The question of whether Hillary Clinton is uniquely polarizing is actually pretty hard to answer. For instance: The metric you use matters quite a lot. If you're going by how many voters "definitely would not" vote for her, she's less polarizing than John Edwards, Rudy Giuliani, John McCain, Fred Thompson, or Mitt Romney. But some say that number is a function of name recognition -- that folks are sure they won't vote for candidates they don't know. So if you're going by favorability numbers, then Clinton's 44% unfavorable is fairly high. But that may just mean she's further along in a process that any high profile Democrat will undergo. At this point in the 2004 cycle, John Kerry's unfavorables were between 13% and 20% -- by the time the election rolled around, he was in the mid-40s, posting numbers pretty comparable to Hillary's.
So that's the question: Not whether Hillary Clinton is more polarizing right this second. Given that everyone knows who she is, that simply has to be true. But whether she'll be more polarizing than John Edwards after eight months of haircut and hedge fund smears, or Barack Obama, after an election full of madrassa insinuations. Clinton's numbers probably reflect the end point of that process -- she's been smeared with maximum energy and efficiency for 15 years now. Edwards and Obama haven't, but if either captures the nomination, the GOP's attack machine will boot up, and do to them exactly what it did to John Kerry. If someone has an argument for why, at the end of that political war, they'll be less polarizing than Clinton, than that's a fair comparison. But the current numbers are not.
In short, Klein thinks that the unfavorables of Kerry, Edwards, and Obama would increase into the 40s by Election Day, while Clinton's "probably reflect the end point of that process" and wouldn't increase further.
I agree with the former point, but not the latter. As commenters on Klein's post note, there's no reason to think that Hillary's long political career will prevent her unfavorable numbers from increasing. She hasn't come under sustained political attack since before the Monica Lewinsky scandal -- how could months of negative campaigning against her not push up her unfavorables? They might not go up as much as other Democrats' would, but each extra percentage point is far more damaging to her general election prospects than it would be to theirs.
Consider the trajectory of President Bush. By late 2004, he was very well-known and had unfavorables in the 40s, but they've since increased to well over 50 percent and might reach the 60s by the time he leaves office. Hillary's profile is like his during the 2004 re-election campaign and it's more than a year until the general election. How sick would America be of her by the end of her first term?
In addition, as Tapped commenters point out, even if Edwards and Obama would end up with comparable unfavorables to Hillary, more people have a visceral hatred of her than the other candidates, which may boost GOP turnout, organizing, etc.