Oh, some programmers are definitely better than others. No doubt about that.
I've also found in my interviews that you can separate out the people who are going to be useful almost immediately and learn quickly, from the people who are going to struggle for a very long time to reach a reasonable degree of contribution, just by throwing a few small coding tests their way and letting them interactively work out the solutions.
But my point is, a lot of the advantage beyond that is situational. Take a developer who is massively productive writing hand-tuned driver code for embedded systems, and drop them into a team that's building a machine-learning assisted visual effects pipeline, and that 10x drops down to 0.5x for quite a while. They may learn quickly and adapt well because they're good programmers overall, but that "10x" label is no longer accurate.
It's also quite possible to hire a programmer who can crank out huge amounts of code that only they can understand, and/or that they are unwilling or unable to document, and thereby reduce the overall effectiveness of everyone around them, even as they forge ahead. I've dealt with that type before, and they need a good team around them to interpret, document, enforce standards, and generally push back, or they end up coding their employer into a corner that's very expensive to escape...
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
no subject
Date: 2023-02-03 01:35 am (UTC)I've also found in my interviews that you can separate out the people who are going to be useful almost immediately and learn quickly, from the people who are going to struggle for a very long time to reach a reasonable degree of contribution, just by throwing a few small coding tests their way and letting them interactively work out the solutions.
But my point is, a lot of the advantage beyond that is situational. Take a developer who is massively productive writing hand-tuned driver code for embedded systems, and drop them into a team that's building a machine-learning assisted visual effects pipeline, and that 10x drops down to 0.5x for quite a while. They may learn quickly and adapt well because they're good programmers overall, but that "10x" label is no longer accurate.
It's also quite possible to hire a programmer who can crank out huge amounts of code that only they can understand, and/or that they are unwilling or unable to document, and thereby reduce the overall effectiveness of everyone around them, even as they forge ahead. I've dealt with that type before, and they need a good team around them to interpret, document, enforce standards, and generally push back, or they end up coding their employer into a corner that's very expensive to escape...