I think Simon Peyton Jones has the right idea (paraphrasing): programming is one of the most difficult, complex engineering efforts undertaken by humans. At the scale we do it one can't help but be amazed that it works at all, let alone that it works so well.

I think the reason these compensations get that high is because it does take 5-10 years to become good enough to lead a team that manages something so complex and to do it well so that the results are reliable and consistent. I think the difference with doctors and lawyers is that we're not licensed to practice. We're not a capital-P profession. However we still have to attend conferences and stay relevant but the expense and requirements to do so are on us or the companies we work for: there's no professional obligation to do so.

I don't think we're in a programming bubble if the author means we're in a compensation bubble and that programming is over-valued.

I think the real bubble is complexity. We're seeing a deluge of security breaches, the cost of software running robots in the public sphere on unregulated and very lean practices, and a lot of what we do is harming the public... though by harm I don't necessarily mean only harm to human life -- but harm to property, insurance, people's identities, politics, etc... and we're not accountable yet.

If anything I think we need to up our game as an industry and reach out for new tools and training that will tame some of the complexity I'm talking about... and in order to do that I expect compensation to remain the same or continue to spread further out and become the norm. Being able to synthesize a security authorization protocol from a proof is no simple feat... but it will become quite useful I suspect.

[0] https://www.youtube.com/watch?v=PTSE779n0nI