Split verdict in Georgia?

I'm one of the very few people who (at least tentatively) prefers a split verdict in the Georgia Senate runoffs. And this 538 article suggests it could actually happen (though its very far from guaranteed, since both races are essentially toss-ups).

I'm not going to dive into the debate over possible polling errors right now. But I will note 1) the polls in Georgia were actually very accurate predictions of what happened in November, and 2) though there were serious errors in a few states, the national result (Biden by about 4 points) was within the standard margin of error of what the average of national polls showed (a 7-8 point Biden win). The race looked closer than it was because of how long it took to count mail-in votes, and - of course - because of Trump's bogus claims that he actually "won."

I certainly don't rule out the possibility of a bigger-than-before polling error in Georgia. Run-off races are unusually hard to model, and we've never had one under the highly unusual conditions of this one. But it would be a mistake to just dismiss the available evidence.