This strikes me as careful cherrypicking of "absurd" results to pick only the non-absurd "absurd" ones...not all absurd conclusions from the past turned out to be okay in hindsight

I don't think Ozy is claiming that all absurd conclusions are correct. Rather, Ozy claims that some absurd conclusions are correct. When you just need an existence proof, there's no cherry-picking - you just pick your example/s and you're done.

People who say "it is okay if my moral reasoning produces absurd results" generally don't personally think "that sounds absurd, but I'll accept it anyway"

Maybe they should! My impression is that Ozy does.

Go tell a vegetarian that he should support exterminating all wildlife to end wild animal suffering, and see what response you get

Ozy's a vegetarian, and their position on wild animal suffering is:

short version:

• wild animal suffering v bad

• currently unfixable because we don’t understand the environment well enough yet to not destroy everything

• am much more sympathetic to wild-animal antinatalism than human antinatalism but am still not convinced

Seems pretty open to absurdity to me.

Rejecting reasoning that produces absurd results even if we can't find a flaw in the reasoning is an important way we avoid errors

I'd prefer the framing of applying an absurdity penalty to one's estimated probability, rather than "rejecting" it in a binary way, but yes: absurdity could be a useful thing to weight in one's estimated probability of a conclusion being correct.