Want CNET to notify you of price drops and the latest stories?

Building a new-idea factory

At his new company, ex-Microsoft bigwig Nathan Myhrvold gathers well-known inventors for "invention sessions." Will the idea pan out?

Michael Kanellos Staff Writer, CNET News.com
Michael Kanellos is editor at large at CNET News.com, where he covers hardware, research and development, start-ups and the tech industry overseas.
Michael Kanellos
10 min read
Golf really isn't one of those things ex-Microsoft execs do in retirement, and that's true in spades for Nathan Myhrvold.

Since stepping down as the company's chief technology officer five years ago, Myhrvold has pursued paleontology, a long-standing hobby, by funding dinosaur digs. But Myhrvold has more on his mind these days than searching for prehistoric bones. He's also the founder of Intellectual Ventures, a somewhat mysterious and controversial start-up.

Look at Bell Labs, look at IBM Research. They invent SQL, and Larry Ellison makes all the money from it!

Combining the disciplines of a venture capital firm, a think tank and an intellectual-property firm, the company gathers well-known inventors for "invention sessions," hoping the get-togethers will one day result in hugely profitable patents. Investors are said to include Microsoft, Google and Intel. Myhrvold sees the company as a forum where brilliant minds will have the room and luxury to simply create.

But critics take a dimmer view of the venture. They say it will encourage patent litigation and, paradoxically, discourage innovation. Whatever your point of view, Intellectual Ventures could well become one of the central figures in the ongoing patent debate. Myhrvold spoke with CNET News.com about the decline in invention, his company, and why some people get bent out of shape when the conversation turns to intellectual property.

Q: You've talked about the difference between research and invention in the past. What is the difference?
Myhrvold: Most engineers who work for companies doing technology are paid to build products, not to invent something new. They do invent things new, but the percentage whose actual job title is inventor or whose job function is primarily inventing is very small. Almost no company has people who are really dedicated to invent new technology without worrying about whether it's in a new product or not.

Can you give an example of something in the past 20 years that is an invention, versus something that emerged from research as an extrapolation of existing products?
Myhrvold: The graphical user interface was a hell of an invention. The microprocessor was an invention. Inside every technology product is a litany of different inventions. Nevertheless, it turns out that not only are people not inventing as much, but the inventing that does occur is highly constrained, and it's constrained because companies try to do what will work within their next product release cycle. And academics do either the thing that their graduates could get a thesis on or the things that will get them a paper and get them tenure.

How did we slip into this situation?
Myhrvold: It's risky and people don't want to fund it. Frequently, breakthroughs happen outside the area that you originally started working with. The guys at Xerox PARC had some clear breakthroughs: the graphical user interface, networks, tons of stuff that totally contributed to the PC world as we know it today. But they were kind of in left field from the people who were funding them, which was a copier company.

At Microsoft I had to fight this perception in order to start Microsoft Research. The received wisdom was that it was silly to invest in research because--look at Xerox, look at Bell Labs, look at IBM Research. They invent SQL, and Larry Ellison makes all the money from it!

Being an inventor almost sounds like a 19th century sort of occupation.
Myhrvold: Exactly! The entire 19th century was about doing exactly this. Every important invention was some crazy individual, and it was an established profession. If you were an inventor, people would

recognize what that was. If you go to a cocktail party now and someone says that they're an inventor, it's very strange. Most people's idea of an inventor is that crazy guy from the "Back to the Future" movies.

So how did you come up with Intellectual Ventures?
Myhrvold: What we set out to do is to say, "Let's honor invention and inventions. Let's take a company whose only focus is investing in invention." We're not going to make products. When I say we won't make products, it's not because products are bad--products are wonderful--but the world already has a very well-developed infrastructure for dealing with products and product-oriented ideas and product-oriented companies.

If you have a new idea about a product, you put together a team, go to Sand Hill Road, get some funding from VCs, and you start your company. So I said, "OK, let's go where they are not. If we get great ideas, we'll find somebody else who can incorporate them. We'll license it off, or we'll do some kind of venture, or we'll start a company." The exact form depends a lot on the ideas and circumstances.

We try to recruit inventors before they've had their idea and say, "Look. You're a person who's had a track record of lots of interesting invention. We'll support you before the idea, and you'll get together with us, and we'll try to brainstorm and create some new ideas."

Who are you working with?
Myhrvold: We've got some full-time inventors, but most of our senior inventors are not full time. Mostly, they have a job. They could be a professor, they could be a consultant, they could be retired. Typically, if they work for an existing company, they will sell their (brainpower) to that company and so they can't come and invent with us on the side.

Can you give us some names?
Myhrvold: Edward Jung and I, who were both at Microsoft, work here. And we have a bunch of other folks, Leroy Hood, who invented the DNA sequencer and tons of other things in biotech. He's had his own research institute called the Institute for Systems Biology. But he invents with us.

But a guy like Hood could go to Sand Hill Road and get funded.
Myhrvold: Yes. In fact, many of our inventors have had successful product companies in the past. But ironically, the guys who have had that success are the most likely to sign up, because they know what a pain in the ass it is. It's like five years of your life.

If you're a very inventive person, what I typically say is, "Look, if you work with us, you'll get some equity in your invention, and then we'll go to license it to people later. Net-net, you'll probably get less than if you were founder of the next Cisco, but it won't take five years of your life either. So probably you'll get a lot less for each one, but you'll have so many more ideas and you'll have your whole life." It's a different kind of a trade-off.

What sort of areas are you looking into that have been stymied by a lack of invention or really are ripe for more invention?
Myhrvold: We're quite wide-ranging. Mostly, we're in communications, IT and biotech. In biotech we don't do anything involving new molecules. Mostly what we do in biotech is either tools or biomedical instruments, stuff that involves a mixture of disciplines between computing and biotech or medicine.

Almost like the IT side of the biotech.
Myhrvold: Exactly. And then there's core IT things. We've had quite a bit of effort in electronics and solid-state physics. Nanoelectronics, photonic band gap materials--stuff like that. We tended to try to work in areas where we feel the world isn't doing a great job, and that falls into a couple of categories. One is in areas that interface between two disciplines. There are great biotech people, and there are great computer science people. The world is much (worse) at getting them together in the same place than you'd think.

Another thing we tried to do is we tried to work five years out, because almost all of the world's engineering work is for between zero and three years out. Almost no people in companies are paid to deliberately work more than three years out. I said deliberately just because the product can slip. It can take longer than you think to get it there.

Will Intellectual Ventures operate kind of like the IP firms we know today, like Rambus? Or is it somehow different?
Myhrvold: We don't know. Sometimes you get some intellectual property that is really interesting but it's really different from what the

rest of the world wants to do. In that case what you do is either take that as a hint or you start a company or do some other entrepreneurial activity to make it happen.

Who are your investors?
Myhrvold: We don't comment on who our investors are. You need to have a very long-term perspective to invest here, because it takes you three years before you get a patent.

One thing that intrigues me about the whole patent area is the emotion it engenders. If you bring up patents to some people, they really get wound up.
Myhrvold: Yeah, it's very funny to me, and it's a cultural thing in that it depends on what industry you're from. In the medical device industry nobody gets worked up.

Most inventors don't get paid for their inventions in proportion to their value.
Myhrvold: Try it. Call somebody in Medtronics or U.S. Surgical or anyone of a hundred little start-ups that do medical devices. People generally don't have any problem with the patent system. They may argue about a specific patent being valid or not. The same runs true with biotech. Go talk to biotech guys and ask them if they want the patent system abolished. They say, "My God!"

The one exception--and frankly, if you look at the world overall, it's one tiny blip of an exception, although it's one where I set my career--is in the computer industry. In computing, there has been a strong sense that patents were not the fundamental secret of success of most of the big companies. Oracle is very up front about them, saying they copied the SQL idea from IBM. And certainly Apple and Microsoft both learned a lot from Xerox.

There are exceptions. Intel, IBM...
Myhrvold: IBM has turned them into a vast resource, and it's very hard for me to feel bad about IBM doing that. They supported R&D for a really long time when lots of other people didn't. If one of the ways they justify keeping the doors open at Watson Research Center is licensing the patents, I'd say hallelujah. I had such a hard time making the business case at Microsoft for Microsoft Research initially. The only reason I got Microsoft Research started is that I could convince Bill Gates that we would develop stuff that would go into products and make money. It turns out we did, and I think if you ask Bill today, he'd say Microsoft Research was one of the best investments we ever made.

Here's a true IBM story. People at IBM Research would fool around with lots of things in fundamental physics. They started fooling with lasers very early. Well, somebody there burned themselves on the lasers, at least as the story goes. Then they filed the first patent on the laser ablation of human tissue.

Is the patent still valid for Lasik surgery?
Myhrvold: Yes. IBM made a ton of money on licensing and then ultimately sold the patents to Lasik. And they sold the patents for, like, 20 million bucks! Now, I think that's a great story--that you can justify having some guys screwing around with the latest thing, which was at that time a high-powered laser.

And the person who got burned got a sick day, no doubt.
Myhrvold: They got a nice little plaque, I'm sure. That's actually another thing: Most inventors don't get paid for their inventions in proportion to their value.

How about the complaint that IP companies just file vague patents to generate royalties?
Myhrvold: The patent quality issue is one which does come up, but usually it comes up from people who don't like patents already. It's a rationalization for their position. The majority of patents probably are very valid. The stock market has stocks of companies that are flaky and questionable, right? But does that mean we should just avoid all public security markets?

But the subject of patent reforms seems to come up a lot. Are there particular things that you think could help to eliminate some of the conflict?
Myhrvold: The most basic one is letting the patent office keep patent fees. Instead, Congress tends to divert the fees to do other things, and that's just criminal. I mean, here you've got this government office that has got tremendous demand. The demand is measured by the number of people that pay their fees. Shouldn't you at least use those fees to plow them back into making the system better?

And the examiners aren't paid that well either.
Myhrvold: So there's lots of reform that could be done of that nature. Keep the funding there. Pay the examiners better. A whole variety of other technical points about the patent systems are things that I think would really make a big difference.

How about compulsory licensing? That idea keeps floating up.
Myhrvold: There isn't compulsory licensing in essentially any other part of American life. The only thing that would be similar to that would be eminent domain, where the government can ultimately condemn your property and force you to sell it to them if they truly need it to develop a new highway or something.

What do you think of the complaints of how patent litigation is hurting companies? Some days it sounds like the trumped-up malpractice crisis of the '80s.
Myhrvold: Well, this is even stranger. We actually did a study on this. The overall number of lawsuits for patents is growing, but so is the overall number of patents. So explain that to me.

If you then look at it and ask, what fraction of those lawsuits are due to companies that have no products, the IP-only companies--it's about 2 percent. If you look at it and say what fraction of lawsuits are due to large technology companies, it's about 2 percent.