The question in a sense answers itself. The category of the 'new' sustains itself by seeking its own value in negative relation to its invented other, the 'old', thus duplicating itself ad infinitum. A logic of perpetual new-old keeps the question alive, unavoidable, and thus, it seems to me, theoretically important.
Before deciding whether or not a new form of criticism is necessary, though, it might help to decide whether or not there is something 'new' to criticize in the first place. Assuming such a decision can be made at all, I imagine it can only be made locally, conditionally, and in the context of particular readings. Thus, habits of critical recognition continue to be essential, that is, noting if and in what ways a given work has been "remediated" (Bolter and Grusin 2000: 45) or refashioned from current or older art forms (how a particular hypermedia work borrows from collage or typographic poetry, for example). Manovich's (2001) effort to place new media practice in the context of early Russian montage cinema is one example. Equally important would be recognizing how existing critical terms (like 'image' or 'rhythm', for example, or even 'interactive' or 'experience') both do and do not help account for current reading and writing activities in programmable media.
Assessments of novelty and difference are valid exercises, useful, I believe, if grounded in habits of recognition. But perhaps this particular kind of assessment would benefit from a revised set of questions directed at specific works of hypertext/media, a few of which might be:
Manovich, Lev. (2001) The Language of New Media (Cambridge, MA: MIT Press)