I'm not just talking about swing states in the 2016 primaries though. Bernie won primaries in safe-red states like Kansas, Oklahoma, and Idaho, and he started at the bottom floor in that race while not even at peak popularity for half the voting season. The attitudes in the swing states are not necessarily "moderate", they're just fed the narrative that they need to vote for corporatist dems to sway conservatives or have nothing.
Minus some outliers, Warren is doing pretty similar to Hillary was at this point in time. My point is moreso that incumbents are supposed to have a much bigger bump than this, and Warren doesn't even have the benefit of Hillary's name and popularity. Putting any other factors aside, you'd expect a candidate in Warren's position to only improve and eventually win. The only presidential incumbent to ever win with these polling numbers at this point in time was Obama in 2012, and Warren's generally polling better than Romney was here.
The narrative about 2016 polling isn't what a lot of people think it is. The polls would've predicted an electoral win for Clinton, but it was far closer than what people seemed to say. Florida was well within the margin of error with Trump ahead in some polls. Ohio had Trump ahead in the last few weeks.
The major issue with the way polls were reported here is that two-way polling vs four-way polling gave very different pictures in many states. Michigan and Pennsylvania seemed like safe wins for Hillary in a two-way race, but four-way polling had them in virtual ties toward the end. When you look at four-way polling in Ohio and Florida, they were in a virtual tie in Florida while Trump was safely ahead in Ohio. The results weren't so much about Trump getting more votes than expected, but about the 3rd party candidates actually pulling votes somewhat accurate to the polling, when they typically receive less support on election day. I don't see this same factor coming into play again. (btw, I'm not saying 3rd parties were at fault for the 2016 results like some democrats do, but that the 3rd party voting was just one manifestation of the apathy toward Hillary).
Polling's rarely going to be an exact predictor of a result, but it generally gives you a good picture of what's going to happen. Trump's team knows this, and they'd be stupid to ignore polls just because statisticians predicted he'd lose in 2016. Now we have even more data to go by though, such as approval ratings and the 2018 election, which doesn't make Trump's chances look very strong.