Build to learn, Build to Gain a Generalist's Superpower

I’ve been reflecting a lot on expertise, creativity, and how we learn new things. When you’re a deep specialist in a field, you see everything wrong with it. You know the rules inside out, the best practices, the pitfalls - so clearly that it almost becomes hard to imagine doing things differently. That’s invaluable for refining and perfecting what already exists. Specialists keep the world running smoothly and push incremental progress forward.

But that same depth can sometimes trap you. You get so immersed in “how things are done” that radical new ideas feel risky or even impossible. This is where generalists - or people who come at a problem from the outside: have a huge advantage. Without all that baggage, they can borrow ideas from completely unrelated fields, ask “stupid” questions, and connect dots that experts in their silos might miss.

I didn’t study computer science in college, and my day job has never involved backend engineering. Yet, out of pure curiosity and the itch to “build shit,” I dove headfirst into coding on my own time. Late nights watching random YouTube tutorials, digging through GitHub repos, vibe-coding whatever caught my interest - no curriculum, no boss telling me what to learn next.

That freedom let me explore wildly different areas without anyone defining my path. My employer decided the projects (and therefore the skills) I picked up at work, but side projects? Those were all mine. And that’s how you accidentally become a generalist: by building things you actually care about.

For instance, I’ve never touched event-driven architecture or pub/sub systems in my corporate roles. But on the side, I built trading bots that receive market signals, process orders, and stream them out to multiple consumers so the same trade gets executed across different accounts simultaneously. Was it production-ready? Probably not - definitely not GDPR-compliant or battle-tested at scale. But man, did I learn a ton. I got hands-on with message queues, real-time data flows, error handling in distributed setups, and most importantly, what not to do in certain scenarios. Those lessons stuck way deeper than any theoretical course could have.

Building your own stuff, especially outside the constraints of work: is one of the fastest ways to gain that broad, generalist experience. You end up wearing all the hats: frontend, backend, DevOps, security (even if poorly at first), data pipelines. You make messy mistakes in a safe space, iterate quickly, and bring those unconventional insights back to your “real” work.

David Epstein (not the Epstein trending in news these days) in his book Range: “Why Generalists Triumph in a Specialized World”, shows how, in complex and fast-changing domains, people with broad experience often outperform narrow specialists because they adapt faster and make unexpected connections.

Of course, the ideal is probably a mix: T-shaped people who go deep in one area but stay wide enough to pull in fresh ideas. Specialists provide the depth and rigor; generalists bring the sparks. If you’re feeling like an outsider or your skills feel “messy” and self-taught, lean into it. That curiosity-driven, build-whatever-you-want mindset might just be your biggest edge.