As a hacker who eventually somehow landed a SWE job and did it for ~10 years, I found that management expected something entirely different at that point: they didn't want me to code or design systems, they wanted me to spend time ensuring that younger programmers wrote better code/design better systems, or explain how the system was working to management, or thinking about how to prepare for future changes in technology.
Makes a lot of sense from a business perspective. If a senior person can train 10 more junior people, who then produce higher quality work and stick around at the company to become senior, everybody wins.
it's not a IC->manager transition. Many people in my SWE community are in a similar role as me- senior or staff SWE, usually tech lead, but not manager. Manager role is very different from "advisor of junior SWEs". Some overlap.
Right, especially in small-medium companies it's not practical to isolate yourself completely. Eventually - even if it's years - you end up spending a fair amount of time mentoring juniors.
idk about y'all but I do spend a lot of my time trying to write less code, not more. And not in a, "I want to minimize my effort" way, but in that I write more code related to code generation and task automation than I do on implementing any product in particular.
I like thinking about and solving problems - writing code is the tedious part of implementing a solution to me. The less I have to do of that, the more time I can spend on new problems and new solutions.
my goal now is to spend more time writing good tests and less time writing feature code. I get a lot more enjoyment out of a good test- it means people can add new features with confidence, and I often find bugs in the course of writing tests (typically not even related to the test I'm writing).
In fact, my number one advice to junior programmers I advise is: write better tests. Write tests that prevent production bugs. Write tests that allow people to add a new feature with confidence they aren't breaking an old one. Write tests that give 99% of the signal required in 30 seconds, etc, etc.
That said, I often do love writing code- when it's something I'm interested in, that nobody else has done before, and I know once the code is done I'll learn a fact nobody else in the world knows (I work in scientific ML). That's a dopamine hit that is hard to match. But, after a while of advisor really top-notch juniors, I've learned to enjoy the pleasure of... not having to do everything myself, and seeing smart people do things I couldn't have done, in shorter time.
I am a postdoc, and I am doing everything in my power to make sure I keep doing actual things, not filling paper work and advising people only. I don't mind it once in a while, but I want to do things for a living. Not quite programming but similar sort of deal I guess.
I know it reduces my pedigree, and I think more importantly, stability. That said, I would do with the more stability, but as long as I'm on papers that the big wigs are co-authors on, I'm okay.
Click-baity title. Linus is being modest :-) He absolutely IS a "Programmer" just not a "Code Monkey". A Programmer doesn't only write code, he needs to be a bit of an Architect understanding as much of the whole system as possible, how the end users use it, where the dragons are, how to mentor juniors and delegate effectively etc etc. In short there are multiple facets to "Programming".
Linus is one of a kind. I wish somebody would get him to write a book on his Thinking, Design Methodology and Programming Techniques aimed squarely at the Technical Folks. His record is unparalleled and his no-nonsense style highly effective. All worthy of being studied and emulated by other Managers/Programmers.
This is when programmer start to call them self Engineers, tho Torvalds is way more qualified for such a title than all those mobile web devs out there doing dog work.
14 comments
[ 2.8 ms ] story [ 41.8 ms ] threadI like thinking about and solving problems - writing code is the tedious part of implementing a solution to me. The less I have to do of that, the more time I can spend on new problems and new solutions.
In fact, my number one advice to junior programmers I advise is: write better tests. Write tests that prevent production bugs. Write tests that allow people to add a new feature with confidence they aren't breaking an old one. Write tests that give 99% of the signal required in 30 seconds, etc, etc.
That said, I often do love writing code- when it's something I'm interested in, that nobody else has done before, and I know once the code is done I'll learn a fact nobody else in the world knows (I work in scientific ML). That's a dopamine hit that is hard to match. But, after a while of advisor really top-notch juniors, I've learned to enjoy the pleasure of... not having to do everything myself, and seeing smart people do things I couldn't have done, in shorter time.
I know it reduces my pedigree, and I think more importantly, stability. That said, I would do with the more stability, but as long as I'm on papers that the big wigs are co-authors on, I'm okay.
Linus is one of a kind. I wish somebody would get him to write a book on his Thinking, Design Methodology and Programming Techniques aimed squarely at the Technical Folks. His record is unparalleled and his no-nonsense style highly effective. All worthy of being studied and emulated by other Managers/Programmers.
It's true - I saw his pseudocode to make getrandom() insecure.