The article is behind the NYT Paywall so I don't know his exact argument, I do trust him a lot more than the average data guy though and agree with his caution.
The bigger problem that I have with the larger discourse on polling this time around is that it is frequently stated and implied that the pollsters haven't done anything to fix their issues from 2020. Generally people say this while, as evidence, pointing to the discrepancy between their expectations (and conventional wisdom) and what the polling is showing.
Does that suggest reason to be wary or skeptical? Sure. But I kinda think people are now going beyond that into "the polls don't matter, they are wrong because they aren't confirming my priors." So much of the punditry is backward looking and fighting the last fight, and it loses sight of the fact that no two elections are the same. Even 2016 and 2020, often compared, were much different elections and fell on much different lines.