Claude Design doesn't make you a designer, any more than Figma does

In December 2023, Brian Chesky announced that Airbnb had eliminated its product management function. At Figma Config, designers cheered.

You can understand why. Airbnb’s new model elevated designers to lead. The PM - long seen as the gatekeeper between business and engineering - was out. Designers were in.

This week, Anthropic launched Claude Design. Figma’s stock fell 6%.

That’s the thing about winning in a fast-moving industry. The advantage you celebrate today can become tomorrow’s problem faster than anyone expects. AI has compressed that timeline to the point where nobody feels safe.

Airbnb’s restructure was supposed to be a story about PMs. Two years later, it’s a story about something bigger.

When Chesky made his move, Airbnb had a real problem: the product had stagnated, the org had become fragmented and political, decisions were slow. His solution - a leaner structure with designers at the centre, a CEO who reviews everything, and two major launches a year - was radical. And it appears to have worked. Airbnb is shipping faster, and AI tools have accelerated that further.

Blowing up a fragmented, political organisation might have been the right call. But Chesky’s solution - making himself the decision point for almost everything - is exactly the HiPPO problem (Highest Paid Person’s Opinion) that Itamar Gilad warned about. He does have a stated vision, and Airbnb is shipping faster for it. But the fact that he still reviews almost everything personally suggests the vision isn’t specific enough for others to filter by. It works because of who he is. That’s not a model - it’s a person.

But here’s what the designers celebrating in 2023 didn’t see coming: the same AI wave that was supposed to sideline PMs is now targeting the execution layer of design itself.

Claude Design lets anyone go from a text prompt to a shareable, interactive prototype - reading your codebase and design files, applying your brand’s colours, typography, and components automatically. No Figma. No design ticket. No waiting two sprints.

In that sense, now everyone’s a designer.

Except they’re not.

This is where I think the “AI changes everything” narrative loses its way. The tools collapsing doesn’t mean the skills transfer.

Having access to Figma never made everyone a great designer. Having Claude Design won’t either.

As Karri Saarinen, Linear’s CEO, put it this week: “AI can generate interfaces. It still can’t figure out what feels right for humans.”

Left brain and right brain still matter. Some people have a genuine eye for layout, hierarchy, and how an interface feels to use. Others don’t - and no amount of tooling closes that gap. The collapsing of tools and roles gives everyone the ability to do more. It doesn’t mean everyone will be good at it.

We can see this already with writing, coding, and research. AI has made all of these more accessible. But the quality of the output still varies enormously based on the judgment of the person directing it. That won’t change with design.

This cuts through a lot of current AI noise: the tools becoming universal doesn’t make the underlying skill universal. We still have to pick a lane.

The same logic applies to how product teams work. Faster delivery means we can build the wrong things faster - Melissa Perri and Stephanie Leue have both made this point recently, and they’re right. Discovery still matters. If anything, it matters more.

But if roles are collapsing and PMs are prototyping, that frees something else up too. Engineers can go deeper into highly technical opportunities rather than routine build work. The spike model still holds: PM identifies an opportunity, engineering assesses feasibility - both moving in parallel, both owning their piece. What changes is who does what. But the work still needs an owner.

In tech, we’re told to keep our skills current. Stay up to date. Keep learning. That was always easier said than done. With AI, the pace has become almost impossible to track - let alone stay ahead of.

The lesson from designers celebrating in 2023 and watching Figma’s stock fall in 2026 isn’t that they were naive. Nobody anticipated this speed. The lesson is: don’t get too precious about how things used to work. Stay open - because the disruption rarely comes from where you’re watching.

Be ready to do more. Experiment. See what works for you.

But don’t feel like you have to master everything, and don’t keep chasing the next hype wave just to stay relevant. You can do more than you did before. But there’s a spectrum worth knowing: what you’re genuinely good at, what you can stretch to, and what’s beyond the line where quality drops off.

The tools don’t tell you where that line is.

Find where it is for you.

Product Notes is free and lands every week. Subscribe to get it in your inbox.

Subscribe now