I make a lot of slides in Excalidraw. One night I was building a deck for a video and the toolbar, the sidebar, the little floating panels — all of it — kept getting in the way of the part I actually cared about: the blank canvas. So I did the dumbest possible thing. I right-clicked, hit inspect, and deleted the UI element by hand in the browser dev tools.
And then I just… looked at it. A clean canvas, nothing in the way. I liked it a lot more than I expected to.
That's the whole origin story. There was no grand plan. There was an annoyance, a ten-second hack, and the small surprise of oh, that's better. The same instinct runs through everything I tend to build — the metronome with the good haptics, the bigger language app — start from a thing that bugs me and see how clean I can make it. This one just happened to be the smallest.
The part where I admit I didn't really know how
Here's the honest counterweight before the fun part: I'm a backend dev. At the time I had never really worked with HTML, let alone JavaScript. I knew how to make a server do things. I did not know how to make a browser extension do anything.
So I did what I do now — I described what I wanted to an AI (t3 chat / GPT) and let it write the thing. About forty lines, spread across a couple of files. A keyboard shortcut to toggle the UI on and off.
"a simple excalidraw extension that took me 5 mins to build and 2 hours to package ended up with 100k views… called out by Theo himself"— from my notes on building
To my genuine surprise, it immediately worked. I want to be careful here, because this is the easiest place to overclaim genius and there isn't any. It was forty lines and a good idea. The idea — hide everything, give me one clean canvas, one key to bring it all back — was the hard part, and the idea took ten seconds in the dev tools. The code was the easy part, and I mostly didn't write it.
What I did spend real care on
If there's a thing I'm quietly proud of, it isn't the line count. It's that I refused to ship it half-broken. A toggle that works once and falls apart the moment the app re-renders isn't a tool, it's a demo. So the small stuff got fussed over:
- It re-hides the UI dynamically whenever Excalidraw tries to re-render it back in — so it doesn't pop back the second you draw something.
- It syncs the state across tabs.
- Zero latency on the toggle.
- It works on both free and paid Excalidraw.
- It stays responsive across different layouts.
- One deliberate exception, so the color picker can still appear when you need it.
None of that is hard engineering. It's just finishing — the unglamorous difference between something that technically works and something that feels solid every single time. Then I gave it a name, a logo, and a tagline, published it to the Chrome and Firefox stores, open-sourced it, and made a little demo video. The packaging genuinely took longer than the code, which tells you something about where the real work is.
My two seconds of fame
I posted it in Theo's Twitch chat — Theo of t3.gg, a YC-backed CEO — at about 1:30am my time, mostly not expecting anything. He lost his mind over it, live on stream, and then tweeted about it. His take, roughly, was that companies should hire people who find and solve real problems instead of building convoluted solutions to problems nobody has.
That tweet did 100k+ views. My little forty-line extension, the thing I'd built to stop a toolbar from annoying me, was suddenly being shown to a hundred thousand people by someone I'd been learning from for months.
I'd be lying if I said that didn't feel incredible. But I've thought about why it happened more than the high itself, because the why is the part that's actually useful — and it's not "I'm secretly brilliant." It's something smaller and more repeatable than that.
The lesson I actually took from it
The thesis I keep coming back to is this: when you raise your value even a little, and you have any surface area at all, it shows up fast. Not "build something huge and wait." Just — make one small thing genuinely well, put it where the right person might bump into it, and the work does the talking before you do.
"when you increase your value and you have any kind of surface area, it shows very quickly"— from my notes on building
I didn't spend months on Super Zen. I didn't grind. I noticed something that bugged me, made the smallest clean version of the fix, refused to let it be janky, and shoved it into a stream chat at 1:30am. The surface area was a single tweet. The value was forty lines and a real annoyance solved. That ratio is the whole lesson.
I bring it up not as a trophy — it's a browser extension, let's keep perspective — but as the cleanest little proof I have of a thing I half-believe and am still testing: that the small, well-made stuff is worth making, because you genuinely cannot predict which forty lines find their hundred thousand people. Most won't. This one did. The only way to find out is to keep shipping the small clean things and see what catches.
That's also, more or less, the bet behind everything bigger I'm working on — including the one I've poured actual months into.
Keep going A metronome that feels good · How Flua began