt. Engineers are already lamenting the good old days of “hand written code”.
“I feel like an artisan shoe maker in the age of Nike.” (Chris Arter on BlueSky)
OK Peter, but I am a product person, a UX person, not a software engineer. Why do I care? In the words of Ethan Mollick (whose newsletter is also a great follow!):
… these systems are actually broadly useful to knowledge workers of all types, and, by seeing what they can do (and experimenting with them yourself), I think you can learn a lot about the future of AI (Ethan Mollick on Claude Code)
Your practice will change. Most of the processes, artifacts, tools, processes, job titles we’ve developed over the past 20 years to build digital products were based on the idea that writing software is slow and expensive.
Agile. Figma. Team composition. Service Design. Lean. Squads. Storybook. User Journeys. DesignOps. Is that all out of the window? Not necessarily, but you should act like it is.
The foundations (understanding people, understanding the business, collaboration) matter more than ever, but the specific tactics were built for a world that is rapidly changing.
Over the past 20 years, we’ve built up a beautiful but complicated LEGO building of job titles, processes, tools, work artefacts and business models for a world that no longer exists.
My call to action in 2026 for the product and UX community is this: let’s put our LEGOs back in the box, step out of that room, and go hang out with the engineers, the marketers, the business owners, and most of all our customers, and build new ways of working together.
No fear. You got this.