This is always a controversial opinion, at least in some circles, but: I've long said that the PM role was created solely because companies can't find enough good engineering talent.
That isn't to assert that PMs don't do fantastic, necessary work. More-so that there's a more ideal reality where the engineers should be doing that work; the outcomes, in my experience, are better in orgs where engineering is not just influential in, but responsible for, customers, business outcomes, product planning, etc.
This isn't always possible. Most orgs don't resemble Whatsapp; they resemble Google. A hundred product divisions building a thousand different things. Its already difficult to hire engineering talent to fill an org like this, let alone the additional number you'd have to hire in order to account for the fact that a double-digit percentage of an engineer's role should be PM-like stuff.
So, throw away idealism; outsource that part of the engineer role to another specialty which is easier to hire for: PM. Its reasonable.
But the issue has arisen in that industry has forgotten than this was a necessary sacrifice; its not desirable. We've now got university and bootcamp programs dedicated to training for PM roles. Many companies have entrenched the idea that every team/org needs a PM. Tons of people are now adding financial momentum to this idea that their skillset is valuable and necessary; the ship can't be turned so easily.
At the end of the day, this is a disservice to many engineers, as their people, product, and business skills fall out of practice and they're further relegated into code-monkey positions. Its actually a disservice to PMs as well; they could have been trained in engineering and unlocked far more professional mobility. In the worst cases, it leads to PMs resembling the Office Space "what is it, exactly, you do here" guy ("I bring the specs from the client, to the engineers!" "Then, why don't the engineers just talk to the clients?" "Well, they're not people people!" mhm)