Hacker News Reader

I guess it’s perspective, I’m a career long generalist and I feel confident that I can be put on any team and preform at or better than the average specialist in the same position. I’ve backed this up with every position I’ve had and have built a network of people that know and trust when I say I can do something. I think that’s more important than any specialist because if I don’t know an answer or how to solve something I will study and learn the material until I do.

I tend to agree, and for strong performers/individuals (or just people who are willing to learn quickly/go outside of their comfort zone), it is often possible to become proficient at many ‘specialized’ roles. Especially with softer things, like moving between Java/C#/Kotlin/Swift, or Android/iOS development. There’s a lot involved in each ecosystem, but there’s also a lot of commonality. And, frankly, we shouldn’t pretend like the average contributor on every team is a genius.

There are still roles that probably do require actually specializing to become great at (e.g. graphics/game engine dev comes to mind) but those roles are comparably rare. Most of the Android/iOS/Java/C#/React/JS/TypeScript/etc developers I’ve met could become proficient in any of those other technologies, they just choose not to, perhaps out of comfort. FWIW, in my personal experience, willingness to learn new things/broaden perspective & skill base (a.k.a. being a generalist) is correlated with job performance to the point where ‘not in my job description’ and ‘no I won’t learn UIKit, I’m an Android developer’ stand out as major red flags. “I don’t have time to learn React because I need to familiarize myself with the 8 new types of pointers C++ added this release’ feels a little more reasonable, but again, the great C++ developers I’ve met are still willing to get their hands dirty in other areas.

Tl;dr specialists are (usually) generalists with an attitude problem

I started my career in digital media on interactive installations and I think that is the most influential position I’ve ever had. Every application was different and designed to the clients creative needs; wrote graphics engines, networking engines, transparent h264 format, hardware/software device integration, etc.

I kind of don't know where to put myself on that spectrum. While I'm dealing with a pretty narrow, specialist role - data engineering, there is extreme breadth of tools and approaches in the field. I feel that internally I'm generalist, picking up new tools and ideas with ease. Especially compared to people that mastered internals of Oracle DB for 20 years.

Imho bi / data people need to be generalists to be good. Even if you've only done data, if you're good, i bet you have got inside the heads many job roles, within and outwith IT. You've probably dealt with you're fair share of connectivity and authentication / permission issues. Not to mention being able to translate between strategic objectives and day to day data capture..

It's also something of a continuum. I was an IT industry analyst working for a small firm. This made me acutely aware of consciously picking a point on that continuum. We definitely tended to be more generalist than analysts at big firms who knew narrow areas inside out--and therefore we were better at drawing connections between different topics. By the same token, we needed to be narrow enough to have genuinely deeper knowledge about some set of related technology areas than a more casual observer would have.

I still do something similar within a vendor today and I do find myself having to consciously tell myself "It's OK not to understand much about $X technology area."

There are certainly times when I have to be okay not knowing the whole picture of what I’m working on. I think most problems when broken down enough start to have very simple, similar solutions, or some combination. So if you don’t understand x technology as a whole, you may know a lot about the he individual aspects of the technology.

If a generalist is better at X than an X specialist, do you even get to call them an X specialist?

Yes. Specialization does not imply quality.

What does it imply?

A focus on a single topic/category/skillset. The Ford dealership has a bunch of mechanics who specialize in Ford vehicles, but I find that my preferred mechanic does a better job.

(In this case better job means: fixes it the first time, with no new problems after, and the work doesn't feel sloppy unlike the warranty repair I had to have done by the dealer. In the later case they just reattached 2 panels wrong, something a specialist should be able to do without problems)

The grass is greener where you water it.

Never heard this before, and I love it!

I have a very good friend who leaned into C++ hard about 10 years ago. He's now sitting on one of the languages sub-committees. I think he'll never have any shortage of employment opportunities doing work that he finds interesting. At the same time, I leaned into Ruby hard. I'm now leading a team that writes JavaScript and, quite frankly, I do not have the same confidence. I'm not even confident that I'll be writing JavaScript in the next 1-2 years...

Interesting perspective. I'm on the other end of the coin. Work with mostly C++. Looking at job boards, there are endless opportunities for Ruby and JavaScript positions, not so much for C++ these days. JavaScript will probably be fine for as long as the web exists.

There's a lot of churn in projects / jobs but also in tooling, ecosystem, etc...

I've written web applications in: Ruby, Python, C#, Elixir, and JavaScript in the last 5 years alone. I could technically add Lua to that list, too.

I can get a job done, even jobs that are fairly complex, and sometimes (I'd wager I'm batting 0.500 for whatever that's worth) those projects do have long legs... But I don't have the sort of depth of skill or command of language and specific tooling that I expected to have at this point in my career. That is what feeds into my feelings.

Another way of thinking about this is absolute returns vs risk-adjusted returns. Maybe as a specialist you’d have higher absolute returns, but the probability that your speciality dies may mean your risk adjusted returns are higher for generalists.

You didn't want to be the go-to Y2K guy in 2001. Or the world performance expert in some now obsolete computer architecture.