Mmm, well the harder theory probably originated with Ettingermentum (chapo maybe got the name and vibes), who minted himself w the correct prediction, but it would just be that Dems would do better much better than expected, based on the observation that Trump severely damaged his brand following his loss in 2020, and high propensity voters ("normal whites"), who would turn out better than polls would suggest, were down on the GOP, esp. Trump backed candidates; and that these voters were very energized against Dobbs. (I'm sure Ettinger had sophisticated empirical data supporting this).

I think this is essentially what happened during the midterms, and over all the referenda and special elections, which Dems have won, nearly exclusively, since.

But the normal white thesis may not hold when: there is a high turnout, contest (given normal whites will make up less of the vote share), and when Biden no longer looks like a competent, steady alternative, but, rather, like an incompetent, weak old man (an auxillary theory is probably that, Gaza is hurting Biden, not because normal people care about Gaza or the students (they don't), but because it makes him look like a little weakling who is an embarrassment to American prestige). So, very unclear that Biden has normal white luck in November, but it's probably the only theory of the case for his victory, IMHO, tbqh.