Design links for September 24, 2026

Delegative UX is what lets a person see, track and understand work being done in their name, by an agent.

Curating the right context now matters even more than writing a good prompt.

AI is a boring technology. It's infrastructure — and the sooner we get used to thinking of it that way, the sooner we can stop being astonished and get back to making things.

We are proposing to make this a built-in capability of the web, and we'd love to know if this is worth pursuing and what the exact scope should be.

This week's episode is with Geoff Teehan, co-founder of the legendary design agency Teehan+Lax, longtime design leader at Facebook, and now Chief Design Officer at Lightspark.

A metaphor is a function between two worlds. A good one is a protocol: many ideas meet in one place.

A lightweight library full of beautiful loading indicators for React.

In 2026 the browser does all of this on its own. Here are twelve of those features, each with the workaround it retires, a snippet you can paste into a plain .css file, and where support stands.

Ten thousand hours in, my practice looks very different from the one I started with. Every client, project, and experiment has helped me understand what I'm good at, what gives me energy, and where those things are valuable to others.

It’s easy to agree that we should care about the details. It’s much harder to decide which feature we’re willing to leave out so we can finish the ones we’re keeping.

What started as an effort to refine our spacing tokens ultimately changed how we think about semantic tokens altogether.

If you want to have fluid typography, I recommend sticking with viewport units. This keeps the changes to the elements uniform across similar compositions.

After a couple of years designing agentic products, the thing that surprised me most is how little of the work is text. It's mostly decisions.

Go back to the fundamentals. Study them until you can apply them, then until you can see them — in your own work, in other people’s, in whatever a model hands you. Build the habit of naming what’s wrong and why.

On the value of cultural references, and the cost of losing them at scale.

As more interface work is AI-generated, the output of research and design shifts from documents written for humans to curated context that guides AI.

Evals can tell us whether AI respected the decisions we've already made. They get much less useful when we ask them to make the next one for us.

All 115 elements of the current HTML specification laid out as a periodic table: search, filters by spec section and links to MDN.

As execution gets cheaper, the new cost of creation becomes your judgment, your willingness to keep going, and your ability to see something worth making in the first place.

What worries me is that we have started to accept the smooth, flat average without questioning it.

Most companies will use AI to clear a faster queue. But I think there’s a larger opportunity: to turn a company’s point of view into a living system, then give its best creative people more time to make the kind of work no system can predict or generate.

AI can now generate summaries, themes, documents, even recommendations faster than ever – but how do you know if any of it is actually good?