Machine codes, yes.

Erm. Duplication of code is ok. Removing absolutely every duplicate function is just premature optimization imho.

If you have two different customers with slightly different workflow then go ahead and create two mostly the same functions. When you will have 4 different customers with slightly different workflow, then its a time for refactoring, maybe extract basic same functionality into separate function/object, maybe introduce dynamic workflow using finite automata, maybe extract these functionality to separate modules… but never do it prematurily.

Imho, sometimes ,removing of duplication very much increases complexity and code became hard to understand and hard to modify.

Im not even 100% sure that API is really bad? Some hairness could be easily explained by performance reasons. You don’t want logging to take a massive amount of system resources.

Do we have a term for combination of enshittyfication and LLM?

Hm, I still not sure about this article lesson.

So, main issue is that users of old version can use new feature, but they should not? On a desktop app? But why they sould not be able to do it in the first place?

I mean idea is good, but situation described in article looks like completely fictitious or incomplete…

If SO supposed to be wiki, then why there no clear way to update the answer with new information? Why only the person that asked the question can mark answer as correct? Clearly some person with more expirience should have possibility to mark answer as correct.

Thats actyally very good point. Our phones x100 or x10K more powerful and complex than computers from 90s, but always works and very-very rarely need reboot.