I’ve been working for the last few months at Thatch – a 4 year old startup that started implementing ICHRA / CHOICE for companies as a way to help improve the healthcare system. They’re very mission focused and incredibly people-centric. Did I mention that I moved back to being an Individual Contributor?
For the first time since 2021, I’m a developer again. I’m not exactly “writing code” – but I am solving problems with code in a way that I was a step removed from as a manager.
Using AI has been a really eye opening thing. I’ve both enjoyed harnessing the power AND struggled with some of the verbosity and over-engineering. There are times that I love the idea of having an app that is entirely owned, built and managed by AI. And other times that I think “how will we ensure that things are good down to the lowest level with AI? We can barely hold a PR stack in our heads, let alone the whole code base.”
And that’s when it hit me. One of the benefits of having “microservices” might not be so that we divvy up what code owns what but also divvying up what developers own something–including “AI” as a developer.
Having AI own your frontend might be a great way forward. You can login and smoke test (or write scripts that smoke test) it very quickly and ensure that it stays in good shape.
Having humans owning some state machines that become very unwieldy under AI’s tutelage might keep the bugs to a minimum.
Having humans own financial pieces and PII pieces could be great too! Ensuring that humans are really understanding everything and writing something clear and understandable so that it can be extended rather than having to rewrite large swaths constantly.
All of this sounds interesting and worth trying. I’m hoping to try it out as we continue to forge forward in AI land.