Build to learn
I never studied computer science, and no job ever asked me to build a message queue. Side projects taught me anyway, which is the whole case for them.
Published: ·3 min read·First on Substack
Your employer writes your syllabus. Whatever the next project needs is what you learn this year, and whatever it doesn't need, you don't. Side projects are the only place you pick the course yourself, and that's how most generalists actually get made.
I didn't study computer science, and my day job had never involved backend engineering. Everything I know about it came from late nights with random YouTube tutorials, other people's GitHub repos and a lot of vibe-coding whatever had caught my interest that week. No curriculum. Nobody telling me what to learn next. Just an itch to build stuff.
One example is event-driven architecture. No job I've had has needed it. But on the side I built trading bots that take in market signals, turn them into orders and stream each order out to several consumers, so the same trade executes across different accounts at once.
Was it production-ready? No. It was never GDPR-compliant and never tested at any real scale. But I learned message queues, real-time data flows and error handling across moving parts the way you only learn things when they break on you at midnight. Mostly I learned what not to do, which no course has ever managed to teach me, because courses are built so that nothing goes wrong.
This is why the elective matters more than it looks. A side project makes you wear every hat at once: frontend, backend, deployment, security (badly, at first) and the data pipeline nobody else is going to write. You make the expensive mistakes somewhere they're cheap, then carry the scars back to work.
The flip side of expertise is that it narrows what you can imagine. A deep specialist sees everything wrong with their field. They know the rules and the failure modes so well that doing it differently starts to feel reckless. That's exactly what you want from the person refining a system that already works. It's not what you want from the person trying to invent the next one. Someone from outside the field doesn't carry the rules, so they borrow ideas from places the specialist never looks, and ask the questions the specialist stopped asking years ago.
David Epstein (not the Epstein who keeps trending) makes the long version of this argument in Range: Why Generalists Triumph in a Specialized World. In messy, fast-changing fields, people with broad experience often beat narrow specialists, because they adapt faster and connect things the specialists can't see from inside their lane.
The ideal is probably both: deep in one thing, wide enough to steal from everything else (the T-shaped person, if you like consultant diagrams). Specialists bring the rigour. Outsiders bring the odd ideas. If your skills feel messy and self-taught, lean into it. Those are your electives, and nobody else took the same ones.