We typically assume that it's really important to prevent great harms. And indeed, usually it is. But there are at least a couple of exceptions.
Most obviously, some harms might be outweighed by greater associated benefits. Benatar thinks it's terrible to allow someone to come into existence given all the subsequent harms their life will contain (no matter how overall happy their life will be). That's obviously nuts. These harms are more than compensated for by the overall happiness of the life. So it's only uncompensated harms, or "net harms", that we should seek to prevent.
More interestingly, even net harms may nevertheless not warrant preventing (in a certain way). For suppose the harm is comparative in nature: the harmful event does not put the victim in an intrinsically bad state, but rather harms them in virtue of depriving them of some much better alternative. There are then two very different ways in which a comparative harm could be prevented. You could ensure that they get the better alternative. (That's the good way!) Or, you could prevent the "better alternative" from ever arising as a possibility to be deprived of in the first place. There is generally no reason whatsoever to prevent a harm (however grave) in this way.
For example, suppose you know two facts: (i) your friend is about to pick up a discarded lottery ticket that is, unbeknownst to them, a winning ticket, worth millions of dollars; and (ii) if they do pick it up, it will be immediately stolen by a compulsive ticket-thief who is watching you both closely. The theft would greatly harm your friend, by robbing them of a future of wealth and luxury. (Further suppose the harm is merely comparative: your friend won't feel any great distress at the event itself.) You could prevent this harm from occurring by distracting your friend so that they never pick up the winning ticket in the first place. But preventing the harm in this way doesn't actually do any good, or make your friend any better off than they would have been had the harm occurred. They equally lack a future of wealth and luxury either way. You've prevented this desirable future from being "taken away" from them only by ensuring that it was never theirs to begin with.
In 'The Worst Time to Die', Ben Bradley argues along these lines that even if the death of a human embryo is a grave harm to that individual (on Marquisian grounds of depriving it of a valuable future), whereas preventing them from ever existing (say through contraception) is harmless, we've no moral reason to prefer contraception to early abortion. Death isn't bad in itself, but only in comparison to continued happy living. So there's no reason to prevent this comparative harm in a way that equally prevents the better alternative, of happy future life, from occurring.
A surprising result!
P.S. It may be worth flagging that Bradley's view remains subject to a closely related objection. If early death is a grave harm to the embryo, it would at least seem to follow that we have strong moral reasons to try to prevent such deaths by ensuring their continued life, if at all possible. But it isn't plausible that we've such pressing moral reasons to prevent spontaneous abortions. So embryonic death can't be such a grave harm. The lack of psychological connection to the future life, as required by McMahan's competing Time-Relative Interests Account (TRIA), plausibly explains why this is so.
(See also: why Bradley's objection to TRIA fails.)