[loup-vaillant]

If you really want to do the expected utility calculation, you'd need to ask if the absence of Microsoft and the Gate foundation would have permitted wealth to be created and moved elsewhere, perhaps to another foundation.

Anyway, if the goal is to slay Poverty, a private foundation probably won't cut it. We need to do away with the current fractional reserve banking system, nullify most national debts, localize production and distribution of common goods (especially food), build things to last (no programmed obsolescence), and stop thinking that GDP growth is a good measure of wealth (it was a good predictor, but now we tend to cheat).

But the amount of collaboration needed to do that looks so huge that it may be easier to build a Friendly AI.