Resistance

When every major tech CEO started evangelizing artificial intelligence and using the veil of this new technology to force layoffs, to say I was resistant to adopting this new technology would be an understatement. Not because I wasn't excited about it, but because of how quickly this new tool was weaponized to create tangible rifts in our community. "We won't need developers, designers, or product anymore." It's deeply saddening to see our first reactions coming from a place of anxiety, fear of replacement, and as an excuse to collaborate less.

I can't recall the birth of Geocities, the iPhone, or even WYSIWYG editors like Canva or Webflow causing anywhere near this same level of negative sentiment. I'm not sure if this is unique to paradigm shifts in other professional areas, however, experience taught me that technology workers were uniquely skilled at adapting to change. My gut says the anxiety has more to do with who's pushing for change than the technology itself. Regardless, I have spent the past year putting the noise and hype aside to fully investigate this new technology and how it will shape my work going forward. The irony is, after 14 years in the business — from humble beginnings at a small web design agency coding websites, to massive enterprises servicing millions of users where I never even considered touching an IDE whilst living in Figma-land — my workflow and tool stack looks a lot like it did in 2012.

The Pattern

When the barrier to entry is lowered significantly by new tools, we see similar results: more stuff is made, yet the quality reflects the source. We've used the term "AI slop" to identify the general lack of quality that follows accessible technology, and the pattern is familiar. When desktop publishing put PageMaker in the hands of anyone with a laser printer, we got a generation of self-appointed Paul Rands — the aesthetic ambition was real, but the systematic thinking underneath his work wasn't. The same gap shows up in AI output today. What separates the Cosplay Rands from the real thing is the same thing it always was: they kept working until the thinking caught up with the taste.

AI slop is another manifestation of poor execution as a result of increasingly accessible processes. The barrier to entry of coding, designing, being a "technologist" has just experienced its own PageMaker moment. Each time a production process becomes accessible, there's a lag period where new practitioners have the process but not the accumulated taste that the process used to require to enter.

The deeper irony of AI slop isn't that bad output exists — bad output has always existed. It's that the discourse around AI is itself a form of slop: high volume, low signal, optimized for reaction. Saying you use AI in 2026 carries the same awkward weight as admitting you used clip art in 1993 — the association with the worst use cases poisons the tool. Deleting Facebook five years ago and Instagram at the end of 2024 taught me that the anxiety usually lives in the feed, not the thing the feed is about. Filtering the noise out gave me room to actually evaluate what I was working with.

What Matters

I haven't been more excited to be a designer than I am right now in March 2026. I don't care about Figma, Pencil.dev, Claude vs. OpenAI, or any of that noise. What I care about is: am I still excited by the work I do? AI has helped me return to my original love for web design, a field that ironically requires you to be unemployed to keep up with everything happening. I have years of notes for app ideas I could never touch because of just how difficult it can be to spin something up from scratch, get the logic flowing without hiccups, and ensure the design passes the myriad of accessibility and usability tests. Those barriers are still important, but to me they've always been exactly that: barriers. I don't want to have to think about whether an image tag has a text fallback or whether my website works on every Android phone in existence. What I care about is good design and solving problems. AI has been a great partner — not in doing my work for me, but in doing the work I don't want to think about.

The question is whether AI slop has the same correction mechanism — or whether the feedback loops are different when the tool itself generates the content rather than a human making bad choices with a tool. With PageMaker, the Cosplay Rands had to keep showing up. The tool waited for them to get better. I'm not sure AI slop works the same way when the tool is the one doing the showing up. Until then, I'll continue dusting off my code chops with a new spring in my step.