Agreed. They were talking about this on the FiveThirtyEight podcast last night. The general idea is that the electoral college greatly amplifies polling gaps in traditional races. At this point in the 2012 campaign, Obama had something like a 2-3 point polling advantage, and ended up with +126 EVs over Romney. In 2008, he was at 3-4 points ahead of McCain, and came out with +192 advantage. 2000 was probably the closest squeaker we've had in recent times, with bush polling with less than a 1 point advantage (it was something like 0.7pts if I recall), and won with a mere +5 EVs. If you remember the 1984 landslide, Reagan was polling 9 points ahead of Walter Mondale, and won by over 500 EVs.
At present, Clinton enjoys about a 6.1 point average advantage according to FiveThirtyEight, and 5 point advantage according to RealClearPolitics. Additionally, Trump is not currently have an advantage in
any of the polls that both sites use to calculate their averages, including the LA Times and Rasmussen polls that have had Trump in a consistent lead since the start of the year.
I'm not posting this to gloat or anything, but as
@Il Duce pointed out, all this data points to an inevitable Clinton victory, and the only question is how big of a victory it will be. As well, it remains to be seen how this will affect senate races.
If Trump manages to win somehow, the entire polling industry will need to take a very hard look at their models, and Nate Silver will probably jump off of a bridge.