Generalists vs Specialists

Every developer I have ever hired, managed, or been, sits somewhere on a line between two people. At one end is the Generalist, who wants to know how everything works and get the job done. At the other is the Specialist, who wants to do it the right way and make it a masterpiece. Nobody lives at either end full time. But everybody has a home end, the one they drift back to when they’re tired, under pressure or left alone with a problem for too long.
This post is a journey to find out which end is yours. Not so you can wear it as a badge, but so you can see your own tendencies coming, understand why the person across the desk from you is wired the other way, and work out why a team made up of only your type is a team in trouble.
Fair warning: I’m a generalist. I’ll try to be fair to the other lot. They would want me to be precise about it.
Stop one: which one are you?
Read both of these and notice which one makes you nod, and which one makes you slightly defensive.
The Generalist. You pick things up fast. You’ve shipped in four languages and could do a fifth by Thursday. You are happiest at the start of a project, when everything is possible and nothing is decided. You’ll happily copy and paste something that works, and you’d rather delete it and start again than spend a week understanding it. You get bored at the eighty percent mark. Your code works, and somebody else usually tidies it up later. You find the Specialist slow, precious, and a bit of a show-off.
The Specialist. You go deep. You know one language or one domain better than almost anyone you’ve met, and you’ve read the source. You are happiest in the middle of a project, when the hard problem is finally visible and you can do it properly. You’d rather not ship than ship something you’re not proud of. You notice duplication the way other people notice a crooked picture frame. Your code is beautiful, and sometimes it’s late. You find the Generalist sloppy, impatient, and a bit of a cowboy.
If you nodded at one and bristled at the other, that’s your home end. If you genuinely nodded at both, you’re either very experienced or lying, and I’ll come back to you.
Stop two: the honest pros and cons
I wrote the first version of this list years ago as a hiring note, and I’ve tested it on a lot of people since. It holds.
The Generalist
Pros. Very fast to learn and implement. Flexible, with many languages and patterns to draw on, so they see solutions from other worlds that a specialist wouldn’t. Good at thinking outside the box, which usually means they’ve seen the box before, under a different name. Unfazed by the unknown. Great in a crisis, because a crisis is just a new problem and they like those.
Cons. Unstable code. High duplication. Lower stamina on a long project, because the interesting part finishes early and the rest is a slog. Disregard for standards, usually not out of malice but because they’ve worked under six different sets of standards and stopped believing any of them mattered. Their solutions tend to work beautifully for the case in front of them and fall over on the second case.
The Specialist
Pros. Stable code, in the areas they care about. Rigid standards, which is a pro when you have twenty people in a codebase. They take the time to master a language and keep the codebase clean. Low duplication. When something is genuinely hard, a performance problem, a concurrency bug, a security hole, they’re the person who actually fixes it instead of working around it. Their work ages well.
Cons. Never quite finish. Over-engineer simple problems, because to a specialist no problem is simple once you look closely enough. Instant technical debt of a particular kind: not messy code, but clever code that only they can maintain. Can be hard in teams, through inflexibility, or an insensitivity to the fact that other people are still learning the thing they mastered years ago. Their solutions tend to handle fifteen cases, thirteen of which never happen.
Notice that every pro on one side is roughly a con on the other. That isn’t a coincidence, and it’s the point of the whole post. Hold that thought.
Stop three: why you ended up at your end
Nobody picks a side on purpose. You drift, and the drift has causes. See if any of these are yours.
Your early environment. Generalists are made in small companies, agencies and start-ups, where there is nobody else to do the database, the deploy, the front end and the client call, so you do all of them, badly at first, then adequately. Specialists are made in big companies and serious engineering orgs, where there is a team for everything, and the way to be valued is to be the one person who understands the hard thing. Ask yourself where you were in your first three years. That’s probably where your home end was set.
What you got praised for. If the first time you felt like a real developer was shipping something fast that the client loved, you’re going to chase that feeling again, and that makes you a generalist. If it was the senior dev reading your pull request and saying “this is really nice”, you’re going to chase that, and that makes you a specialist. We become what gets rewarded.
Your tolerance for not knowing. Generalists are comfortable being the least informed person in the room, as long as they can learn fast enough to be useful. Specialists hate that feeling and arrange their careers to avoid it. Neither is wrong. One is a tolerance for ambiguity, the other is a need for mastery, and you were probably born with more of one than the other.
Boredom versus anxiety. This is the one I find most useful. Generalists are driven by boredom. The moment a problem stops being new, their attention leaks away and they start looking for the next thing. Specialists are driven by anxiety. The moment a solution isn’t right, it sits in their head and they can’t let it go. Your worst days as a developer are caused by whichever of those two you’ve got, so it’s worth knowing.
Ego, honestly. Generalists like being the person who can do anything. Specialists like being the person nobody can replace. Both are a bit vain about it. If you felt a flicker of either reading that, there’s your answer.
Stop four: where each of you ends up
Careers sort people, whether they notice or not. Here’s where the two types tend to land after ten or fifteen years.
Generalists become technical leads, founders, CTOs of small and medium companies, solutions architects, consultants, the “head of engineering” at a company with eight engineers, the person who runs the agency, the freelancer who can take any job. They’re also, increasingly, product engineers: the people who can go from a customer conversation to a shipped feature without a handover. Their value is breadth and speed, and the roles that pay for that are the ones where a lot of different problems arrive every week and most of them are not hard, just unfamiliar.
Specialists become staff and principal engineers, platform and infrastructure leads, database people, security people, performance people, compiler and language people, the one person at a big company who understands the payments system. They write the libraries the generalists import. Their value is depth, and the roles that pay for that are the ones where one hard problem matters more than a hundred easy ones, and getting it wrong costs real money.
There’s a trap for each. The generalist’s trap is plateauing: twenty years of being adequate at everything and excellent at nothing, in a market that has started paying less for adequate. The specialist’s trap is obsolescence: twenty years of mastery in a thing the industry moved on from, or a thing an AI now does well enough. Both traps are avoided the same way, which is by borrowing from the other type. The generalist needs to pick one thing and go deep for a while. The specialist needs to look up and learn something they’re bad at.
Stop five: the type you’re not
Here is the part I actually want you to take away.
If you’re a generalist, the specialist on your team is not slow. They are carrying the risk you’re not thinking about. The stability of the thing you shipped in a weekend depends on somebody caring about the stuff you didn’t: the edge cases, the migrations, the security, the next person who has to read it. When they push back on your pull request, they’re not being precious. They’re the reason your fast work gets to stay shipped. Say thank you more often than you do.
If you’re a specialist, the generalist on your team is not sloppy. They are carrying the momentum you’re not generating. Every project has a stretch where the interesting problems are solved and it just needs to get done, and the generalist is the one who gets it done while you’re perfecting the thing that was already fine. When they merge something that offends you, ask whether it offends the customer. Usually it doesn’t. They are the reason your beautiful work ever meets a user. Say thank you more often than you do.
And both of you, try the other end for a month. Generalists, take the gnarliest problem in the codebase and own it until it’s actually right, not just working. Specialists, take the next three small tickets and ship them before you’ve finished understanding them. It will be uncomfortable. That’s the point. The best engineers I know are home-end generalists who learned to go deep when it mattered, or home-end specialists who learned to ship when it mattered. If you nodded at both descriptions back at stop one, that’s probably you, and it’s probably because you did this on purpose at some point.
Stop six: why we need both
Remember the thought I asked you to hold: every pro on one side is a con on the other. That’s not a flaw in the two types. It’s a description of a complete team.
A team of generalists ships fast, pivots easily and builds a codebase that nobody can maintain by year two. A team of specialists builds a cathedral and misses the market. Put them together and you get something that ships and lasts, as long as the two sides respect each other enough to listen. The job of whoever leads the team is to make sure they do, and to make sure neither side is allowed to win every argument.
Neither type is better than the other, though both will argue otherwise. In the extreme, both are miserable to work with, because teamwork is foreign to them and other people get in the way. So when I hire, I’m looking for people who have a home end and know it, and who have made some effort to visit the other one. I’ll take a self-aware generalist or a self-aware specialist over a brilliant extremist every time.
The business perspective, because somebody has to say it
The harsh reality is that most of the time, nobody else in the business will see your code or care about it. What they care about is:
- Does it work?
- Is it on budget?
- Is it on time?
Specialists hate this list, and generalists use it as an excuse. Both are wrong. It doesn’t mean write shitty code and deliver terrible solutions, and it doesn’t mean the clean solution doesn’t matter. It means the clean solution matters because of the list, not instead of it. Stable, consistent code is how you stay on budget and on time next year. Shipping is how you have a next year. Don’t over-engineer, but make sure it’s stable and consistent, and never forget where your lunch is coming from.
If you want the practical version of that last line, I wrote a whole post about over-engineering and what it really costs. It’s the specialist’s failure mode, written by a generalist, which I suppose makes it a team effort.