Do you want the list of things that haven't changed since I started programming as a kid?
I still write a lot of Python and shell. I still use Linux, mostly systemd. (I did have to use initscripts early on and do not miss them at all. systemd's way easier.) The entire networking stack has changed only minimally, with only early versions of SSL/TLS becoming obsolete. All the software I use to do my job is the same (A terminal, Vim, Firefox, GNU coreutils, and a smattering of other tools). I still use the same cloud services and databases. The skills I learned in school are equally as useful today as they were then, especially math and CS theory.
The only major shift during my career has been migrating from Linux VMs to containers on clusters (first on Mesos, then Kubernetes). Having administered both at scale, Kubernetes is a lot easier.
> I still write a lot of Python and shell. I still use Linux
All of those listed things have changed. Python went from 2 to 3, number of shells has multiplied like rabbits. Linux changed in leaps and bounds. It got async, BPF, different tech stacks, and window environments.
> Do you want the list of things that haven't changed since I started programming as a kid?
Sure basic principles have remained the same, but huge chunks of ecosystem have been transformed. Perhaps we are looking at it from different perspectives.
It's kinda like ecosystems. Amazon tropical rainforest looks the same now as it did 1000 years ago from birds eye view. But on the ground entire species came and went, and the species living there changed dramatically, in quality and quantity.
My issue is, employers don't want general knowledge, they want an easily slottable asset. You need to know their tech stack. Even if it changes. Hence why LLMs are so in demand.
Python 3 is not that different from Python 2. Yes, if you're a mindless interpreter designed to specifically parse and evaluate Python 2 code, you're going to have a hard time with Python 3. Thankfully, software developers are humans and should not have this issue.
The number of shells that are relevant to any workplace I have worked in have been two, possibly three if you're working somewhere heavily Apple-based. All three of these shells are POSIX-compatible. Other shells exist, but are entirely optional and rarely used (I say this as someone using one of those other shells).
Linux has changed more, and is probably the biggest source of change in the list that the previous poster described, but as they point out, most of these changes make managing Linux systems simpler, and shouldn't really cause significant problems.
I think you're strongly exaggerating the scale of changes here.
> My issue is, employers don't want general knowledge, they want an easily slottable asset. You need to know their tech stack. Even if it changes. Hence why LLMs are so in demand.
I'm a frontend developer. Every job I've had, I was not familiar with the specific framework they were using when I went in for that job. In every case, that's not been a problem at all, because I've been able to clearly demonstrate that I understand the foundation of frontend development, which is far more important than the precise technologies used in the company's framework of choice. My experiences in general are the complete opposite of what you describe here: every single employer that I've worked for has wanted primarily deeper general knowledge rather than specific knowledge of their specific stack.
Maybe this is a market thing — maybe things really are that different here in Europe — but your descriptions do not ring true at all for me.
I still write a lot of Python and shell. I still use Linux, mostly systemd. (I did have to use initscripts early on and do not miss them at all. systemd's way easier.) The entire networking stack has changed only minimally, with only early versions of SSL/TLS becoming obsolete. All the software I use to do my job is the same (A terminal, Vim, Firefox, GNU coreutils, and a smattering of other tools). I still use the same cloud services and databases. The skills I learned in school are equally as useful today as they were then, especially math and CS theory.
The only major shift during my career has been migrating from Linux VMs to containers on clusters (first on Mesos, then Kubernetes). Having administered both at scale, Kubernetes is a lot easier.