I tend toward your side of things. In your first 0-n years of experience, you're growing extremely quickly (hopefully). You're working with a lot of new tech, new people, if n is greater than 4 or 5, hopefully at least two companies if not three. But by the time you have n-teen years of experience, it's much more like you've been repeating the same year of experience over and over again.
Growth, especially if you have a decade of experience (and not one year ten times), is hard. I think especially at the upper end it takes dedicated effort.
I stopped learning programming skills on the job about 7 years ago.
Today, Im still learning new languages, frameworks and libraries.
I'm also still growing in regards to,
Management skills, project skill, leadership skills, self time management, effective habits, not to mention work life balance things as my life now requires it.
One thing I have noticed is that (prior to the pandemic) new positions in my geographic area were almost always limited to the kind of roles that could be successfully handled by programmers with 3-5 years of experience. There simply weren't many higher-level roles available.
I'm in a standard tech area outside of Silicon Valley - plenty of standard enterprise Java jobs with a slightly smaller number of C# positions. My experience here has been that I seem to be able to make lateral moves into similar jobs that don't particularly offer more technical or career growth.
This relates back to your original comment about not seeing correlation to gains in output and skill versus experience in years 12-15 for working engineers. Most job roles in my area simply don't require more output or skill than a software developer gains in those first 3-5 years.
I would guess that for people who do observed a significant correlation in growth in output and skill over those longer years of experience are biased by seeing engineers who were in positions that required growth in output and skill.
Another weird effect (at least in my group - which tends to have long tenured employees - I've been here 12 years) is that I see engineers get stuck in specific responsibilities centered around institutional knowledge. For example, being the devops type for three legacy systems that will exist forever. If a new need comes up, my group seems to prefer hiring a totally new person to own that new project. We have actually lost a couple of employees who (rightly) were frustrated by not being able to get off of legacy support projects. These are the programmers with 20 years of experience but really have that initial 3-5 years plus the same year of experience * 15.
In my case, I had personal reasons for staying in the area and have had to manage my own tech growth. Typically by side study and programming. Some of this has led to new work in the day job but I realize I traded career growth that I would have very much enjoyed for stability. I have also have enjoyed a tremendous amount of flexibility, some autonomy in choosing dev stacks, and collaborating with medical researchers.