This week, I built a working prototype with Lovable. No code needed. Nice UI, working buttons, mobile friendly - all without writing any code. When I showed it to my client, one of the devs said: “Ok, we are doomed now!” He said it jokingly, but… The hard truth? AI is coming for all of us - designers, developers, product managers. Not a question of if, but when. And which role dies first. Don’t close this email thinking “not another AI doom post.” I’m not here to scare you. I’m here to help you be the last person standing when AI hits our industry hard. Let’s look at which tech role is most at risk - and what to do if it’s yours. The Case Against Product ManagersAfter 7+ years in tech, I’ve watched PMs run around collecting feedback, making roadmaps, and saying “no” to feature requests. But what if AI can do all that? Here’s why PMs might be first to go: AI is already doing PM workHave you seen tools like ProductBoard and Cycle lately? They don’t just organize work - they suggest what to build next. They analyze user feedback automatically and tell you what features will have the biggest impact. Claude and ChatGPT can write PRDs that look just like human-written ones. Feed them some user interviews and market research, and they’ll spit out a decent product spec in seconds. PM work is mostly pattern matchingMost product decisions follow patterns:
This is perfect for AI. It’s just connecting dots and following rules. The metrics are clearSuccess for PMs is measured in numbers - conversion rates, retention, revenue. AI loves numbers and can optimize for them better than humans. When the CEO asks “will this feature increase revenue?” an AI can run simulations based on past data and give a more reliable answer than a PM’s gut feeling. Real talkBut real talk: A big part of a product manager’s job is stakeholder management, making sure everyone feels heard, and to translate between clients, founders, devs and designers. I can’t see a world where AI replaces PMs. However, It will be heavily AI assisted, where all the technical and documentation work is automated. So PM is probably not the first one to go. The Case Against DesignersAs a designer myself, this part hurts to write. But we need to face the facts. AI is getting scary good at visual designMidjourney, DALL-E, and Stable Diffusion are making amazing visuals. Adobe’s Firefly can generate UI components that look like they came from a senior designer. Tools like Galileo AI can make entire app screens from simple prompts. And they’re getting better every month. Design patterns are… patternsLet’s be honest: 90% of what we design follows patterns. Login screens, dashboards, settings pages - there are only so many ways to design these things. Please, never sit your ass in front of a blank Figma file and try to reinvent the signup flow. Users know exactly how to sign up. They know where to look for the “forgot password” link. AI knows these patterns too. And it can apply them perfectly every time. User testing is being automatedTools like Maze and UserTesting are adding AI features that can analyze user behavior without a designer watching hours of recordings. Soon, AI will tell you exactly where users get confused and suggest design changes to fix it. Real talkThe most at-risk designers are the ones who just make things look pretty. If all you do is visual design and basic UI work, AI is coming for your job fast. But designers who understand business goals, can talk to users, and translate complex problems into simple solutions? They’re safer. Design thinking still needs a human touch. For now. The Case Against DevelopersMy dev friends hate when I talk about this. But after my Lovable demo, they can’t deny it anymore. Coding is being automated fastGitHub Copilot writes 40% of code in files where it’s enabled. And that’s just the beginning. Tools like Replit Ghostwriter can write entire functions from comments. New tools like Devin and Claude Code can build whole applications from simple descriptions. Even my Lovable prototype shows that non-devs can now build things that used to need “real coding.” Frontend is getting easierMost frontend work is becoming drag-and-drop. Webflow, Framer, Bubble, and now AI-powered tools are making it possible for designers like me to build working websites and apps. The days of hand-coding HTML/CSS are ending for most projects. Backend is nextSure, backend seems safer now. Database design, API architecture, and security are still mostly human work. But look at what’s happening with tools like Supabase, Firebase, and AWS Amplify. They’re automated backend services that need very little custom code. Once AI can handle security edge cases and complex data models, backend devs will feel the pressure too. Real talkThe average developer spends hours on Stack Overflow looking for solutions to problems that have been solved thousands of times before. AI doesn’t need to search - it knows most common solutions already. Junior and mid-level devs who mostly implement standard features are most at risk. The ones who just turn designs into code will be replaced first. But developers who can architect complex systems, understand business needs, and solve novel technical problems? They’ll stick around longer. The question is: how many of those roles does a company really need? So, who will die first?After looking at all three: I think in the next 2-5 years, it will hit designers the hardest. Only those who can really understand the user and build interfaces that excite will stay. Just making pretty software is already not enough anymore, but will be done fully by AI soon. However, understanding the user, and finding new ways to solve their problems is something I can only see an experienced, humane designer can do. PMs are safe for me. Maybe the role itself will shift, and devs will do more and more PM work, but you can’t really automate the human component of the job. Now, looking at devs. Oh man. Sorry to all my dev friends but. It will be suuuuuuper hard to replace devs, and our technology is not there yet, but it will. And then PMs and designers can just code frontend, backend, whatever. This is the only role where no direct customer/user contact is required, which means it doesn’t really require a human. That’s why I think the basic Frontend, backend or Full-Stack Dev will go. However, machine learning and computer vision will be more and more important, and people will also move towards quantum computing, so there’s light at the end of the tunnel. To conclude:
Adapt or DieI’ve been pretty direct in this newsletter. Some roles are going to disappear. It’s not a question of if, but when. And honestly, we’re all at risk in some way. But this isn’t about fear. It’s about opportunity. Every technology revolution has eliminated jobs, but also created new ones. The printing press put scribes out of work. The industrial revolution replaced craftsmen. The computer made file clerks obsolete. Yet here we are, still working, still creating. The difference now is how fast things are changing. We don’t have decades to adapt—we have years, maybe months. So what should you do right now?
I started this newsletter with a story about how I built a prototype that made my dev friends nervous. But the truth is, I’m nervous too. We’re all in this together. Let’s help each other evolve instead of competing for the last few jobs that look like the ones we have today. Please Please Please, reply and let me know what YOU think. I’d love to be proven wrong! Stay valuable, Nik |
Weekly insights on user research, business metrics, visual design, and team collaboration - curated in 7+ years working with industry leaders like Telekom & Deltia AI.
Basically what the title says. Social media discussions about the best no-code tool for websites have been heating up more and more. And I can’t stand it anymore! So let’s clarify something… I’ve been designing websites for over 7+ years now, and after trying out almost every website builder, I got stuck on two: Wordpress Webflow Onepage Squarespace Wix Framer So I decided to build the same website template on both platforms and see how they compare with each other. The result might surprise...
After 7+ years in product design, I’ve seen designers obsess over the wrong metrics. They track everything: time on page, scroll depth, click rates, heat maps, and dozens of other data points. Yet when the CEO asks how their work moves the business forward, they freeze. Here’s the truth: You only need 5 metrics to prove your design’s value and become indispensable to any company. Not 20. Not 10. Just 5. Let me show you which ones actually matter and why most of what you’re tracking is just...
The way I work has drastically changed in the last 3 months. AI is everywhere in my process now, spearheaded by Claude. And honestly… I might stop using Figma in the very near future. In this issue I’ll be sharing how my design process got impacted by AI (in a good way), how you can get better design results with tools like Claude, and what needs to happen for me to ditch Figma once and for all. How & where Claude took over my design process I never had the feeling AI will ever replace me. I...