When writing code stopped being the hard part, I took a job at an animal hospital

Once a coding agent cuts the cost of building software to a fifth of what it used to be, writing the code stops being the hard part. What gets scarce is domain knowledge. So to make my app more useful to the people who rely on it, I went to work at an animal hospital.

I wanted to build an app that understood your cat

A year ago I was working at an AI startup in the Bay Area. Everyone around me was shipping AI apps, coding agents kept getting stronger, and I caught the bug and started making my own things. One of them was Furwise (毛孩指南). (I wrote about that whole first year in another post; here I’ll stick to this one thread.)

The idea came from my first couple of years with a cat. I adopted an adult cat named Hennessy, had no clue what I was doing, and basically lived on Google. I looked up which litter to buy, whether his vomiting was normal, why he tore around the apartment in the middle of the night. One time I got anxious enough to text my family a photo of a paw pad that looked a slightly different color, asking if something was wrong. Then he was diagnosed with FLUTD, a urinary tract problem, and I had my first trip to an animal ER. On one of the follow-up visits the vet asked when it started and how his peeing had been the last few days, and my family and I just stood there scrolling our phones, one of us digging through the photo roll, the other through our chat history. We had the information. We just couldn’t pull it together.

It made me wonder why no app did this. You’d download it, and it would walk you through what your cat needs, remind you when vaccines and checkups are due, and let you jot the daily stuff down as you go: feeding, water, weight, medication, all on one timeline the whole family shares. Once people put their data in, that data could actually be used. I was thinking bigger than that back then. I built a knowledge base, tried some RAG, and wanted it to become something like an old friend who’s kept cats for years, one who knows cats and knows yours.

So the app shipped. Now what?

I brought in a designer I’d worked with before, we built the app together, and we put it out.

The problem was what came next. On top of marketing it, I wanted to make it deeper, to set it apart from the other apps like it. A designer’s instinct is to understand the pain and pin down the need before sketching anything, so every time we talked about a new feature, he’d ask me who it was for and what problem it solved. The deeper those questions went, the more often I had no answer. All I had was my own experience with one cat, and what I ran into isn’t what every owner runs into. Cats come in different breeds, different temperaments, different chronic conditions. What owners actually get stuck on, what vets actually care about, I had no real feel for. I was guessing. You can’t work that out sitting at a desk.

One day I took my cat in for a checkup and saw a help-wanted notice at the front desk: assistant needed, as long as you don’t mind blood or mess. I thought it over on the way home and applied. The idea was simple. There I’d meet real owners every day, watch how vets work, and learn firsthand how to care for a sick animal. I started the job. Not as a vet, just someone doing the grunt work and quietly watching.

What the clinic taught me

Once I was actually there, I picked up a lot of things you’d never learn from a desk.

I look after the hospitalized cats: diabetic ones, ones with liver trouble, kidney cats, urinary cases, gastroenteritis, all kinds. I give meds, syringe-feed, keep an eye on whether they’re eating, drinking, peeing, pooping. I help hold them still, and within what the law allows, I get to watch the vets operate. Even the record-keeping is more careful than I’d pictured. For feeding alone, you weigh the food before and after, so you know exactly how much the cat really ate.

I also got to watch how a vet figures out what’s wrong with a cat. They read the bloodwork, look at samples under the microscope, and read the animal in front of them: how he looks, how he holds himself, even how he smells. The same symptom can point to several different things.

The longer I spent there, the clearer I got on what this app can and can’t do. Not one of the things a vet leans on to make the call is something I can capture. So diagnosis stays with the vet, and the app has no business touching it. What it can do is close the gap between the vet and the owner: help you keep good records, know what to ask when you’re in the exam room, and get through the care that follows once you’re back home.

I think this is the thing worth trying in 2026

I’ve been writing code for almost ten years. Since coding agents showed up, the barrier to building has dropped a lot. For the kind of not-especially-complex product a solo developer usually makes, getting it built is now something almost anyone can do.

When the tech is no longer the moat, whether you can get inside an industry, come away with knowledge other people can’t reach, and find the real pain and the real value in it, that’s what’s most worth a solo developer’s shot in 2026.

Furwise has maybe a dozen or so daily users right now. The two who pay are strangers from other countries I’ve never met. The numbers are small. But the stretch of time I’ve spent working at that animal hospital for it, throwing myself into a completely different world and learning with my own hands how to care for a living thing, has already made it worth it.