[ummonk]

>First, I wish we'd talk more about comp in terms of a tuple of (person, company). There are many companies that will never pay above a certain amount for software because even great software devs just don't move the needle for the business. So I think it's as much a question of where one works as the person's individual characteristics.

True. An engineer can create much more value at a company with a thriving feature-driven business than at a stodgy low profit company.

>It's not just a matter of cranking more code faster. It's domain expertise, knowing what's hard and what's easy, what hard things are worth doing well,

A lot of that changes fast though - and the highest paying subfields of software seem to be new and/or fast-changing, e.g. web dev and ML, as opposed to say kernel programming). I don't think a person with 25 years of experience is like to have an advantage in things like domain expertise over someone with 15 years of experience in these sorts of fields, or what will be hard or easy in modern projects. The experience you gained 20 years ago is not going to be as valuable as the recent experience you gained in the past several years.

>how the social dynamics of teams help or hinder progress, and what's been tried before (both successfully and unsuccessfully).

Right, but this is getting into skills that are valuable in the management ladder, not the IC ladder. I doubt you'll see the same leveling off in pay at 15 years of experience when you look at engineers & engineering managers as a group.