This week I built two apps and wrote none of the code. For the first time in a long time, I’m not burned out, but, like, actually speeding up.
The first app was a card game for my family. We play an obscure one where you trade cards with your neighbors and the whole point is making your opponent fold under pressure by cracking a smile. Now that we're all older with our own families, it's harder to spend evenings together—even on vacation. So I thought: I wonder if I can recreate that face-to-face experience with WebRTC and have it work on hotel wifi? Turns out, yeah. I outlined the constraints, and Claude Code brought it to life. I've built game servers before, and what Claude wrote was about as good as what I would have done by hand. But I don't care. I just wanted to see if WebRTC could carry a bluff. And now my family has a new way to connect when we can't be in the same room.
The second app was for my brother. He just started a new job, and his seasoned colleagues struggle with writing things down. It's not that they don't want to. It's that they don't have anything good to write in. Their options are Word or Excel on a shared drive, or some internal system buried behind three layers of auth and MFA. That's not an exaggeration. And because data exfiltration is the compliance team's biggest fear in his line of work, any request for a modern SaaS product is dead before the RFQ hits their inbox. What they do have is Chrome extensions and an old-school Windows Z:\ drive. When he told me about it, I wondered: could we write a Chrome extension that uses the Z:\ to build a collaborative note taking app? An hour later, we had a PoC. It worked unbelievably well. We then learned Chrome extensions can be locked down to deny all network activity, which is another box checked for compliance and security.
Did I write code? No. Did I build something? Absolutely.
Here's the thing I want to be clear about: I'm not anti-code. I care about fit. What I don't care about is being part of a tribe. I'm not a language elitist. I'm not a framework cultist. I'm not an AI boomer or a doomer. I just want to go from question to answer and skip the parts that were never giving me joy.
I don’t want a woodshop where I cut dovetail joints all day. I want a list of projects and the right tools to finish them… like, once. Because what I really want to know is how things work. I’ve hung drywall. I loved every minute of it the first time. I never want to do it again (unless I know you and you need help—then I’ve totally got you).
But, that's just how I've always related to software. Growing up, I watched How It's Made. I watched those old History Channel shows about how big projects got built. I wanted to know how things worked. And when it comes to knowing how computers work, learning to code was a necessary side effect.
Here’s what changed: the cost to try something—to learn something—is basically zero. The gumption tax on drawing the rest of the owl is basically gone too. For me, that’s the real unlock.
The reason I’d burn out is that with each bigger project I took on, the percentage of interesting parts grew smaller as the necessary grunt work grew larger. Early on in my career I vacuumed up every project that came my way because it was new and exciting. Now, I just want to test little ideas that exist in the context of a larger canvas. I want to solve constraints like a puzzle and think about what could work. But, finding the time and energy to build the whole world just to test one little idea is crushing.
With Claude Code, I’m not investing time. I’m investing attention. I focus on the parts I care about. I don’t stall out when the work stops being fun. I don’t feel like I’m broken because I don’t force myself to power through the parts I don’t like. I pass the concept from brain to Claude and it implements it for me. Meanwhile, I’m already working on the next problem.
It’s a wonderful thing.
✌️