Kant has several arguments against various broadly Cartesian versions of ontological arguments (by Descartes, Leibniz, Wolff, and Baumgarten), but the most famous is the being-is-not-a-real-predicate argument:
Being is evidently not a real predicate, that is, a conception of something which is added to the conception of some other thing. It is merely the positing of a thing, or of certain determinations in it. Logically, it is merely the copula of a judgement. The proposition, God is omnipotent, contains two conceptions, which have a certain object or content; the word is, is no additional predicate—it merely indicates the relation of the predicate to the subject. Now, if I take the subject (God) with all its predicates (omnipotence being one), and say: God is, or, There is a God, I add no new predicate to the conception of God, I merely posit or affirm the existence of the subject with all its predicates—I posit the object in relation to my conception. The content of both is the same; and there is no addition made to the conception, which expresses merely the possibility of the object, by my cogitating the object—in the expression, it is—as absolutely given or existing.
This has a certain amount of plausibility, as Kant's thaler example shows. If we are arguing about whether you have a hundred thalers, for instance, then obviously 'your hundred thalers' has to mean the same thing regardless of whether we conclude that your hundred thalers exist or not.
But a great deal of the plausibility depends on the fact that we are basically treating all attributions of being as having exactly the same account, and (unsurprisingly, given his other commitments) Kant takes this to make it a logical notion that describes not an object itself but how the object is related to our thinking. If, however, we accept the view that 'being is said in many ways', this entire line of thought becomes much less reasonable.
When we say that "being" does not add to the conception of other things, or that it does not perfect it or make it more real, we mean it in something like this sense: reference to being, in some way, is involved already in every conception. If I offer a hundred dollars, but then go to say, "And I will do one better than that: the hundred dollars will also have being", you would likely be puzzled at the implicature that the hundred dollars originally offered did not have being. And this is quite general; if I talk about a something as possible, I am talking about it as a possible being. Whenever we conceive of something, we conceive of it as being in some way, somehow; it at the very least has being-as-an-object-of-cognition.
If, however, being is said in other ways, then the argument doesn't carry through for other things. In the context of the ontological argument, we have to get to a conclusion in which we are saying that something exists or has actual being; this is not at all the same as being-qua-conceived, and and the argument that the latter is not a real predicate adding to, perfecting, or making more real the object fails. The ontological argument involves a movement to more than just being-qua-conceived, and therefore it becomes a live issue whether being in this 'more' sense is a real predicate, perfection, or whatever.
In some sense, this is not a criticism of Kant; the being-is-not-a-predicate argument is not Kant's only argument against ontological arguments, the argument arguably does work against at least some Cartesian-ish versions of the ontological argument, and Kant himself has a lot of reason to argue this way, tied to the basic features of his transcendental idealism. But 'being is said in many ways' in and of itself gives a non-Kantian a reason to reject this particular line of argument as inadequate.
