I am not a fan of shooting the messenger. And there are some valid things to consider in this article. But there are also significant flaws as well. I'll take them by category:
1. THE POLLS ARE WRONG - The article even admits that the pollsters made some changes to their methodologies since 2016 (when they weren't that far off). They've only gotten better. They still aren't perfect. But they are better
This is important because while the polls have gotten better (and there is more scrutiny as well to which are good and which are bad polling organizations), Joe Biden's lead is significantly larger than Hillary's. At this time in 2016, the national poll of polls had Hillary ahead by about 2.7% (she won by about 2.1%). Biden is ahead now by about 8.4%. The battleground polls are the same. So for Trump to win, there would have to be a polling error several times the magnitude of 2016.
I would also note that the internal polls are showing the same thing as the public polls. We know this because campaigns plan around the results of their internal polls. And we see both Biden and Trump in what were easily red states last time such as Georgia, Texas and Iowa.
2. THE SHY TRUMP VOTER IS REAL - Maybe. But first, the article makes no effort to quantify these voters. How many are there? I have seen no evidence, either in 2016 or 2020 that this is a significantly large number.
Also, Trump is now the incumbent President who has demonstrated (along with his supporters) a willingness to target those who oppose him. This has led to a phenomenon that the article doesn't even mention, which is the shy Biden voter. I know they exist because I know some personally. Particularly people who live in rural, red areas where coming out publicly for Biden would literally be dangerous. But these people know how they are voting on election day.
Is the shy Biden voter numerous enough to make a difference? I have no evidence on that point. Which means I have exactly the same evidence that the author of this article has for his proposition.
3. THE YARD SIGN TEST - As someone who has run for office numerous times and ran the campaigns of many other candidates, I can say definitively that there is no such test.
In very partisan areas, people put up yard signs of the popular candidate to virtue signal, and avoid putting up yard signs of the less popular candidate to avoid trouble. I've seen many campaigns where the candidate who ultimately looses has double, triple or more yard signs than the eventual winner.
The voter registration gap is a better argument. It's unclear if this is Democratic voters who have been voting Republican for years finally switching their parties or something else. However, I can say that so far, exit polls of those who have already voted show that 12% of Republicans are voting for Biden, while only 4% of Democrats are voting for Trump. So it's not clear numerically that this registration differential matters.
4. KEY DEMOCRATIC VOTING BLOCKS ARE CRUMBLING - Trump has made modest gains among young African American men, by which I mean that their turnout is lower than Biden had hoped. However, black turnout among older African-Americans is through the roof and they are virtually all voting for Biden.
Beyond that the article fails to mention key Republican voting blocks that are crumbling. For example, Biden is performing up to 20 points better than Hillary in some polls among senior citizens. The combination of Trump's botching of the COVID situation along with his mocking Biden for being "senile" and his general demeanor have just killed Trump among a group he won big last time and is losing big this time.
But seniors are not all. As Trump himself refers to at his rallies, his support has utterly collapsed among suburban women voters. Also, Trump never did well with young voters, but their numbers are up significantly from 2016.
All of this is not to say that Biden can't lose. Anything can happen on election day, or more likely, Trump could have some success suppressing votes in the courts after election day. But if there was such a thing as a disinterested observer who looked at all the available data, they would have no choice but to conclude that Biden is going to win going away.