April 6, 2026

Tips for Designers on Making it Through 2026 to the Other Side

If you’re in tech and you don’t have at least a little career anxiety right now, you are either overly confident or you aren’t paying attention. Engineers are jittery. PMs are jittery. Designers and Researchers are jittery. The simplest way to explain what’s happening is:

The assembly layer is going away.

The assembly layer has been around since the beginning of digital product development. For engineers, it includes things like writing a function that calls an API and stores the results in a database. For PMs, it includes culling through experimentation data and collecting significant results into a slide deck. For designers, it includes specifying a button hover state for the thousandth time. For researchers, it involves going through open-ended survey results line by line and tagging each result.

“White-collar assembly work” is the digital equivalent of all of the various trades that go into building a house. It is not the conception and planning of the house, but rather the drywall, flooring, electrical, and dozens more essential, skilled functions. This work is noble and important, but in the digital world, a lot of it is rapidly being eaten by AI.

The next question to ask yourself then is am I an assembler?

I would argue that at least in design, most of us are some percentage assembler. It could be 5%, it could be 100%, but it’s rarely zero. I run the largest design team at Microsoft AI, and even I spend some of my time nudging things around in Figma, trying to make decks look sharp, and performing some truly lizard-brain level data entry.

Orchestral Manoeuvres in the Dark

In order to prepare for a world where the nature of the work is changing so dramatically, Microsoft and many other companies are moving towards more of an Individual Contributor empowered environment. Some people assume this means all managers are going away. This is not the case. What it actually means is:

  1. A true dual-track career path. Tech companies have been saying managers and ICs are rewarded evenly for 15 years now, but it’s often not the case in practice.
  2. A requirement that every hire, whether IC or manager, is deeply involved in the product-making details.
  3. A shifting of decision-making to the people closest to the work, where context is richest.

When people think of an orchestra of Individual Contributors, they picture everyone as a violinist. Perhaps you, as a lifelong IC, consider yourself a violinist.

What it really means though, is that everyone is a conductor now. It is your job to wave your arms expertly, and the machines will play the violins. As many violins as you want. And trumpets. And pianos. And sousaphones. Even in the dark, while you sleep, if you so desire.

What some companies are getting wrong right now is assuming those conductor jobs are mainly for PMs. Perhaps that is what’s behind this odd looking State of the 2026 Job Market survey. The data comes from a company called TrueUp, which I hadn’t heard of until today, but that’s probably just a me thing. Perhaps the data reflects reality, or perhaps design jobs aren’t accurately tracked by this company, but either way, this is not what I or a lot of my colleagues at other companies are seeing. If anything, most cross-functional teams are more underwater on design than on other functions.

Don’t You Forget About Me

Regardless, one thing that has absolutely changed is what we look for in the designers we interview. If you are in the job market now or will be at some point in the future, this post is designed to help you shape yourself into a better candidate for the design jobs of today and tomorrow.

As I first mentioned a year ago in The Future Favors the Curious, we are at a moment in time when orientation towards the future may supersede hard-earned experience with the past. Decades of wisdom in design can be a wonderful foundation on which to build even more expertise, but it can also mean unhelpful dogma about what the product-making process must look like. What’s helpful? A chicken-sexing like feel for identifying experiences that will and won’t resonate. What’s unhelpful? An insistence that static redlines be scrutinized by multiple levels of leadership and then handed off to engineering.

Some parts of AI-native product-making require rewiring your brain to know what is helpful and what is wasteful. How often have you seen a flawed project that an entire cross-functional team has worked on for weeks or months only for you to say “ugh, we could have saved thousands of person-hours if you did X, Y, and Z earlier”? How would your reaction change if you saw the same work but an enterprising designer got it working with live code in a single afternoon? “What a giant waste” turns into “That’s a helpful starting point!”

I should mention that the goal here is not to build a company full of product-makers who collaborate entirely with their computers and not their teammates. Ex-colleague Jess Rosenberg wrote a great piece that you should read in full. My favorite passage:

“I do think there’s something worth mourning in the transition from ‘we’ to ‘I.’ The solo builder, empowered by AI, is a kind of miracle and also a kind of loss. The miracle is obvious: democratized creation, reduced gatekeeping, the long-tail flourishing of ideas that never would have found support in the old institutional structures. The loss is subtler: the slow atrophying of the collaborative muscle, the gradual forgetting of how to think with others, the replacement of human to human dialogue & conversation with prompting.”

Let us not sit idly and create environments where people talk passionately to their computers and only reluctantly to their teammates.

Let us also not turn ourselves into output-obsessed monsters, spending 16 hours a day spinning up as many agents as we can, unable to sleep without wondering whether we could burn even more tokens on some other phantom obligations. AI Pyschosis appears to be real, and it’s important we modulate ourselves away from it, towards balanced lives in the real world. The goal isn’t to work 9am to 9pm, 6 days a week forever. It’s to work a healthy workweek — whatever that means to you — and have microchips carry out your well-conceived plans asynchronously. Don’t forget that the machines are supposed to work for us, and not the other way around.

Everybody Wants to Rule the World

So what has changed in the candidates we look for today? You probably guessed it, but it comes down to more conducting and less assembly. Essentially, can you guide a small group of humans and a large group of GPUs to produce the very best experiences in the world in timeframes unheard of until 2026? A year ago, it was enough to simply be experimenting and comfortable with LLMs. Now, it’s hard for me to imagine hiring anyone who is still just dipping a toe in. The uplift in superpowers is just too great to keep doing things the classic way. If you’re a designer and you bristle at this, think of how our friends in engineering are feeling. We may not even be looking at code at all in two years, but at least we’ll always be looking at, critiquing, and playing with design.

Aside from conductor skills, we’re also focusing on people who:

  1. Take token design just as seriously as pixel design.
  2. Are comfortable in engineering-dominated environments like VSCode and Terminal.
  3. Don’t wait for others to tell them what to do.
  4. Push the frontier forward, rather than just ape what very-online trendsetters are doing.
  5. Make working software, not just the blueprints for it.

On that last point, Joel Lewenstein, my counterpart at Anthropic, had a really great answer to a question on the By Design Podcast about design being “downstream of engineering” at most companies. Joel said:

“Actual usable software is the lingua franca of Anthropic, and whoever can make that drives decision making, ideation, and roadmaps.”

Traditionally, the people who have ruled the tech world have been mostly engineers and very technical PMs, but leading roles are now available to literally everyone. Designers are in one of the best positions to raise their influence under these new conditions.

Let’s Dance

So how do you, as a design candidate, stand out and prove you have these modern skills that companies are looking for?

  1. Do a self-assessment to see where you might be behind. Don’t rely on hiring managers to see your hidden potential. Other candidates for the roles you’re applying to are likely including entire working applications they built, not just static mockups of projects from five years ago. Don’t fret about who might be more advanced than you. There will always be someone. Just make sure you are comfortable with the new instrument panel and have some tangible artifacts to prove it.
  2. Rebuild your portfolio site if necessary so it contains a very small number (3-5) of things you’re proud of. Using the LLM of your choice, you can rebuild your portfolio in a single night. You should probably spend more time on it than that, but the days of procrastinating weeks of handcrafted new HTML/CSS/JS are over. Coaxing an LLM to present your portfolio in a compelling way is itself proof of fluency.
  3. Follow the tried-and-true format of presenting design work. What problem were you trying to solve? How did you know it was a problem? What variety of solutions did you try? What solution did you land on? Was it successful and if so, how did you know? This format is straightforward and hiring managers are used to evaluating it. It will ensure the depth of your process is communicated.
  4. Many designers, whether right out of school, or 30 years into their careers, haven’t done “fully AI-native work” at their companies or on real projects yet. Don’t let that stop you. Build a casual game. Create a single-purpose app for a problem you’ve always wanted to solve. Hiring managers would rather see thoughtful, modern work on small or even imaginary projects than unimpressive work for billion dollar companies. Show what you can do, not just what the people who currently pay you ask you to do.
  5. Don’t be shy about applying for positions you “are not qualified for”. Many companies (including us) post more positions at the Senior and Principal level because our systems follow an inflexible “one role, one level” policy, and we’d rather see seasoned candidates if we are forced to choose. That said, I’d hire a fresh-out-of-school designer with a great start and a demonstrably high ceiling over someone with 30 years of experience building unimpressive things. If your stuff is good, trust me that no one is going to laugh at your application. If anyone, you’ll probably stand out even more.
  6. As always, applying cold is worth a shot but rarely the best way in the door. If you see a position or a company you’re interested in, find out exactly who the hiring managers are, and figure out a way to be helpful to them. Maybe you have an idea for a new feature or a way to fix a broken part of one of their products. Produce a working prototype of your solution. Maybe you are organizing an event they might feel honored speak at. Introduce yourself and ask them. They may be too busy to say yes, but the point is to break through the pack. Your ability to stand out with a hiring manager is really the first test of your creativity, and you’d be surprised at how many of your co-applicants (99%?) barely even try.

With those tips in mind, here are some of the many roles we have open at Microsoft AI right now:

Should I Stay or Should I Go

In trying to decide what legendary ’80s song to end this little post with, I think The Clash sums up the psyche of the modern designer best:

“If I go, there will be trouble. And if I stay, it will be double.”

A lot of us look around at all of the potential damage AI can cause and our first reaction is to turn away from it. Every time another slop cannon publishes a piece about how much more prolific they’ve gotten since getting the agents to fight each other, The Way of the Luddites seems more attractive than ever.

Then you think back to all of the previous inventions that threatened designers — lithography, the Linotype machine, Photoshop, the internet itself — and you remember this train ride has always been bumpy. Diving into the next part of the journey with both feet isn’t going to speed up or slow down the train, but it might end up taking you to places you never imagined… and that, for now, is worth the ride.

Leave a Reply

Your email address will not be published. Required fields are marked *

Subscribe by Email

... or use RSS