You're forgetting one crucial problem. Younger developers may or may not have more energy, but one thing they don't have is experience. Never, ever, underestimate experience.
Something of a false mutual-exclusion. Due to the massive influx of cheap computing hardware in the late 90s, There are 17-22 year old people who have been programming and studying math & CS for 10 years.
100% true. And this should be obvious: none of get paid to write code. We get paid to solve other people's problems. Experience that matters is experience of doing that, Working in your own projects - unless you have actual users, who can confront you with scenarios you didn't expect - isn't really that transferrable.
By the time I graduated I had already worked on 3 development teams in fortune 500 companies. As much experience as people who graduated years before me? No. More experience than people my age who were in just about any other discipline than me? Definitely.
Young developer non-professional experience also commonly dwarfs what you'd expect in other industries. Try doing molecular biology at home.
I don't disagree that non-professional experience is valuable and that it certainly depends on the developer. For example, it's incredibly easy to work on open-source projects or have code on github that's in circulation and supported. And when hiring for a team, they're definitely good signs.
My main point was that the developing software on your own is very different from working with teammates and having deadlines and processes, where you're often required to choose the best of a set of bad options rather than having the time to do things right or where you're thrown into a legacy codebase that can't be broken but must be fixed.
So - if we happened to be ~16-20 in 1990 and had a good start with a commodore/amiga/vc-20 whatever - is that relevant?
Sure, the experience is bound to be more similar to current hardware, but you can easily be 40 now (I'm now, early thirties) and claim a little more experience.
The next argument could be 'is experience > $someThreshold enough'?
I'm with you, but what's more important is whether whoever is paying the salaries gets that. If you don't have engineers running the show, you can almost bet on management underestimating experience.