The polls are wrong as long as they keep clashing with electoral reality. Nate Copper’s article is heavy on poll data but flimsy on electoral anecdotes: a county election in 2020 and New York Elections with inconvenient data lopped off (The recent elections to replace George Santos).
The shift the polls are claiming are so seismic that it begs the question why this unprecedented shift is non-existent in basically every post-dobbs election. And let’s not forget the fact that these polls present other, nonsensical trends to like the elderly shifting hard to Democrats too: a shift that can’t easily be waved off by the usual “The shift is only in voters that only vote in presidential elections” excuse.