This is well-put by Julian Sanchez. A strategy of defensive patents only works if the “innovation” you’re patenting isn’t worthy of patent protection:
But now think about how defensive patents work. Companies aren’t buying them—or buying into the services of companies like Intellectual Ventures—because they provide otherwise unavailable technical insights. The point, rather, is to acquire (or have access to) a bundle of patents that any potential litigant who sues you is likely to be “infringing” in their own products. Like nuclear weapons, the point is not to actually use them—but only to be able to threaten to use them if anyone else should deploy theirs against you.
This only works, however, if other companies are almost certain to have independently come up with the same idea. A patent that is truly so original that somebody else wouldn’t arrive at the same solution by applying normal engineering skill is useless as a defensive patent. You can’t threaten someone with a countersuit if your idea is so brilliant that your opponents—because they didn’t think of it—haven’t incorporated it in their technology. The ideal defensive patent, by contrast, is the most obvious one you can get the U.S. Patent Office to sign off on—one that competitors are likely to unwittingly “infringe,” not realizing they’ve made themselves vulnerable to legal counterattack, because it’s simply the solution a good, smart engineer trying to solve a particular problem would naturally come up with.
The whole idea of patents is, as Aaron Swartz pointed out to me yesterday, pretty odd. A copyright forbids copying. A patent forbids independent discovery as well. And it does so even though it’s famously common for certain ideas to be “in the air” and independently developed by different people.