While your point about ideology drift over time has some merit, I don’t believe it to be significant enough to materially consider.

https://www.journals.uchicago.edu/doi/abs/10.1086/706889?jou... (https://doi.org/10.1086/706889 if you want to grab the full paper from SciHub)

“Consistent with previous research but contrary to folk wisdom, our results indicate that political attitudes are remarkably stable over the long term. In contrast to previous research, however, we also find support for folk wisdom: on those occasions when political attitudes do shift across the life span, liberals are more likely to become conservatives than conservatives are to become liberals, suggesting that folk wisdom has some empirical basis even as it overstates the degree of change.”

If only a worldwide event would take place where the entire business world deemed it acceptable to change policies so employees can work remotely, then realize they can save boatloads of money by not paying inner city office rents and reduce employee salary expenses by vastly increasing their applicant pool, coinciding with significant increases in inner-city violence sparked by rioting protestors demanding to get rid of the police or else they'll burn the entire city down......oh wait

I'd point out that it's not the polls that are predicting such things. Rather it's people who are interpreting those polls.

Polls are just a snapshot of opinion in time. They are useful and can provide information about public opinion, but they certainly aren't predictive in the way you're stating.

Making predictions like "Biden is 80% likely to win" is a guess. Possibly based on polling data, but are not a poll.

A poll reporting[0] that the respondents of that poll prefer one candidate over another is not a prediction, rather it's data.

[0] https://projects.fivethirtyeight.com/polls/president-general...

Some rural counties that are know to be republican strong holds by a huge margin are favouring Biden by ridiculous amounts. Its not going to happen. Its obviously an error, and the accumulation of these errors are the final polls. (Source https://www.pscp.tv/w/1YqKDpqkYBOKV)

Its true that the democrats are winning the early mail in voting. But since republicans don't trust mail in voting, as Trump told them to vote in person. They will come out on election night, by a large margin.

“Early voting turnout continues to shatter records, as sky-high voter enthusiasm meets the realities of the United States' creaky machinery of democracy amid a pandemic. That means long lines in some places and administrative errors with some mail ballots, but a system that is working overall, according to experts.

More than 26 million people had voted as of Saturday, according to the U.S. Elections Project, a turnout-tracking database run by University of Florida political scientist Michael McDonald. That's more than six times the number of votes cast by the same point in 2016.”

Doesn’t that mean older generations have more influence?

https://www.pewresearch.org/fact-tank/2020/04/28/millennials... (Millennials overtake Baby Boomers as America’s largest generation)

Young voters outnumber older voters, but this is mutually exclusive from an aging populace and a declining birth rate. You can chalk it up to population momentum and previously higher levels of immigration. It does mean Millennials will hold on to their electorate influence longer though.