In 2024, a product team at a large tech company shipped a feature in six weeks. Three of those weeks were design. Wireframes, user flows, three rounds of usability testing, a pixel-perfect handoff spec.
In 2026, a product team at an AI-first company ships a similar feature in four days. The engineer prototypes it live. An AI generates the variants. Design review happens on a working build instead of a Figma file.
Same outcome but wildly different process. And this shift is creating a paradox that most people in our industry haven't fully reckoned with.
The data
Here are the numbers that should make every design leader pay attention.
Lenny Rachitsky published his biannual job market report last week. Engineering openings are at 67,000 globally. PM roles are at a three-year high. Design? Flat since 2023. About 5,700 open roles worldwide.
That plateau has lasted three years while every other function climbed.
The same week, Jenny Wen, head of design at Claude, went on Lenny's podcast and said something that ricocheted around the design community: "The design process that designers have been taught, we sort of treat it as gospel. That's basically dead."
And Nielsen Norman Group released their annual State of UX report with a title that reads like a thesis statement: "Design Deeper to Differentiate."
Three signals in one week, all pointing in the same direction.
What's actually happening
The easy conclusion is that design is shrinking. That AI ate it. That engineers don't need us anymore. That product managers can do our jobs.
The reality is more interesting. Design is changing along two dimensions at once.
The first is *production*. The part of the job that was about building specs, pushing pixels, and creating handoff documents is being absorbed by AI tools and by engineers who now move fast enough to prototype in code instead of waiting for mocks.
The second is *exploration*. A product manager or an engineer can now spin up a wireframe from your design system in minutes. They look at it and think: done. The design work is finished.
Except it isn't. What they've created is an *iteration*. And iterations are easy. Any function can produce a plausible-looking screen now. That doesn't mean the screen is usable, optimized, or worth building. It means someone assembled components. That's collage, not design.
If your value as a designer was primarily in production or basic exploration, yes, the math is harsh. If AI lets one designer (or a PM) do the work that used to require four, companies will employ fewer of them. That explains the flat job numbers.
But here's the paradox: the need for genuine design thinking and craft is becoming more critical, not less.
Same as it ever was
The best designers have always worked across every phase of the product development lifecycle. This part hasn't changed. What's changed is that more people are noticing.
A great designer improves product utility alongside their PM. They push for superlative execution with their engineering partners. And yes, they drive the work that is traditionally owned by design: the explorations, the iterations, usability assessment, task flow analysis, design specification, and primitive prototypes.
The difference now is that these boundaries are becoming explicit. The best designers I work with can automate the creation of a PRD. They understand where to focus the team's efforts, not just what the interface should look like. At some of the most cutting-edge companies, designers lead project initiatives end to end.
The same thing is happening on the engineering side. We have designers checking their work directly into pull requests, committing code to GitHub. They aren't replacing engineers. Their work simply overlaps.
When engineers can build anything in days instead of weeks, the bottleneck shifts. It used to be: "Can we build this?" Now it's: "Should we build this?" And: "Of the forty things we *could* build this sprint, which three actually matter?"
That's a design question. It always was. We just didn't notice because the old bottleneck, production speed, masked it.
The new model
The old model treated design as a *phase*. It happened between strategy and engineering. Designers owned a step in the process, and their power came from controlling that step.
The new model treats design as a *lens*. Applied continuously, by everyone, at every stage. Designers don't own a step. They own a standard. Their power comes from raising the bar of what "good" means across the entire product.
I lived this at Airbnb. I was the GM of the Guest team, which was responsible for 97% of the company's booking revenue. The role required me to think like an owner: understand the business, learn the basics of finance, develop vision and strategy, and deeply understand the customer. None of that was traditionally "design work." All of it made me a better designer.
NNGroup's report nailed it: "UI is still important, but it will gradually become less of a differentiator." When every app can look polished, because AI makes polish cheap, the differentiator moves deeper.
Deeper into craft. Deeper into soul.
Here's the thing about a PM or an engineer spinning up wireframes from a design system: the output is functional. It works. But it doesn't have the details, the nuance, the moments of delight that make someone actually love a product. It's a wireframe with a coat of paint. It doesn't breathe.
If you're building purely functional software, something internal, something that doesn't need to compete, something where nobody cares whether it sparks joy, then sure, have a PM generate some wireframes and ship it. That's fine. That's also what we're starting to call AI slop.
But if you want to build software that people love? That's a different game entirely.
I learned this at Slack. Slack is an enterprise product. Enterprise software is supposed to be boring. But people *loved* Slack, and it wasn't because the feature set was radically different from the competition. It was the craft. The micro-interactions. The personality in the copy. The way the loading screen made you smile instead of making you wait. The feeling that someone who cared had touched every surface.
Compare Slack to Microsoft Teams. Teams is functional. It does the job. But it has no soul. Nobody has ever described Teams as lovable.
I saw the same pattern at Airbnb. Airbnb and Expedia solve roughly the same problem: help people find a place to stay. Expedia is functional. It works. But Airbnb made you *feel* something. The photography, the storytelling, the way a listing felt like a window into someone's life rather than a row in a database.
The great designer is also the best craftsperson and has the best taste. They bring the soul to the product. They can spot AI slop from a mile away, and they are entirely allergic to it. They take something functional and turn it into an object of love.
That, at least in the near term, is not something I see changing.
What's next
So what does this mean?
For design leaders: Stop defending the old process. The designers clinging to their five-step diamond framework are fighting for territory that no longer exists. Let it go. Instead, invest in your team's judgment and taste. The scarcest skill in a world of infinite production capacity is the ability to look at something and know, quickly, intuitively, correctly, whether it's good. Embed your designers everywhere. A design team that reviews work after the fact is too slow. Designers need to be in the room, in the pull request, and in the prompt at the moment decisions are being made.
For IC designers: Learn the business. The best designers I've worked with understand revenue models, customer acquisition costs, and retention metrics. They don't just know what looks good. They know what *works*. Expand your toolkit. If you can write a PRD, if you can commit code, if you can facilitate a strategy session, you become the kind of designer no AI can replace. And above all, sharpen your craft. In a world flooded with AI-generated "good enough," the designer who obsesses over the details, the transitions, the micro-copy, the moments between the moments, becomes exponentially more valuable. The gap between functional and lovable is where your career lives now.
For everyone: Go deeper. Surface-level design (layout, color, typography) is increasingly commoditized. The people who thrive will be the ones who understand systems, incentives, edge cases, and human psychology at a level that no prompt can replicate.
There's a pattern in the history of technology worth remembering.
When desktop publishing arrived in the 1980s, people predicted the death of graphic design. Anyone could make a newsletter. Why hire a designer?
What actually happened was the opposite. The flood of amateur design made professional design *more* valuable, not less. When everyone could make something, the gap between "something" and "something great" became the most important gap in the market.
We're at that moment again. AI is the new desktop publishing. It democratizes production. It commoditizes the average. And it makes the exceptional matter more than ever.
The design process is dead.
Long live design.