I"ve always sbeen suprised by Deleuze's embrace of empiricism. I intiially thought that it was a return to an old style empiricism in the guise of a reaction against concepts in Anglo-American philosophy, or an appeal to the primacy of lived experience in phemonenology.

But the former (traditional empiricism) sits at odds with Deleuze's conception of philosophers as creators and inventors of concepts. In Dialogues he writes:

I have always felt that I am an empiricist . . . [My empiricism] is derived from the two characteristics by which Whitehead defined empiricism: the abstract does not explain, but must itself be explained; and the aim is not to rediscover the eternal or the universal, but to find the conditions under which something new is produced (creativeness)...Empiricism starts with ...analysing the states of things, in such a way that non-pre-existent concepts can be extracted from them. (p. vii)

And the difference from the latter? I guess that is indicated by a transcendental empiricism (ie., the conditions
of experience are not themselves given in sensible experience) being an ontology that would grant primacy to difference rather than identity.

Update Jan 3 2006
In the a href="http://www.levibryant.com/transemp.pdf">article referred to above Levi R. Bryant says:

On the other hand, transcendental empiricism has epistemological implications insofar as knowledge too must be formed in a process of individuation. In other words, we arrive at a knowledge of difference through processes that are themselves ontological and productive of difference. As a consequence, knowledge cannot be conceived in disembodied representational terms that would maintain a strict distinction between the knower and the known, but must instead be thought in terms of becoming, wherein the thinker and the thought themselves form a new difference.