A Million Save Two Utils-Ten Thousand Lose A Hundred

I think in such a circumstance pure utilitarianism breaks down. You shouldn't just multiply and thus come to the conclusion that the first option is better. Not once the downside to one group is very high versus very low to the other. The problem of course arises if the decision is arrived at via people's votes. Once the downside to those harmed reaches a certain point, adding more people who will very slightly benefit from that downside should not swing the decision. The problem, of course is that if the decision depends on people's votes, the wrong side will win unless half of the people who stand to gain a little, vote against their "interests". They often do do that, either because they evaluated the situation incorrectly, they were lied to by experts who persuaded them that the other side was better for them, they already had more than enough utils so it was no big deal to be magnanimous (eg celebrities) or they simply wanted to be nice.

But often less than half vote this way, especially if the small group who are hurt are people they don't like. Yet another problem with "democracy". "One person, one vote" sounds nice. But is it really a good thing if a voter who very slightly prefers x completely negates a voter (or someone who can't vote) who desperately needs NOT x?