I don't have a tight argument.

One thing that I see as relevant is the impact of funding bed net distribution relative to the impact of usual work, and another thing that I see as relevant is the amount of social value that founding Google produced relative to the founders' earnings. But these things don't address small probability edge cases in which specific targeted interventions turn out to be many orders of magnitude more important than most activity.

I hope to write about this more in the future.