The goal was never to build something useful. It was to remember how to play.
From drawing black circles to shipping apps
TLDR: Every time I chased external goals - clients, business outcomes, speed, LinkedIn labels - my creativity died. Every time I returned to making weird, useless things just to see what happens, it came back. AI didn't save me. It just removed enough friction that I could start playing again. The goal was never to build something useful. It was to remember how to play.
This is not a technical talk. You won’t learn how to best use AI or optimize your workflow. This is just my personal experience navigating the past year of creative chaos, and how I managed to not completely lose myself in it.
A kid who wouldn’t stop drawing
My name is Bogdan. I’ve been in design for about 10 years.
Let me go back. Way back - about 30 years ago, when I was a two-year-old in kindergarten. I was an aggressive kid. Hitting other children, scaring my teachers, doing all sorts of things that worried everyone around me. And to make it worse, I started scribbling black circles everywhere. You can imagine the concern.
Then, at some point, I just... changed. Nobody knows what happened. My teachers were relieved. I stopped hitting kids. I stopped drawing black circles. But I never stopped drawing.
If you asked me what I could actually do in life, drawing was the only honest answer. That’s been true for basically 30 years.
Conceptual art and the idea that changed everything
The natural path from there: high school of arts, then Academy of Arts. And in that period, two people became important. One was a friend from high school - he’ll come back later. The other was a professor who introduced me to conceptual art.
Before him, I thought art meant being able to draw really well. He dismantled that completely.
Through conceptual art I got introduced to several artists, but the one who stuck with me was Lawrence Weiner. His idea was simple and unsettling: the idea is complete in itself.
His quote: “The artist may construct the piece, the artist may have the piece fabricated, the piece need not to be built.”
I still don’t fully understand this. But it lives in the back of my brain and it keeps me going. Because it keeps asking: does the thing need to exist as an object to have value? Or is the experience of having the idea already enough?
My master’s thesis was built around this - something I called “unmaterial,” looking at things through the lens of quantum physics, trying to understand that when an idea appears in your mind, it’s already in some sense materialized. And maybe you don’t always need to go further than that.
This period gave me a deep sense of calm. I learned to sit with ideas and not rush to produce.
Web design and the slow death of creativity
Then I finished my degree and had no idea what to do.
That friend from high school showed up again. He introduced me to web design.
At first it was incredible. A completely new canvas. I had no idea about UI rules, UX principles, any of it - and that felt great. That period was honestly the most alive I’ve felt while working.
But the more I learned, the duller it got.
The more I optimized for clients, business outcomes, safe delivery - the more the spark faded. Four, five years of stagnation. I knew my foundations. I could deliver. Clients were happy. But inside, I was just a robot running through familiar motions.
That’s the cost of letting external metrics fully run your craft. You become competent and dead at the same time.
The spark, and then immediately: anxiety
That changed about a year and a half ago. I watched a YouTube video - a conversation between Ryo Lu (head of design at Cursor) and Tommy Geoco. They talked about building, failing, building again - and how the process itself is what’s worth chasing, not the finished thing.
Something turned on.
I had a surge of ideas. I wanted to build everything at once.
And then - immediately - anxiety. Because I wanted to do so much that I got overwhelmed. New tools every week. Everyone claiming theirs is the best. The speed of the industry started crushing that same spark it had just reignited.
So I stopped. Took a step back. Asked myself: what can I actually build, realistically, right now?
The receipt app and learning to breathe
Two things came out of that question.
The first was a receipt scanning app - built purely for myself. There were tools on the Serbian market for this, but none that worked the way I needed. So I built one. It solved my own problem. Nobody else uses it.
But that was the point. It proved something: I could finish things. I just needed to stop chasing the speed train and actually breathe.
At the time I was using Xcode for native iOS development, and this was where I found a weird flow. Working in Xcode with Claude, I almost never hit token limits. I could iterate endlessly. The environment forced a kind of deliberate, specific communication - you can’t just say “move this a little to the left.” You have to name the class, describe the container, be precise. That precision made everything better.
One thing I’ll say clearly: if you don’t know the basics, you can’t prompt properly. Natural language alone doesn’t cut it. The AI isn’t the problem when things break. It’s almost always you.
Useless Notes: from zero to the App Store
The second thing was Useless Notes - the app I actually published.
The concept: a canvas where you can drop notes, bookmark links, burn things down. The idea was to visualize how cluttered we are with information we never look at again. The more you add, the more useless the whole thing becomes. That’s the point.
My animation workflow during this project: I’d go into Figma Make, describe what I wanted to build, then ask it to output the Xcode-compatible version of the code in the chat. Figma Make works in React, so I’d take that output, go back to Xcode, and say: “Add this animation but accommodate to our codebase.” That one sentence matters. Without being that specific, nothing worked cleanly.
Then came the moment I almost quit.
I was weeks deep, close to publishing, and I accidentally reverted the wrong commit. In Xcode, there’s no simple undo for this. I was furious. I sat there wanting to break something, eventually just... cried. Told myself I was done.
Then I opened Gemini, explained what happened, and recovered the commit in two minutes.
I’m a bit of a drama queen. Almost never AI. Almost always you.
After that, I committed properly (still badly, but better), kept going, and shipped the app. Will anyone use it? Probably not. That was never the point. The point was the process.
Pixelb and Claude Code
After Useless Notes, I finally tried Claude Code - specifically to build out my web portfolio.
I didn’t finish the portfolio in time for this talk. But I did build a character.
Claude Code named him Pixelb. He’s the main character of my website - a pixel art figure who follows your mouse, wants to hang out with you, gets nervous when you open a case study, headbutts icons when you click them, and disintegrates into particles through a portal into my personal corner of the web.
The whole portfolio is intentionally eclectic. I couldn’t decide between minimal and 80s arcade, so I mixed things that shouldn’t be mixed. I broke UX rules on purpose. Some things are misaligned by a pixel, just to make you notice.
That’s who I am right now as a designer. I’d rather trigger a reaction than play it safe.
Money and tools (the Serbian way)
Practically: I run two subscriptions. Google Pro for Gemini access (which also gets me into Windsurf for free, Gemini models only). And Claude. That’s it.
The workflow is messy: Claude chat for ideation, Figma Make for prototyping animations, Claude Code for building, back to Windsurf or Cursor for small edits. It’s not optimal. But it keeps costs low and keeps me moving.
I compare AI to calculators. When calculators appeared, people didn’t forget math because they were dumb - they forgot because they stopped practicing. Same thing will happen with AI. If you delegate everything, you stop growing. The tool can only extend what you already have.
For dumb, repetitive tasks - delegate. For things that matter to you, sit down and go through the process. The speed narrative is everywhere, but speed isn’t the point. Play is the point.
What I actually learned
Not how to prompt better. Not which tools to use.
The real lesson is simpler: opt out of the speed-and-status game and fall back in love with the process.
Every time I chased external things - clients, business goals, industry labels, LinkedIn metrics - I lost myself. Every time I went back to making something weird and useless just to see what would happen, the spark came back.
Conceptual art taught me early that the experience of having an idea might be more valuable than materializing it perfectly. That seed took 15 years to become useful.
AI and code just removed enough friction that I could finally build the things in my head without needing a team or a budget. The friction removal is real and it’s valuable. But if you use AI as a replacement for thinking, you’ll get overwhelmed, anxious, and lost - same as before, just faster.
The goal was never to build something useful. It was to remember how to play.
The more you play, the better the people around you feel - because you feel good. The more you explore, the more you learn. The more you learn, the more you can share.
That’s it.