[AndrewKemendo]

Does it really make sense to make that kind of optimization scale to the point that absolutely necessary work fails to have meaningful value?

I think the reason this is a problem is that there is no foundational vector for progress which we can universally agree upon a measurement metric for. So capital allocation becomes the default measure, and money becomes a proxy for value.

So in that sense, what is most easily measurable is more valuable, and something which is easily measured and increases the unit of measurement by the largest factor becomes the most valuable.

Eg. If one unit of money is input and some multiple of that unit is the result, the system with the least amount of work or fastest iteration time (high margin) becomes the place where resources pool.

I'm skeptical whether humans could even define in a meaningful or measurable way how to better allocate resources than individual resource allocation on a common unit of measurement aka "markets."