The best case for saying that “gaffes matter” is that actual voters are persuaded to change their minds because of the gaffes. If they don’t, then it’s tough to argue that “gaffes” are really “game-changers.” And, in fact, usually voters don’t change their minds. See, for example, Michael Tesler’s and my analyses of the impact of “the private sector is doing fine.”
The best argument you can make about these gaffes is sort of a woolly counterfactual: “Well, if it hadn’t been for the release of Romney’s video today, Romney would have been able to accomplish X, Y, and Z, which would have helped him win the election.” Like any counterfactual, there is some plausibility—yes, Romney would rather talk about the unemployment rate than these comments.
But like any counterfactual, it’s predicated on assumptions about what the world would have looked like without these comments. And given the tenuousness of any such assumptions, and the (at best) small effects that single events in any presidential general election campaign tend to have, I would stop well short of calling this video “devastating.”