Why I’m Giving Up My Design Title—And What That Says About The Future of Design
Why I’m Giving Up My Design Title—And What That Says About The Future of Design
Jul 28, 2025
“The secret of change is to focus all of your energy not on fighting the old,
but on building the new.” — Socrates
I vividly remember my first encounter with a private experimental version of GPT-4. It wasn't just the content of the responses that grabbed my attention; early versions often produced errors, making it clear that this was still very new technology. What truly struck me, though, was its eloquence. I thought about how language is uniquely human, the very thing that makes us the most advanced species on this planet. If we could replicate that power in an artificial system, regardless of its sentience, the potential would be enormous—solving everything from complex problems like protein folding, streamlining intricate enterprise workflows, and yes, even doing your taxes.
We were then—and still are—at the beginning of a massive shift driven by artificial intelligence. Despite the skepticism, I've seen enough to believe this technology is fundamentally resetting the paradigm of work and human interaction. My background is in computer science with a specialization in human-computer interaction. Initially, I was drawn to design because I saw its potential to deeply enhance our relationship with technology, moving beyond pixels into meaningful interactions. Design, to me, was never just about aesthetics—it was always about how things actually worked.
But over the past three decades, design as a discipline simply hasn't evolved fast enough to keep pace with rapid technological advancements—a point I’ve thoroughly argued in my 2023 piece, The End of Design Report. I won't rehash those details here.
Instead, I want to direct your attention forward. I’m convinced the future will belong to those who understand AI deeply enough to wield it effectively—those who know exactly what these models can do and how to shape their capabilities. This mindset demands a completely new skill set and tools that are vastly different from those we've used before.
In this piece, I want to lay out five core arguments—each pointing forward, not backward. They are about Paradigm, Innovation, Leadership, Distribution, and Wealth. These five shifts are not just my personal reasons for leaving a design title behind—they're the unavoidable forces shaping the next chapter of our industry. Ignore them, and you risk irrelevance. Engage with them, and you gain clarity about where to aim your ambition.
The Paradigm Argument
Agents are going to be everywhere.
In the early days of smartphones, many companies simply took their websites and wrapped them in apps. Few early apps utilized the smartphone's native capabilities—like GPS, multi-touch, gyroscope, or even the camera. It took time to recognize that smartphones required a completely new way of thinking, ultimately leading to personalized apps and the famous phrase, “There’s an app for that.”
Today, we're similarly in the early stages of Agents and AI more broadly. The versions we currently see aren't what agents will look like in a few years. We're in an experimental phase where everything is still evolving rapidly. But the direction is becoming clearer: traditional web interfaces and even apps as we know them may soon begin to fade away.
Brian Chesky of Airbnb captures this shift well: “The Holy Grail is becoming more like an AI travel agent that’s the ultimate agent that can learn about you and understand you,” he says. “It doesn’t just ask you ‘where are you going’ or ‘when are you going’ but understands who you are and matches you precisely to your travel needs.”
Think back to Craigslist—each section of that simple site eventually evolved into billion-dollar businesses. We learned to solve hyper-specific problems with dedicated interfaces. Yet, over time, these interfaces inevitably became bloated, repeating a cycle of simplicity to complexity and back again. Look at any app store today and you’ll find hundreds of email, calendar, news, and finance apps, each doing just one small thing slightly better than another. This approach of 1:N surface-focused (or UI) innovation is fundamentally flawed.
Agents offer a radical alternative by placing control directly into users' hands. Instead of navigating through endless interfaces, finding a good Airbnb could be as simple as having a conversation with an AI agent. The UI could be generated on the fly, tailored specifically to your preferences; an N:1 model. No more clicking around, no endless tabs, no frustration.
Yes, websites and apps remain relevant today in a multi-modal world—Perplexity's Comet and OpenAI's Agents prove there’s still utility in streamlined web browsing. But these are interim steps designed to ease transition. The long-term vision is fully personalized and dynamically generated user experiences. The entire traditional role of product designers, creating static UI in Silicon Valley offices that work for billions of users, is becoming increasingly irrelevant; when the Agent can simply generate the UI it needs for every single user.
This new agent-driven AI era demands radically new skills that don’t yet exist in design as a discipline. It requires diving deep into how agents are architected, understanding how models are fine-tuned for decision-making, and mastering the orchestration layers and dependencies between agents to achieve complex goals. These are skillsets design simply does not currently encompass.
The Innovation Argument
Design Doesn’t Drive Innovation. It Supports It.
One of the most frustrating patterns I’ve seen across the industry is this: design rarely leads innovation. It struggles even to keep up with it. Despite years of industry talk about “design-led” companies, the reality is this: in most rooms where real innovation happens—design isn’t leading. It’s following.
Look at OpenAI. Every livestream, every major announcement, every product reveal—who’s on stage? Engineers, researchers, founders. Design is never present, because they supported the work, not led it.
This is not unique to OpenAI. Perplexity has a standout brand, but the product itself is driven by engineering and product. Lovable, one of the fastest-growing apps in history, is founded by engineers.
Yes. Airbnb, Figma, Notion, and Linear are incredible companies with design founders. But Field (Figma) and Zhao (Notion) both have a computer science background and only one, Airbnb, is a Fortune 500 outlier.
Across the board, the discipline today celebrates differentiation in experience (e.g. Notion is an elevated experience relative to others), but the discipline ignores the engineering and product logic that made those systems possible. We confuse polish with leadership. The divide is structural. Product and engineering operate in strategy and logic. Design often argues in delight and flow. And in business, strategy and logic win.
If design led and won, amazon.com would look very different. Instead, it’s rooted in strategy, logic, numbers, and efficiency.
I’ve seen this firsthand. I’ve led integrated teams across product, design, and engineering. The fastest, highest-leverage teams always share one trait: they don’t design products in a silo. But this maturity is shockingly rare despite knowing better. Most companies still treat design as a function downstream of decision-making—brought in to make things usable or beautiful, not to decide what gets built or why.
And now, AI is accelerating that divide.
In the intelligence era, what defines innovation is not the interface. It’s the infrastructure. It’s the model. It’s the system, its reasoning. The orchestration. Design didn’t show up to that conversation. It stayed in the Figma file. Even Figma Make, is a sorry excuse of “a canvas for prototyping cool ideas”, rather than actually building upon an AI-first visual metaphor to ship products are scale.
Meanwhile, the pipeline of designers has ballooned—thanks to bootcamps and credential factories churning out UX talent with no systems grounding. The result? Pixel pushers arguing for “delight” in rooms where people are solving for inference cost, latency, and architectural complexity. Innovation is delegated.
The Leadership Argument
Orchestration Costs Are Zero, and Busy Work Is Dead.
In the old world, scaling your impact as a designer meant scaling the surfaces you influence. As you progressed through the ranks—Senior to Principal, Principal to Director, Director to Head of Design—your craft was increasingly displaced by coordination. You became a negotiator, a timeline manager, a translator of ambition through Product and Engineering partnerships. You were promoted not for building the product, but for convincing the organization that your team should.
Eventually, at the executive level, design leadership became synonymous with orchestration. Aligning partners. Synchronizing calendars. Reviewing work at varying levels of quality and fidelity. You were held accountable for outcomes, but spent your days arbitrating inputs. The job became a matrix of stakeholder management, dotted lines, and quarterly planning sessions. It was dignified busywork. Well-compensated, yes—but soul-draining nonetheless.
AI blows this model apart.
In a world where agents can reason, generate, evaluate, and deploy, you no longer need a large team to build something remarkable. You need two people: one who understands systems and one who understands the user. Better if they’re the same person. The orchestration cost has dropped to near-zero. So why would anyone choose to spend their time coordinating when they could be creating?
This is the crux of my shift. When anyone in a company can spin up an agent to design, code, and ship a feature, the leverage no longer comes from how many people you manage—it comes from judgment. Strategic decision making. Agency. Intuition. The people who will rise are not those who manage teams of fifty or five hundred, but those who can build products that change a roadmap with two prompts and a weekend. (Exaggerated for dramatic appeal but this is where we’re headed).
There is now a faster, cleaner path to impact: find something broken, spin up a solution using state-of-the-art tools, and ship it. If it works, you don’t need to ask for permission. The metrics will speak for themselves. This is the rise of the zero-to-one generalist. The era of ladder-climbing through coordination rituals is fading. What matters now is velocity, clarity, and contribution.
This new world despises a calendar full of reviews, design crits, review meetings, and 1:1s. It emphasizes a repo with commits that matter. And promises the joy of shipping to return to your work. That joy unmediated by PowerPoint, politics, or process. That’s not a demotion. That’s liberation.
The Distribution Argument
Design Is No Longer a Differentiator. It’s the Baseline.
One of the least talked about—but most profound—outcomes of the last decade of design advocacy is this: taste has been democratized. People have gone beyond simply being in Apple and Android camps. Today, they expect higher quality experiences from packaging to digital surfaces. Some of the best product managers and engineers I’ve worked with have an exceptional sense of what good design is.
There was a time when every new Y Combinator startup looked like someone tortured an intern into generating a logo using Clipart. Today, thanks to a generation of exposure to good design—and better tools—most founders have internalized the basics of aesthetic judgment. First impressions matter, and now, they’re trivial to get right.
Component libraries like Tailwind, shadcn/ui, and Radix have collapsed the design stack. What once required a full design team handcrafting a system in Figma, exporting specs to Storybook, and obsessively QA-ing the front-end… now takes a few lines of code. Spin up a repo. Drop in some components. Tweak the palette. Ship something that looks eerily close to Linear or Notion in a weekend. This reuse of a strong baseline is absolutely fine no matter how much people critique it. Because at this stage what you’re trying to prove is your idea. You can always hire a branding agency to build your brand later which is what is currently happening.
Look at what Cohere did early in their journey. Hired Pentagram to build their brand, front loaded a strong design system. Then moved onto focusing on the hard stuff negating the need to hire a large design team internally.
Design hasn’t disappeared. It’s been commoditized. Hire Pentagram if you can afford it. Or find an exceptional solo designer to deliver the core brand and design components.
Take a tour of today’s YC cohort websites. They don’t look amateur. They look polished. Cohesive. Designed. Not because they hired great designers—but because great design has been packaged, standardized, and made accessible to anyone with a GitHub repo and an hour.
And it doesn’t stop there. Want a logo? Prompt a visual engine. Need assets? Run a Midjourney batch. Want to rework your design system? Ask GitHub’s Copilot Spark to refactor it.
The distance between an “ugly prototype” and a “delightful experience” has collapsed. Design hygiene is no longer a differentiator—it’s the baseline.
The best founders today are not designers by training. They’re builders with taste. Ivan Zhao (Notion) didn’t hire great designers to build Notion’s aesthetic—he embedded taste directly into the product as its author. That’s what design looks like when it’s native, not outsourced. AI now lets you make it native.
We’re 18 years past the iPhone. The world has absorbed what good digital design looks and feels like. To claim that only designers can perceive beauty—or that only trained designers can produce it—is not just outdated. It’s arrogant. The argument here isn’t that you don’t need designers. It’s you just don’t need as many as you needed before because our collective tastes are a lot higher now.
If design were the true differentiator, everyone would be using Mac and Keynote. Not Windows and PowerPoint. Distribution matters. And AI speeds up distribution for anyone with an idea. People with ideas don’t need to hire designers anymore to distribute their idea at scale. All they need is ‘npx shadcn@latest init’.
The Wealth Argument
Nobody is paying Designers $10M, let alone $100M anytime soon.
The smartest people in the world are racing to win the generative AI race. Every CEO today is redirecting capital and conviction toward attracting technical talent—those who can wring measurable gains from foundation models. These people are not pushing pixels. They are engineering the fabric of future capabilities.
Scarcity drives compensation. In the generative AI era, scarcity has shifted upstream—from pixels to parameters, from aesthetic sensibility to the ability to manipulate token flows and GPU throughput. The gravitational pull of compensation now favors those who can shave 20% off inference costs or architect a routing mechanism that doubles throughput on the same silicon. They aren’t looking for another “Head of Design.” They’re looking for someone who knows how to orchestrate agents, compress context windows, or tune weights on a weekend.
The market sends strong signals. When model engineers, safety researchers, and retrieval architects are earning $2M–$10M total comp, it’s not a trend—it’s a redistribution of value. These roles earn disproportionate pay not because they’re esoteric, but because they create superlinear impact. One person fine-tuning prompts and retrieval can 10x retention. Another who optimizes kernel performance can cut cloud spend by millions. That’s margin. That’s moat.
And the skills are stacking. You don’t just write prompts anymore—you design dynamic routers that adapt prompts on the fly. You don’t just call APIs—you wire tools into agent graphs with fallback branches and reflection loops. You don’t just ship interfaces—you fuse model capabilities with user goals in real-time. This is a technical practice. And design, as it stands today, doesn’t teach any of it.
What does that mean for someone who can adapt to this technical depth? It means a return to technical roots. To see the next decade as a golden window for those willing to go deep—into CUDA, into DAG orchestration, into eval harnesses and alignment pipelines. The tools are new, the leverage is enormous, and the compensation follows the talent. Design has been normalized; AI orchestration and engineering, for now, has not.
The truth is: no one is getting rich designing prettier UI kits. And the “OpenAI paid Jony Ive $6.5B tells you why Design is important” is a pretty lame rebuttal.
People are getting generationally wealthy inventing new agentic abstractions, compressing inference cycles, and scaling frontier models safely. That’s where the gravity is. That’s where anybody should aspire to be. With AI enabling and augmenting you as an individual, there’s a far more compelling reason to chase this frontier. No reason not to.
What Comes Next
The road ahead isn’t smooth. It’s steep, deeply technical, and humbling—but it's also increasingly being augmented by AI. Abstracting away incredible amounts of complexity empowering you to do more. I’m not stepping away from design because I’ve mastered it. I’m stepping away because I haven’t yet mastered what comes next. Truthfully, nobody has.
This transition isn’t about abandoning one identity for another. It’s about evolving—unlearning what no longer serves us and embracing the disciplines that will shape the future. There’s a new skill tree ahead: model internals, agent architectures, memory hierarchies, prompt flows, evaluation loops, and infrastructure that determines how products think, behave, and scale.
It’s daunting. But it’s thrilling. And more than anything—it’s fulfilling.
I get to ship whatever I want. And when I do, the reward is deeper than shipping pixels—it’s the satisfaction of making an intelligent system work the way I intended. It’s harder. But it’s better than the endless meetings, performance reviews, and organizational thickets that define much of leadership today.
That’s why I’m switching careers. From Head of Product Design to Member of Technical Staff.
This isn’t a farewell to experience, clarity, or elegance. It’s a return to first principles. I want to get closer to the metal—to shape the primitives, models, and agents that will define how tomorrow’s software is built.
We need more people at the intersection. Builders who understand agentic flows and elevated experiences. Designers who can reason about trust boundaries and token windows. Researchers who can make complex systems usable—without dumbing them down to a chat interface.
A few decades ago, Steve Jobs asked John Sculley a life-changing question: “Do you want to sell sugar water for the rest of your life—or do you want to come change the world with me?”
Today, every designer faces a similar fork: Do you want to push pixels for the rest of your life—or do you want to change the world with your ideas?
If you choose the latter, the tools are finally here. And the future is up for grabs.
