Though Romney won Michigan, Douthat says he "lost his general election narrative": 

This is not where Romney expected to find himself at this point in the campaign, with months and months of careful positioning undone by several weeks of gaffes and defensive political maneuverings. But between his verbal miscues and his clumsy attempts to defend his right flank on policy, the likely Republican nominee is suddenly headed for the kind of political and ideological cul-de-sac that losing presidential candidates often end up occupying.

Wilkinson counters:

I think the primary is obscuring just how strong [Romney] might turn out to be as a centrist candidate in the general election. 

Ezra Klein isn't sure:

The latest Politico/George Washington University poll, for instance, finds, "Romney is bloodied after nine contests, five of which he has lost. Only 33 percent of independents view him favorably, compared with 51 percent who see him in an unfavorable light. In a head-to-head match-up against Obama among independents, Romney now trails 49 percent to 37 percent." Losing ground among independents suggests a real weakness in the general election. But it might be meaningless. Those independents might simply be reacting to the primary, and they'll come around when Romney transitions to his general-election campaign.

Tomasky thinks the improving economy is likely to doom Romney in the general, should he make it that far:

[Romney] will return to the campaign trail with Republican professionals nervously wondering: What awkward thing is he going to say this week? And can he win Ohio? What happens to us if he can’t? And somewhere in there, namely Friday, the January jobs report—expected according to early accounts to show at least another 170,000 or so private-sector jobs gained, and the jobless rate maybe dropping again—is going to come out, and if it follows expectations, it will diminish his rationale that much more. 

Cohn notes that Romney is losing low-income voters:

In a Republican primary, or at least this Republican primary, you can prevail by losing among all voters making less than $100,000. But it’s tougher in the general election. Romney and his advisors can take comfort in the fact that the downscale conservatives who voted for Santorum will generally support the Republican nominee, whoever it is. They may not love Romney, but they hate Obama, and that will be enough to get them to the polls. Still, Romney has to win over at least some middle class votes to win in November. And he’s shown very little ability to do that.

Jonathan Bernstein believes the nomination fight is over:

[Romney] didn’t need Michigan – he would have had a good night even if he lost by a few points there, since he would have gained in the delegate count – but it helps make it more clear than ever that he’s going to be the nominee. The short-term question now is whether he can get a reasonable bounce out of it and into Super Tuesday next week, on March 6. As I’ve said after each round, that depends more on the spin war than on the actual election returns. 

Drum thinks the candidate will be bruised and battered by the convention:

Romney is still the presumptive nominee, the winner by default because everyone else is unthinkable. And after limping through the spring and finally staggering into the convention like a punch drunk Rocky Balboa, guess what? Not only will he have to face Apollo Creed in the main event, but it looks like the Greek Streak, Olympia Snowe herself, might be pecking away at his kneecaps the entire time.

Cassidy sees a tough road ahead:

If Romney isn’t quite restored to his position as "Mr. Inevitable," he’s "Mr. Very, Very Likely." But with Super Tuesday only days away, it is difficult to see how he can change the underlying dynamic of the race, which is slowly but surely dragging him down. What he really needs is an early end to the primary contest, but of that there is still no sight. His victory in Michigan was narrow enough—with 98 per cent of the vote counted, he was leading by 41.1 per cent to 37.9 per cent—for Santorum to harbor hopes of beating him in Ohio, the biggest prize next week.

Alex Altman looks to Super Tuesday:

There’s a week for the winds to shift, but the Super Tuesday forecast is mixed for each of the candidates. Romney will be heavily favored in Massachusetts, Vermont and Virginia, where only he and Paul are on the ballot, thanks to the organizational deficits of their rivals. But polls show Santorum with large leads in Tennessee, Oklahoma and the key bellwether of Ohio. Meanwhile, Gingrich has grabbed a big lead in Georgia, the biggest delegate prize of the night and the state where he served in Congress. A Super Tuesday split would ensure that the bruising Republican nominating fight proceeds indefinitely.

And Noah Millman sizes up all the contests in March:

Even after winning Michigan, Romney could still wind up running a brutal gauntlet in March, and looking a lot more like a loser than a winner at the end. But even in that realistic-worst-case, he’ll still be well ahead in delegates. And he’ll still be the only one with a plausible path to the nomination. And after California, New Jersey and Utah vote, he’ll have it. Assuming he still even wants it by then.

My initial take on the night here.