I'd argue that 2016 would have gone differently if Clinton hadn't been projected to win so strongly the day before the election, because some democrats leaning voter just decided to skip the vote since they though it was in the bag. This time around I don't think you can make that assumption, since the turnout was exceptionally high. I think it's important to note that vast majority of people care very little about politic and are extraordinarily uninformed (there's something like 10% of the electorate that can't even give the most basic information about the two party), so a lot of people who barely know what's happening voted this time around. Many of them probably just voted for whoever they heard about the most (ie Trump, incumbent candidate have massive advantages because of effect like these). These people most likely do not answer in polls which can skew results when participation is high.
The problem is either the statistical methods, the polling procedures, or Trump supporters dodging or outright lying about polls.
By the way, if you're concerned about "liberal arts" pollsters, deep state, universities who will give your address to Antifa. We would legit be fired on the spot if we did this, our work would be discredited, and we would have to be a shill for one side or the other for cranks.
I think what happened is that they undersampled people(Trump supporters), and then extrapolate that smaller sample to assume all Trump supporters think this way, and then just increased the margin of error, and called it a day.
The problem with that is that if 40 Trump supporters or Caucasian men with no degree were polled vs an expected 50, then those 40's opinions would be assumed to be the 50.
The above may work for polls about issues, but not about voting patterns.
There was one poll where the only African American they could find was a Trump supporter, there was another foreign policy poll by a very respected professor on Israel-Knesset members that didn't even poll any Muslims-Knessets because it was hard.
Also fuck Amazon-M-Turk, a bunch of liberal college students taking polls, and people assumed it's the gospel.
There was also another poll where they oversampled foreign policy hawks and one where they undersampled young people who were foreign policy hawks. If there were no foreign policy hawk we wouldn't have Lindsey fucking Graham, Rubio, and Cotton, if there were majorities of people saying we should nuke the Iranians, the Trump admin may have gone to war with them to distract from Covid.
I hope the polling community learned a lesson from all of this, but they won't. Every concern will be brushed off with an excuse or it will just work.