What this says is turnout models changed. People did not change their minds. The pollsters changed their minds.
They decided more of this group or that group is going to vote than they had thought was so a month ago. This, in their minds, they attribute to laughter and being “of color” and younger. But understand what this means. It doesn’t mean any voters changed their minds. It means the pollster changed his and decided more of group X and group Y are now going to vote and were not before.
Where is the reservoir of votes for this? People that they thought would not vote at all. It’s not Trump voters switching. It’s not Independents switching. It’s purely extra votes conjured up in the pollster’s mind from a lower overall turnout number.
Sorry, but turnout is irrelevant when you have ballot harvesting.
In the midterms, we watched the in-person vote follow the “likely voter” polls indicating a red wave. But the mail-in ballots, harvested by various groups, swung the results to reflect more of the “registered voter” polls that took place.