My Next Side Hustle Isn't Software

It feels weird to be a software developer right now.

I remember when the first computer vision models started getting good. I was excited. I built a startup around that wave and eventually sold it. New capabilities meant new niches, new tools, new businesses that didn't exist six months ago. That's how it always worked — technology got better, and the people who could build with it got new opportunities.

When I got beta access to OpenAI a few years ago, I felt the same thing. Cool tool. Interesting possibilities. I'd just ridden one AI wave to an exit — surely the next one would be even better. I could see how it would make developers more productive. What I didn't see — what I don't think most of us saw — was that it would eventually make developers less necessary.

But here we are. And the thing that's become painfully obvious, the thing I think a lot of us are quietly reckoning with, is this: the world is going to have more software than ever and need fewer software engineers to build it.

It took me two years to see it clearly. My old playbook for building indie software products stopped working. The micro-PE fund I ran — XO Capital, where we bought and operated small niche tools — stopped working too. Distribution dried up — X changed its algorithm and I can't figure it out, LinkedIn reach cratered, the channels I used to rely on to get eyes on things just… stopped performing. I've been stuck, and I think a lot of developers have been stuck in the same way without knowing why.

I'm seeing it from the inside now, too. I joined a YC startup as a dev, and we're shedding software subscriptions left and right. Retool? Deleted — replaced with something we vibe coded internally. SSO tooling? Turns out you can do a lot with Google Admin and a few scripts. So much of the back-of-house software that used to sustain small SaaS businesses is just getting replaced by internal tools that take an afternoon to build. The customers are disappearing.

Two years of running plays that used to work

I've been building software products for years. I knew how to find niches, ship fast, get users. The indie hacker playbook was straightforward: find an underserved problem, build a clean product, charge $20–50/month, grow quietly. It worked. A lot of us built real businesses that way.

For the past two years, I kept trying to run those same plays. And everything kept failing. Not spectacularly — just quietly. Nothing stuck. I blamed myself. Maybe my ideas weren't sharp enough. Maybe I'd lost my edge.

But it wasn't me. The ground had shifted underneath all of us.

Our moat — the ability to build software — used to be rare and valuable. If you could write code, you could create something useful and charge for it. That skill gap was the entire competitive advantage.

Now anyone can vibe code a decent app in a weekend. AI can scaffold the kind of nicehe product that used to take us weeks. The supply of "small useful software" has exploded, and getting distribution for something small and simple is brutally hard. You're not competing against a handful of indie devs anymore — you're competing against everyone with a laptop and a Claude subscription.

There will always be hits

I want to be clear: I'm not saying software is dead. Every quarter there are new YC companies, new breakout tools, new developers who find the right niche at the right time. There will always be winners.

But for a lot of us — especially the ones who built our playbooks in the pre-AI era — the math has changed. The edges we relied on have been commoditized. And if you're sitting there wondering why your old strategies aren't working, you're not alone.

So I started exploring

Here's where my story takes a turn. Instead of trying to force another software idea, I started looking for problems in the physical world — something that touched real life, had real barriers to entry, and was a little safer from the forces I just described.

I ended up starting a microplastics water testing lab. I'm not going to pitch it here — that's not the point. The point is that it's the first non-software project I've ever built, and it's been the most fun I've had in years. I only have 5 sales since I launched two weeks ago, but honestly? Only a handful of my software projects ever made more than $0, so I'm already pretty stoked.

The tech skills don't go away in a business like this — they become an unfair advantage inside a physical operation rather than being the entire product someone can replicate in a weekend.

Maybe you should try something new too

I'm not saying every developer should go start a lab. But if you've been banging your head against the wall wondering why your old strategies aren't working, maybe the answer isn't to try harder at the same game.

Look at the physical world. Services, testing, logistics, things that are annoying and messy and require real expertise. Bring your technical brain to those problems. Get out of your comfort zone.

The best time to start a software company was probably 2012. The best time to be a technical founder solving physical-world problems might be right now.

I'm still early. But for the first time in two years, I feel unstuck.