If you spend any time on LinkedIn or Twitter, you've probably seen too many of these posts at this point: A designer shares a screen recording: "I just built a full app with Cursor! No code, and no developer. Just me and my AI." The comments from other designers are all fire emojis and "this is the future!" The comments from developers? "Cool, now try shipping that to production." "Have fun when a user hits an edge case." "That's not an app, that's a demo."
Then a week later, a developer posts the reverse: "AI designed my entire interface! I didn't need a designer at all." The designers show up: "That looks like it was designed by a spreadsheet." "Did you test this on a single real person?" "The hierarchy make zero sense, but sure… you nailed it."
Both posts get thousands of impressions, along with comment sections that turn into a warzone… and everyone seems to frame it the same way. It's either creatives vs. AI, or developers vs. AI.
That's not actually what's happening in those threads though… I'd suggest that it's not humans against the machine, but actually creatives against developers with AI as the weapon they're swinging at each other.
Two brains, one constant battle.
We're witnessing a left brain vs. right brain situation playing out in real-time. Designers typically think in systems of feeling, which all includes content hierarchy, user flows and behaviors, storytelling, along with brand and taste. To most people the work looks visual even though there's a tremendous amount of thinking and strategy that goes into those decisions.
Developers usually think in systems of logic… a ton of thought goes into system architectures, data flow, error handling, performance, and scalability. The output looks like lines of code, but the work is heavily structured and well thought out. A good developer isn't just making something run, they're deciding how it should be organized so it can be easily maintained, extended, and debugged by someone else six months from now.
Both sides have spent years building judgment that's invisible to everyone else. A designer can tell you a layout "feels off" but can't always articulate it in terms a developer would consider rigorous, and a developer can point out how an architecture decision is going to cause pain later but can't always explain why that matters to a designer who just wants the button in the right place.
The tl;dr here about what AI actually did… it gave each side the ability to produce the other side's output very easily. Designers can now generate code that compiles, and developers can generate layouts and styling that look good enough. Naturally, both sides are looking at each other now asking themselves "Do I even need them anymore?"
Each discipline needs each other now more than ever… but explaining why requires admitting something uncomfortable about what each side gets wrong.
What designers usually get wrong about code.
"It works" is not the finish line.
Code that runs but can't be maintained, scaled, secured, or tested is just technical debt wearing a nice dress. AI will generate this for you all day long… it'll give you a React app that renders beautifully in the browser, but falls apart the second a large amount of users start interacting with it. If you don't know enough about code to recognize that, you'll end up shipping apps and experiences thinking you just replaced your developer.
The hard part of development was never about typing lines of code… it was deciding what's worth building, how it needs to be structured, how edge cases need to be handled, how to properly protect user privacy and data, and much more. When these things are not considered, AI is moving ahead making decisions that'll need to be untangled in the future. And in a really harsh way, the designer using AI to code may face a difficult situation if they ever decide to build an app with users… how are they going to fare if (or when) they experience their first data breach?
I've seen my fair share of designers demoing AI-generated apps that look incredible. However, my questions will always be: How is authorization being handled? What happens when the user puts in a bad input? Is this experience accessible and follow WCAG standards? Can someone else get into the codebase and understand what's happening? The silence is deafening if a designer isn't taking some of these into consideration… I get it though, evaluating code quality and performance is usually not part of their job. The designer just wants the experience to work, they want to guide users down the optimal journey, and sometimes it feels like developers are gatekeeping the entire way.
What developers usually get wrong about design.
"It looks fine," but that's not the finish line either.
A UI that renders but doesn't guide users down the optimal experience, build trust, or communicate value is essentially a beautiful layout that doesn't convert anyone. AI is very good at handing you a UI with clean typesetting, decent spacing, and possibly a recommended icon library... but if you don't know enough about design to recognize where the user might get lost, you might mistakenly eliminate your whole UI/UX team thinking you've found a way to optimize your workflow.
The hard part of design was never about pushing pixels around… it was about understanding what a user needs versus what they want, how information gets communicated, what to leave out of the design versus leaving in, and how to simplify the user experience so a customer can actually buy the thing you're selling. AI doesn't magically make you trendy, it gives you a best guess based on patterns it's seen before… and if you don't have the eye to evaluate whether that guess is right or wrong, you're possibly shipping mediocrity and championing efficiency.
Similar to designers, I've seen plenty of developers show off AI-generated designs that look polished on the surface… but once you test it with real users, they start talking about how confusing it is to find what they want, the flow doesn't make any sense to them, and they end up bouncing because they don't know what you're trying to tell them. The designer's job is to help define that journey, so anyone trying to build their site or app without that thinking beforehand is possibly introducing a business problem that'll be hard to rectify. Hopefully they don't lose too much money before figuring out how to fix the UI/UX issues.
It's been painful watching both sides.
I've been working between both design and development for my entire career. At T-Mobile, I'm usually going from Figma to VS Code to stakeholder presentations in the same week. Through Empac, I design the UX, build the front end, write the back end, and maintain the design system everything runs on. When I'm designing, I'm already thinking about how it'll all be built… and when I'm coding, I'm already thinking about whether the design is going to hold up.
Most people don't work this way… most careers specialize, and there's nothing wrong with that. But sitting in the middle for this long gives you a perspective that's hard to get anywhere else. The gap between "I can produce output" and "I can make good decisions" is massive… and AI is making that gap even wider by empowering everyone to create "good enough" designs and/or infrastructure.
All of this feels like that one kid back in grade school that made a volcano for their science fair project. The volcano depends on that one shot of vinegar hitting the baking soda and boof it works. We have designers one-shotting their apps, developers are one-shotting their interfaces, and it's becoming incredibly easy to create a bunch of garbage because one side is trying so hard to save time by eliminating the other.
So... what's actually going to happen?
The low end of both fields are becoming more compressed… if your entire job was "take this mockup and code it pixel-perfect" or "take these requirements and lay out a basic UI," AI can easily handle that now… and unfortunately that work is going to become obsolete. The value will be realized on the higher end of the field though… where someone who can architect a system and evaluate whether the interface makes sense, or someone who can design an experience and understand what's technically feasible will become incredibly valuable. We'll start seeing more people that can understand, and appreciate, both the design and technical parts of building apps and experiences thriving in these new environments and workflows.
The people who should be most concerned aren't the ones worried about AI coming after their specialty… also known as the ones who only know one side of the process and have zero interest about the other. If you're a designer who's never opened a code editor, or a developer who just got a lot more capable of producing "good enough" design by spitballing into a text prompt, you both should be very concerned about your role. Both futures depend on understanding what each other are doing so you can become more well rounded in your craft.
...and what's next?
To be completely honest, the anger in those LinkedIn comment sections are not really about AI… it's more about identity. Designers spent years mastering their craft, and now some developer is posting "look what I designed without you." Developers spent years mastering their craft, and now some designer is posting "look what I built without you." Both feel like their expertise is being trivialized by someone who frankly doesn't understand what the fuck they do.
That feeling is valid. But the response shouldn't be hostility, it should be curiosity.
Start learning about the other side to know what good looks like... a designer who understands technical constraints makes better design decisions, and a developer who understands design principles writes better interfaces. This was true before AI, and it's going to be true moving forward. The cost of ignoring the other discipline is a lot lower, but the consequences of ignoring the other discipline is incredibly high… and it'll bite when it hits you the hardest.
AI didn't start a war between humans and machines, it accelerated a tension that's always existed between two groups of people who build things very differently. The way through isn't to pick a side, it's to get good enough at both to know when the AI output is actually solving the problem... and when it's just a really convincing demo.