I suppose I take it on faith that there's a lot of room for more advanced technology before we hit mathematical limits.

Yes, yes, much progress can (and will) be made fomalising our intuitions. But we don't need to assume ahead of time that the progress will take the form of "better individual utilities and definition of summation" rather than "other ways of doing population ethics".

In hedonic utilitarianism, yes. Are you making this claim for preference utilitarianism as well? If so, on what basis? If we don't give credit for creating potential people, isn't most people's preference not to be killed enough to stop preference utilitarians from killing them?

Yes, the act is not morally neutral in preference utilitarianism. In those cases, we'd have to talk about how many people we'd have to create with satisficiable preferences, to compensate for that one death. You might not give credit for creating potential people, but preference total utilitarianism gives credit for satisfying more preferences - and if creating more people is a way of doing this, then it's in favour.

If existing people understand the repugnant conclusion, then they will understand it is a likely consequence of creating all these people is that the world loses most of its culture and happiness, and when we aggregate their preferences they will vote against doing so.

This is not preference total utilitarianism. It's something like "satisfying the maximal amount of preferences of currently existing people". In fact, it's closer to preference average utilitarianism (satisfy the current majority preference) that to total utilitarianism (probably not exactly that either; maybe a little more path dependency).

So I don't see what you mean when you say this reasoning "pre-supposes total utiltarianism".

Most reasons for rejecting the reasoning that blocks the repugnant conclusion pre-suppose total utiltarianism. Without the double negative: most justifications of the repugnant conclusion pre-suppose total utilitarianism.