During their conversation about the National Science Foundation, Drs. Robert J. Marks and Paul Werbos took a moment to discuss patent law. Their short rabbit trail ends with a grim analysis of patent law today: it’s now “almost impossible for the little guy” to secure a patent – and expensive.

Patent law was considered so important to the nation’s founders that they enshrined it in the Constitution, granting Congress the authority to protect the rights of inventors to their inventions (see: Article 1, Section 8, Clause 8). “Granting exclusive rights to the inventor is intended to encourage the investment of time and resources into the development of new and useful discoveries,” explains Cornell’s Legal Information Institute.

In 2011, then-President Barack Obama signed the America Invents Act (AIA). It was the most significant change to patent law since 1952. The bill passed through Congress with overwhelming support from both parties. It was widely hailed as a great achievement to fix some of the problems faced by the patent office and to bring about innovative progress. But it did not pass without its critics, who argued that it would only benefit Big Business while hurting smaller, independent inventors.

This portion begins at 15:08. A partial transcript, Show Notes, and Additional Resources follow.

Paul Werbos: I think I got something like seven patents about just the late 1990s, and those patents already address many of the technologies of the future that they still haven’t fully caught up to because they’re just a whole lot of pieces.

Robert J. Marks: Oh, how many patents do you have?

Paul Werbos: They’re probably expired by now, but probably about, I don’t know, seven, eight, something like that.

Robert J. Marks: Did you get these when you worked for NSF or are these after?

Paul Werbos: Yes.

Robert J. Marks: So it’s assigned to the government, I suppose?

Paul Werbos: You asked about how NSF is doing it.

Robert J. Marks: Okay. Yes.

Paul Werbos: And I’m glad I won’t have to get into every gory detail because they got ups and downs like the whole world. But for many, many years, the Office of General Counsel, the number one lawyer for NSF, there was a guy named Charles Brown. And when I think of the greatest glories of NSF in its absolute peak period, Charles Brown, the general counsel, had a lot to do with it. He went to the Harvard Law school. He learned about freedom. He learned about free speech, intellectual truth. I mean a whole lot of very basic principles. He understood the spirit of the Constitution. And his way of dealing with regulations and rules, he respected the rights and the privacy of everybody and the integrity of the system. So at some point the idea came up, gee, I’ve got a few ideas. What should I do about patenting? And I talked to him.

Paul Werbos: And he said, “We now have a clear policy.” And in those days we did. He said, “The clear policy is if you come up with a patentable idea, we want you to get a patent. And it won’t be an NSF patent. We will have to make a decision, whether the idea came from your work as an NSF employee, or did it come from outside of your job? And if it came from your work, the government has an interest, which is like a no free use of the thing. But even then you’re allowed to patent and we encourage you to patent because if you don’t, the Chinese will patent it two years after you invent it and we won’t be allowed to use it,” he said.

Robert J. Marks: Is that right? Do we have a patent treaty with China?

Paul Werbos: Well times have changed. Patent law has changed.

Robert J. Marks: Okay.

Paul Werbos: Patent law has changed. I’d probably have 10 more patents if they hadn’t changed the patent law.

Robert J. Marks: Okay.

Paul Werbos: But they have substantially changed the patent law in ways that are not so good for small inventors in all fairness.

Robert J. Marks: Yes. I acted as an expert witness in a number of cases involving kind of computational intelligence, artificial intelligence, and things have changed. I mean, it’s almost impossible for the little guy anymore.

Paul Werbos: Right. So the only thing the little guy can do is talk to the people in China who are willing to pay for that kind of stuff. And a lot of that is going on.

Robert J. Marks: I have a friend who has a bunch of patents and he says, all the patent is good for is it gives you the right to sue people. And that’s basically what it is. And it costs big bucks to sue people.

Show Notes

• 00:42 | Introducing Dr. Paul Werbos
• 01:27 | Overcoming NSF disinterest in neural networks
• 07:30 | Convolutional neural networks
• 10:54 | Deep learning
• 15:08 | How is the NSF doing?
• 21:53 | Recent major advances in machine intelligence

Additional Resources

• Dr. Paul Werbos at IEEE.org
• Paul Werbos’s website
• Paul Werbos’s PhD dissertation introducing error backpropagation used today to train artificial neural networks
• Talking Nets: An Oral History of Neural Networks at Amazon.com
• Paul Werbos’s 1972 Proposal to Harvard for Backpropagation and Intelligent Reinforcement System
• Perceptrons: An Introduction to Computational Geometry by Marvin Minsky and Seymour A. Papert on Amazon.com
• The National Science Foundation’s website
• Geoffrey Hinton, British-Canadian cognitive psychologist and computer scientist
• DDDAS: Dynamic Data Driven Application Systems at NSF.gov