[vintermann]

> There's no scientific formula for "how much of this result of a bunch of actions that multiple people took over the course of a few years is attributed to each of the people"

I can think of a few. You've got things like Shapley values. But it's not a "neutral" way to attribute outcomes to actors.

It's funny actually, I read about Shapley scores ages ago, and then the go-to example was basically political corruption: assume a bunch of political parties with varying vote weight but no principles whatsoever, aiming to secure a majority to split a "prize" among themselves. But looking at Wikipedia now, it's practically presented as a method to guarantee fairness.

Either way, there's no neutral measure of value (or for that matter, effort) either. What a dollar gets you depends 100% on who else has dollars and how much, so productivity or efficiency can never be separated from distributional concerns.

[saghm]

Fair enough, I wasn't precise. There isn't one single objective way to do this math because, as you point out, different assumptions will lead to different results, and it makes it impossible to do neutrally. My reaction to the parent comment was mostly due to the rigid view that the way things work now is inherently right and proper rather than there being any number of plausibly defensible allocations of credit (and therefore the value generated), and I think many of them are probably a lot more defensible than what we have now.