Well that’s just flat wrong. Here’s an example of an association of worker collectives in Spain known as the Mondragon Corporation which employs 75000 people:

https://en.m.wikipedia.org/wiki/Mondragon_Corporation

I mean in the age of feudalism I’m sure people thought the idea of electing their kings would have sounded equally bizarre.

Well that's an interesting data point. Ray Dalio's Principles is another interesting data point on an alternative way to run an organization. These seem to me variations on a similar theme though.

They don't at all make the case for anarchy. The hierarchies are still there and still necessary.

I think you are both touching on the question of "if capitalism is so great, why do firms exist?". You might find the work of Ronald Coase interesting in that regard (see https://www.economist.com/leaders/2013/09/07/the-man-who-sho...). The basic idea is that transaction costs for smaller decisions become too high, so it makes sense to have a figure with authority to make certain transaction decisions by fiat. I think Mondragon still has such authority figures, but it has a more "humanist" corporate structure, composed of co-ops, rather than the shareholder-directors-mgmt-and-workers structure in an american c-corp.

corporate structures do matter.

Pretty much. Essentially if you run a thought experiment through several permutations you can roughly compute that this is the case. I always find blowing things up to extremes usually makes certain things come into focus quite well. For example if the organisation has 100,000 people across a multiple geographic regions and then it is discovered that one of the bathrooms have run out of toilet paper, it's not really feasible to call a company all hands meeting to decide who should buy some more toilet paper, where they should buy it from, which brand it should be. If everyone is involved in every decision that pushes productivity to absolute zero the larger the firm grows, though it possibly works for extremely small cases.

By the same token flip the lever hard in the other direction. Imagine management never trusts any expert or IC to ever take a decision on their own. Things get produced but quality plunges abysmally.

There are many configurations things can be in, but from some basic thought experiments you can work out that there is likely a certain operational envelope with a handful of parameters that produce stable, workable results inside the operational envelope and either unstable or unworkable results outside of it.

Some other thought experiments you can try are imagine a world in which no one is trustworthy. Well, it's close to impossible to do business because you have to spend so long figuring out who to do business with that you can't do business with anyone. Imagine a world where nobody ever cheats you and always presents an accurate, trustworthy description of what transaction will take place. Business flies along because you can do away with a lot of overhead that goes in to accounting for that stuff. So it seems trust is an important parameter. Firms make sense from the perspective of having people available (you can trust you have resources to meet demand) and who you trust to be able to complete the work. If you have to go to market every time you can't necessarily guarantee the company you are doing business with will stay in business or keep you as a client or any number of things. So there's a trade-off being made there and there are certain inflection points where on choice stops making sense and another one starts.

Just thinking about it now it seems that the parameters include, but at not necessarily limited to:

  communication bandwidth/overhead
  trust
  alignment of incentives

Which is why no anarchist (and anarchism necessitates the removal of the state) argues for "anarchism in one country", not only because it's oxymoronic but because it's unworkable for the reasons you mentioned and more.