I read the whole thing and the author is regurgitating the Democrats’ lines word for word.
National matchup polls at this time are meaningless.
National matchup polls at this time are generally somewhat meaningful but not precise, and not a 100% guarantee. A poll saying that Pee Wee Herman (I once joked about Al Franken similarly, but that was in his SNL days) would lose to Hillary is probably accurate, even if the Witch's margin of victory might be off by several percent.
Polls today on the Cruz/Hillary, Trump/Hillary and Jeb/Hillary (Jeb can still dream, can't he? It's his turn, and Mommy wants it for him.) margins are similarly meaningful. History shows that polls at this point are a much better predictor than flipping a coin. History also shows that polls can be wrong on the outcome - that's what the political campaign is supposed to influence.
In Trump's case in particular, historical spring polling results are less likely to be meaningful than usual. In national elections in particular, polling data from the spring are more likely than usual to be meaningful. I have my own guesses, but I will not claim they have any validity, since we have no data on a candidate like Trump.