Showing posts with label patents. Show all posts
Showing posts with label patents. Show all posts

Saturday, November 10, 2007

Do Patents Retard Innovation?


Is the patent system broken? Supposedly a way to protect genuine intellectual property and spur innovation, patents these days seem most likely to wind up being used as a weapon of business warfare, and may actually retard innovation in many cases. Vonage and Research in Motion come to mind, as many observers think the patents Vonage is said to have infringed should not have been granted in the first place, and RIM had to pay what amounts to greenmail so its carrier and enterprise partners would not suddenly have to make all BlackBerry services "go dark."

In fact, it seems to be common these days to attempt to patent common business practices, obvious to anyone in the field. That leads to patent "trolls" buying up intellectual property and then suing companies as a business model.

Suing is a repugnant business model. And most patents seem trivial or--to a layman--overly broad. It is important to foster innovation and reward effort, and some innovations fit that bill. But isn't it obvious we ought to encourage people to work on really hard problems, and reward them, rather than encouraging lots of trivial stuff? Sure, it sometimes is hard to distinguish between an idea of significance and "prior art."

Now there's a big, socially useful problem that Google ought to be able to help with.

Whether it is the patent system or the way it gets used in business, something is out of whack. One might argue it is a necessary evil. Perhaps it isn't so necessary (at least the way currently practiced), though perhaps it often is evil.

"Tokens" are the New "FLOPS," "MIPS" or "Gbps"

Modern computing has some virtually-universal reference metrics. For Gemini 1.5 and other large language models, tokens are a basic measure...