I agree completely what you are saying here, I think you mostly worded my thoughts far better than I can, thanks for posting.
On problem formulation, as a long practicing Systems Engineer myself, this is a known phenomenon, often customers and clients don’t know completely the problem they seek to fix, so we often have to do some work to help characterise it, by working with them, using what we usually identify as a stakeholder analysis (Or Stakeholder Model).
Trouble is if they don’t realise the extent of their problem, they are often not motivated to bring in Systems Engineers until it is way too late to fix things in an ideal way.
If it is any help to anyone, I would strongly advise enlisting the free help of ChatGPT itself at the earliest opportunity to assist analysing any problem. It already knows all the Systems Engineering knowledge I know, and more. In fact it is by far the most capable Systems Engineering and scientific Research Assistant we could ever hope to have, and it is learning such that it will become the one in charge in no time at all, imho.
I know that makes me at least partially redundant, but I don’t care, as I also know it knows how to fix the global energy problem, which will remove all scarcity, and thus also any need for us to earn a crust by doing something inane just to make the profit needed to live.
So I am already celebrating, considering myself already working for it :)