At the recent MANCEPT workshops I had some fun discussions with a couple of defenders of the idea that, while we've reason to improve the lives of people who exist, there's no reason to bring awesome lives into existence. (As Johann Frick put it in his very interesting paper, "our reasons to confer well-being on people are conditional on their existence.")
This strikes me as a rather bleak, depressing view of the value of life. Johann compares welfare to promise-keeping: there's no value in making-and-keeping promises, it's just that once you've made a promise, you'd better not mess it up. Likewise, it seems on these views, there's no real value in good lives, just the risk of their going badly which needs to be avoided.
I think the big worry with these kinds of views is the seemingly unavoidable conclusion that sentient life, as a whole, is regrettable. If God creates a world with a billion blissful, flourishing lives, and one (antecedently identifiable) very slightly bad life, his act of creation is deemed to be wrong on net: the bad life counts against it, and the good lives don't count in favour.
Indeed, as others (especially Matthew R.) brought out in conversation, those who deny that (good) existence has value will find it difficult to avoid full-blown anti-natalism. If future generations are sufficiently numerous, the aggregate badness of the (proportionally very few) bad lives will eventually be arbitrarily large. Assuming that any countervailing values -- including preservationist values, and the value existing people get from having offspring, etc. -- are bounded, they seem likely to be outweighed by the aforementioned harms. (Even if we're harmed by the end of humanity, we are not harmed as much as those possible future individuals who would otherwise have bad lives.)
To avoid these unacceptable implications, we should simply deny the procreative asymmetry on which they're based. Contrary to common belief, granting that good lives have real value does not have any comparably untoward implications. It should be a philosophical no-brainer.