I actually didn’t claim big swing state margin, I was banging on for days that Florida was closer than all the +3/4/5 polls out there, despite putting it in personal prediction. Also didn’t bash Selzer’s poll.
The point here is polling is an imprecise science at best, if it can be called that, and there’s a million reasons it can go wrong. Change in demographics, change in self-identification, change in media consumption. Pollsters can usually get a fairly representative picture at large, within the MoE, but they are finding it harder and harder, even the best of them, to poll micro pockets of population, or irregular voters who are either not in or hard to reach. That’s a question they have to grapple with and we must be more cynical as a result, but it doesn’t put them on the same level with pollsters who set out with partisan interest in mind, or those who claim they were correct and espouse a theory based on said result while being off more than literally everybody else (USC/Dornsife).