I haven't posted here in almost five years. That's wild to type out. The last thing I wrote was a Cypress testing tutorial in September 2021.
Life happened. Work happened. A lot of building happened. I just never sat down to write about any of it.
So here's the update.
A startup in stealth
For about a year now, I've been working with a startup on something we're building in stealth. I can't talk about what it is yet. What I can say is that it's the most exciting thing I've worked on in my career. Some days my brain genuinely hurts from imagining what's possible with what we're building.
More on this when we're ready to go public.
I don't think I'll be writing code much longer
A friend asked me recently if I was scared about how AI is disrupting software engineering.
Scared? It gets me freaking excited every single day.
Here's what I believe. In about five years, software engineers as we know them are going to disappear. I don't mean the profession dies. I mean the job description changes so much that what we call a "software engineer" in 2026 won't exist anymore. Companies won't hire people to write code. The writing-code part is becoming automatic. If all you can do is be a "web developer," it's going to get harder and harder to stay relevant.
But the flip side is way more interesting. AI getting better opens up problems at a scale that wasn't possible before. And those problems need people with different skills. People who can ship at 10x speed. People who know how to use AI tools (not just LLMs, the whole ecosystem around them) to experiment with ideas, iterate, pivot, build things that would have taken a team of ten.
Writing code is not the bottleneck anymore. Having a vision is. And then: how good are you at using the tools and the speed that AI gives you to get there?
That's what I'm honing my skills towards.
Everything I ship these days, the technical side matters less and less to me. I want to get good at orchestration. Building foundations for systems that AI can use. Whether it's for product development, marketing, SEO, A/B testing, outreach. Every new thing I work on, I learn something about getting more out of AI tools. And I systemize what works.
My prediction: in two years, I'll personally stop writing code entirely.
Half-joking. But only half.
Experimenting a lot
The way I've been developing these skills is by building stuff. A lot of stuff. Side projects, experiments, random ideas. I never thought of myself as an entrepreneur, but here we are. Most don't go anywhere. But each one teaches me something about working with coding agents, about how to set up projects where AI does the heavy lifting and I stay in the driver's seat.
I've gotten noticeably better at it over the past year. And these skills carry directly into my day job too — every experiment I run on my own time makes me more effective at work.
DishCost
One experiment that stuck: DishCost.
I got curious about how restaurant owners manage their costs. How they track ingredient prices, figure out what each dish costs them, decide if their menu prices actually make money.
Turns out it's a real mess. The existing tools charge $150-300/month, require long contracts, and are built for big restaurant chains. If you're an independent restaurant owner running a single location and working 12-hour days, your options aren't great.
So I'm building something simple for them. You put in your recipes and ingredient prices, it tells you what each dish costs and whether your margins make sense.
It's early. But it feels different from my other side projects. I spent a long time wondering how people come up with ideas for things to build. Turns out you just have to pay attention. The more I learn about the restaurant industry, the more I respect the people in it.
That's where things stand. Busy in a good way. If you want to follow along, I'm around on Twitter.