No-one knows what they are doing

yeah, but, preconceived notions are not necessarily wrong. I think there's a reason why Software companies disrupt each other a lot more than they disrupt other industries. its because of their obsession with the beginner's mind when whole industries have been created and studied for a long time. Sometimes, starting from first principals is a terrible waste of time and capital.

>> Not everything is designed, or planned. Sometimes the result is just the sum of a bunch of not necessarily coordinated decisions.

I'll go one further. Almost always. Organizations are basically distributed systems; you theoretically could have fully coordinated decisions with an elegant and efficient network topology (i.e., the decisions involve only those people with relevant information, and don't pass through umpteen unrelated routing layers), but I've yet to see that. Invariably those orgs favoring coordination (CP from a CAP perspective) are slow and inefficient, with too many meetings seeking signoff from stakeholders who have no relevant knowledge or responsibility. The more 'agile' orgs (including suborgs routing around that kind of inefficiency) end up with eventually consistent systems, more AP, and, well, yeah, they're necessarily not coordinated. Hopefully they still have clearly defined areas of responsibility, interfaces, and useful sets of abstraction, but those abstractions are leaky, and the interfaces themselves developed as needs arise.

Because people don't understand what a CS degree is and what it isn't, and that's assuming most CS degrees are fairly similar (they aren't).

A CS degree isn't a 3-5 year road to understanding the protocols and nuances of webdev. Personally, I was able to choose courses in webdev and took zero of them. I came out of it with zero knowledge of how routing works in practice, even if I could guess it in theory.

What a CS degree is (or at least should be), is a set of courses teaching fundamentals, how to think in logic, solutions and pseudocode, and being able to map this decently well to some language. Which means there's a high guarantee they will pick up the skills necessary on the job.

And yes, I know some people get through degrees without being at least okay programmers. I do consider that a failure of the system.

IMHO, the "don't worry, nobody knows what they're doing" advice, has a sort of "Laffer curve"-esque quality to describe its relative danger (or inverse curve, for utility).

On the lower end of experience/expertise (I acknowledge they're not actually the same thing), it is minimally dangerous and highly useful advice -- in the sense that, it helps lower the barrier to entry for beginners into a field. Making something approachable, is generally worth it, even at the risk of giving beginners undue overconfidence or letting them make early mistakes. Just getting someone to even start (and making it easier for them to stick to something new, and not give up too early because it's hard), is more important than doing "the right thing" (pragmatism > perfectionism)... assuming guardrails exist to prevent beginners from doing fatal damage (not true in all domains).

In the middle range of experience, is where this advice becomes maximally dangerous and minimally useful -- it doesn't just give false confidence, but specifically to those in this range who "know just enough to be dangerous." The advice can become an impediment to learning, if someone comes to believe "I don't need to learn anything more, since no one else knows anything, and therefore nobody can teach me anything." People who get stuck here, are the most dangerous.

At the high end of experience/expertise, the advice actually becomes useful again. The expert can appreciate the dangers of ignorance, while also appreciating the need for them to manage/mentor/guide beginners in the field. Likewise, the expert generally re-learns a certain level of humility, because now that they "know everything there is to know", they also appreciate how little they actually know in the grand scheme of things (e.g. see cosmologists or particle physicists).

I seem to recall some kind of saying, something about how hard and precious/valuable it is, for an expert to be able to re-learn the ability to approach problems with a beginner's mindset, with naive curiosity and unburdened by years of accumulated fears or biases? So anyways, the "nobody knows" advice becomes useful again, and helps experts renew their perspective.

This is what Jordan Peterson refers to when stating that Postmodernist claim about reality having an infinite number of objects that can be interpreted in an infinite number of ways is a fair valid claim.

And nature, somehow, made us evolve with the machinery optimized to adapt to a finite number of interpretations.

Hence, our cognitive machinery is made up of "Maps of Meaning" that with a finite number of meanings, make feasible for us to navigate the infinite sea of ignorance of what we can't make a sensical meaning of.