Louis, very nice article, I love the question you have raised, and how you raised it.

Let me take the analysis down a slightly different path:

In one sense we have already past peak human. If we consider what the vast majority of what humanity did 200 years ago it was farming, and if we consider the collective capabilities of all the industrial infrastructure on a modern farm, then that infrastructure far outpaces 100% of human output of 200 years ago using only a tiny number of humans. So it already beat us.

But we don't think about it this way, since humans USE those tools and thus the per-human output is bolstered above what the machine only output might be.

Perhaps a better way to look at this is to look at the fraction of costs for modern products flow to human labor, vs capital infrastructure (including AI). When we view it that way humans get the benefit of the tools (AI) as they are built, but the loose credit as the tools become so good, we no longer need the humans at all, and can just perform using machines (AI) alone.

This is a different measure of peak human, but perhaps a more important one. After all the plow beat a human, and so does the calculator at their given tasks, but who cares? What really matter is how much money we are spending on plows and calculators vs. on human labor.

Peak humanity is hit when the total dollars paid to human labor peaks, and when that happens it will have devastating consequences for humanity.

Personally on that measure I think we have a decade or two, it just takes time for new businesses sans humans to get built even after the core the tech is in our hands. but it does not take too long, these thing generate trillion dollar races to be first.

I think this other measure of peak humanity relates to your measure. My measure gets reach after yours. even if a machine can write songs or NY times article better than 99% of all humans, we will still employ humans to do these things since the output of man+machine > machine alone. So even after peak humanity is reach by your measure, it will seem that it has not been reached, since 100% of the articles will still be better when authored by humans+machines.

At the point where my measure starts to inflect, then the balance of power against humanity at large will really be shifting.

VERY INTERESTING THINKING! thanks!