With Intellectual Ventures, Nathan Myhrvold Out to Create “Invention Capital” Industry—and Stop Hurricanes, Malaria, and Global Warming in the Process (Part 2)
Yesterday, we ran the first half of a sit-down interview with Nathan Myhrvold, cofounder and CEO of Intellectual Ventures, the Bellevue, WA-based invention laboratory and investment firm. Myhrvold, the former CTO of Microsoft (and an Xconomist), placed his current company’s goals in the context of venture capital and private equity, arguing that there is a real need to create what he calls an “invention capital” industry.
In what follows, Myhrvold talks about the lessons he learned in forming Microsoft Research, the differences between research and invention, some ambitious and far-out projects from Intellectual Ventures (e.g., invisibility, geo-engineering), and the motivation behind his firm’s upcoming expansion into five Asian countries.
Xconomy: Before we get into specific projects and inventions, what all did you learn from Microsoft Research that’s applicable to Intellectual Ventures?
Nathan Myhrvold: I have a theory that R&D is a great investment, a fundamentally good business. Using the human mind to go from nothing to something is a hell of a trick. And there’s nothing fair about it. A guy like Einstein can come up with all these things, but so can people who aren’t actually all that smart! There are people dumber than Einstein who’ve made amazing contributions.
So I believe you can make money with research, or invention. But you need a certain scale factor. Let’s say I have this idea called life insurance. If I just insured your life, it wouldn’t be worth it to either one of us. Insurance is fundamentally a risky bet, and to make it reasonable, what you’re buying and selling is variance. You need to have a large end limit to shrink the variance down. With Microsoft Research, I came to the conclusion that research could have been enormously profitable for Bell Labs, IBM, and others. It was profitable, but it could have been even more profitable. Xerox PARC could have made Xerox one of the most valuable companies on Earth. But most people screwed it up.
And after screwing it up, the lesson was mislearned that it’s impossible to be successful in this way. Most of Silicon Valley turned away from the notion of trying to do anything new. The implicit attitude was, hey, that’s why Stanford exists, somehow they’ll come up with new ideas. We’ll wait until that occurs. And then when companies got bigger, the size of Oracle or Sun or Apple, they said, “Well, keep doing that. Screw it, we’re not actually going to do anything really exciting.” And I thought, no that’s the wrong thing to do. If you have the scale at which you can afford to wait 5 to 10 years for a result, that was the key thing. If I say, invent something or do valuable research tomorrow, that’s an impossible task. But if I say, support 100 really smart people working really hard for 5 years, something great will come of it.
X: That’s what you had at Microsoft, because of its size.
NM: At Microsoft, we had the resources to do that. So I talked Bill [Gates] into starting Microsoft Research. It’s been hugely successful; they would say it’s one of the best investments they ever made, enormous customer value and shareholder value…To sum up, Microsoft Research is based on a similar idea [as Intellectual Ventures], with one twist. There, all I had to do was convince one man, and we could go ahead. After I retired from Microsoft, I wanted to keep going. I no longer had the one man to convince to do the whole thing. If you think about how to replicate the model, even if I’d gotten Bill to give me more money to do something else, that wouldn’t be the replicable model. So that’s where I came back and said OK, how could you do this on an even broader scale?
It turns out the way the world does this on a broad scale isn’t by saying this will be done by a government agency or by Bell Labs, a research lab funded by a monopoly business. In fact, the modern way to do it is to create one of these marketplaces where large investors are willing to put a small fraction of their income towards really risky things. And so if you create the right model, the world’s pension funds have more money than Microsoft or any of these entities. So if you create that venture capital-like, private equity-like thing, you have an even bigger pot of money than I did at Microsoft Research. And a few other constraints perhaps, but this was something that was ultimately going to have even more leverage than what I could do there.
X: So, besides fundraising, how is Intellectual Ventures fundamentally different from a research lab?
NM: Research is a valuable thing, and pays off. But there’s a difference between research and invention. The typical researcher picks a problem, and works on it no matter what—the LHC [Large Hadron Collider] guys are going to work on it no matter what. There’s lots of stories of someone working for 25 years before they get their great breakthrough; there’s also people who work for 25 years and never get that breakthrough. That’s part of the game.
Invention is different. Invention is about the act of creating something useful. So you don’t work for 25 years, from my perspective. The goal is to invent something, not to work on a problem. We view invention as something distinct from research and distinct from engineering. It can be inspired by either one, but when we say “invent,” we literally set out to create inventions. At the end of the day, if we learned something but didn’t create an invention, that’s great, it’s OK—but that’s not invention. Just like you’re a journalist, and you can have interviews that result in you learning things, that’s great—but ultimately if you don’t produce written articles, someone’s going to say, “Hey!”
X: Can you talk a bit about the invention process? Have you figured out the best ways to vet ideas and produce inventions?
NM: We have a bunch of ideas about that, and it turns out the rest of the world does too. If you look on Amazon for “creative thinking” or “innovation” or “thinking outside the box,” there are people saying, “I have the recipe for how to develop more ideas.” Maybe they do. I’ve bought enough books like that to cover this table. I can’t say I’ve brought myself to complete most of them. I think there’s a lot of different ways to be creative. We try to get interdisciplinary groups together. We try to have a lot of homework done up front so we’re smart about the issue. We try to have unique points of view and different ways in. If we get stuck, we move on—we might come back and revisit. I’d rather invent something than just sit there and all be scratching our heads.
We have a lot of different methods. There’s the “answer in search of a question” method: we take something from [scientific journals] Physical Review Letters or Nature or Science and say, hey this is hot off the press, is this good for anything? We try to do radical new approaches to old problems. So if someone comes up with a new idea, does this have real implications? Sometimes you find something that does.
Another [approach] is “wouldn’t it be great if…?” We have a number of surgeons who are inventors. A practicing surgeon is focused primarily on the health of his or her patients. They know what’s available in the market, but they don’t know what’s really possible. So we get them to ask for things. Don’t worry if you don’t know how to do it, just ask for something. And we’ll see if it’s possible. So we’ve come up with some pretty amazing things that way.
X: OK, so let’s talk about some more specific invention projects. Earlier (see Part One from yesterday) you talked about a new kind of nuclear reactor. Can you give us some other examples?
NM: We’re working on a huge diversity of things. Lots of medical devices, we’ve been doing cooperative work with the Gates Foundation on looking at problems of global health and development around the world. So we have a bunch of malaria projects. Malaria is not a very lucrative disease because even if you’re in the tropics, affluent societies generally don’t have it. Nevertheless, we’ve done a bunch of work on it, we’re pretty excited about that. The patents there have not issued yet. When you file a patent, for 18 months nothing happens, and then the patent office publishes [the application], so we’ve got nearly 650 published applications. My guess is the malaria things we’re working on haven’t published yet. We really got onto malaria [just] last January in a really serious way.
In solid-state physics, we have this way of classifying things as a set of phenomena that are “Maxwell domain” [classical physics] and “Schrödinger domain” [quantum physics]. But can you do things with Schrödinger that you would have done with Maxwell, and vice versa? So as an example, most of the ways people generate or capture photons [light] are Schrödinger domain. They’re lasers, LEDs, hot wires. On the other hand, you can make little antennas for all these things. You can effectively use a [classical] approach. So we have a whole invention area around optical antennas.
And “metamaterials” [structures with weird, unique properties]—we learned of a fundamental breakthrough, and we acquired the patent. We recruited the best guys in the field to be inventors with us, and we’ve done a ton of inventing in metamaterials. Now, metamaterials is a zero billion dollar market. You could write an article 5 or 10 years from now saying, look at those idiots, they spent all this time betting on this crazy thing. It may well be true, but we got in early enough, we got fundamentally interesting ideas—metamaterials seems to be powerful enough—it’d be a little like betting on the transistor in 1957. It would have been a good bet. We’re hoping there’s something here. As long as we don’t put all our eggs in one basket, that’s fine.
X: This is getting a bit abstract. Metamaterials have gotten attention recently because they might possibly be used to create an invisibility cloak. Is that the kind of invention you’re talking about?
NM: Invisibility—that’s one of the cool things you can do. This is an idea that literally could have happened in 1870, and why didn’t it? Eh, who knows? It’s pure classical physics. As it stands, it just happened a few years ago, there was huge controversy, people saying it was fundamentally impossible. Well, it turns out it isn’t impossible at all. So that whole thing is a terrific lesson. It’s the ability to create a negative index of refraction or gradient lenses of bizarre properties such as “transform optics” [related to the technical area of conformal mappings].
We’ve come up with a couple cool applications for invisibility, and we have some patents there. The trick there is, literal invisibility in the “Harry Potter” or “Romulan cloaking” sense is too hard and there’s not that big a market at the moment. What are other things you could do which would match the fact that this will be difficult and expensive, and you can probably only get it in a limited frequency band? We’ve come up with some. The obvious thing is how to avoid getting speeding tickets. But there’s more commercially interesting things than that. But [our patents] are not filed on them yet.
X: What about some other really ambitious projects?
NM: We’re working on hurricane stopping. We’re doing a lot of computer modeling. It’s so simple that it basically has to work. We’re doing the modeling just to make it really clear. Here’s the fundamental idea. Hurricanes occur when you get a very hot layer of surface water. I think most hurricane researchers would tell you 26 [degrees] Celsius is the cutoff. If the surface water never gets above 26, or 26.5, you get no hurricanes. And if it gets hotter than that, you get hurricanes.
Now the reason you get hot surface water is a funny runaway effect, like upside-down convection. Convection in a pot works because you heat the bottom of the pot, the water gets really hot, therefore it rises, and as it rises it stirs the water up. But if you heat water from the top, it doesn’t work—the top gets light, it doesn’t want to sink, it stays there. So that creates a funny temperature profile that’s stable against further convection. So if you could just stir the water, you’d have no hurricanes. So then the question is, can you come up with really cheap, very large-scale-effective ways of stirring the water and keeping the [surface] temperature low? The answer is yes.
We have a bunch of geo-engineering ideas. That’s a fascinating field. Both hurricane amelioration and geo-engineering are very controversial. In climate change, there are disaster scenarios that are plausible. Not proven, but plausible. The Greenland ice sheet collapses, the Gulf Stream stops, and the world goes to hell in a handbasket. If those are real, then don’t talk to me about conservation. There’s no possible way to conserve, or reduce or cut back or cap and trade. If there’s any real probability that a disaster will happen, then it will happen even if we stop today, because there’s such an overshoot in the system. If we stopped emitting carbon today, we’d have a century overshoot, and we’re not going to stop today.
So the only way to stop one of these disaster scenarios is with a geo-engineering intervention. It turns out we have three completely different ways of reversing global warming. Three totally different effects, each one of which is for relatively low cost—millions, not billions of dollars. Some activists hate geo-engineering because they say, look, people will use this as an excuse to keep polluting. I see that as a bit like saying, “Don’t stock defibrillators, I want people to have an incentive to stop eating glazed donuts.” People are going to eat the damned glazed donuts no matter what; if you have defibrillators around, you can save some lives. You ought to have that emergency response. Within a short timeframe, you can take the climate and dial it back to whatever you want.
X: If you’re talking about low costs, I take it you’re not into the idea of building a giant sunshade in outer space, or something on that scale.
NM: No, not the big sunshade. That’s geo-engineering, but it requires 23rd-century technology. There’s nearly trivial things I can convince you would work. Even if there are some “gotchas,” the gotchas are way easier than dealing with the Gulf Stream shutting off. So if you ever got remotely close to one of these disaster scenarios, this is what you ought to do. They’re all very simple physics, all things I think pretty much have to work. But hey, they’re not that expensive to test, so don’t take my word for it on whether they work! This is the kind of thing where the world ought to say, you know, let’s get some stuff out there and test them…It’s unclear what business model there is for us or our investors, but they’re kind of fun to think about
X: OK, so those are some pretty far-out projects. What are the more down-to-earth inventions you’re working on?
NM: The stuff we have that is here and now is lots of interesting surgical inventions. One of our inventors came to an invention session and said, suppose you have a tumor releasing metastatic cancer cells. How many times do they circulate [in the blood]? Turns out about a million times to circulate before settling down. You have lots of opportunity for intervention, six months to a year. There’s a finite amount of time, so it makes sense to intercept [the cancer cells]. And you don’t have to intercept every cell. If it were a short time, you couldn’t come up with effective way to get rid of them. But circulating a million times, if I take any vein, that cell will come by here pretty often. You’d like to have something that says, I can go put something in an IV in someone’s arm and have a good shot at intercepting. You should have a routine thing: as soon as someone suspects [cancer], put this catheter into a blood vessel that has the cancer detection mechanism. In the meantime, work out a surgery date and start on chemo, but if you can actually prevent metastasis from this day forward, it’d be a fantastic therapy to have. Our malaria work is similar to this. There it’s not about detecting—your red blood cells are infested with these things. We have ways of determining which red blood cells are infected. Eventually we could have a therapy based on mechanically getting rid of those cells.
X: Switching gears, I wanted to ask about your upcoming expansion of Intellectual Ventures into Asia, this fall. What are your plans there?
NM: The basic idea is a third way to come at the same idea of investing in invention. Most inventing organizations like universities and nonprofits have a tech transfer office or a technology licensing office. Most institutions in Asia don’t. Our goal is to create an entity to do two things. First, be an outsourced tech transfer agent for inventing institutions that don’t have one. We can provide the same functionality that you’d get from a technology licensing office. We evaluate ideas, we pay for them to be patented, we provide a little bit of funding for them, and then we license them. There’s no reason that smart people in Asia shouldn’t be able to get the same traction that Stanford provides.
The second thing, which ironically looks like there’s just as much opportunity, is we’ll also partner with tech transfer offices here [in America]. We’ve already done deals with some major universities in the U.S. where they say we already have a tech licensing office, but we wouldn’t mind partnering. It’s a bit like the way many venture capital firms will partner in a deal. They won’t go in alone, but they’ll come in with a friend. Or two private equity firms will get together to do a big deal. They have ambitious projects that they’d like to do, but they’d like to have an investment partner alongside. That’s what that deal is about. The difficulty is that we have to go open up offices all over the world. But we’re gearing up to do exactly that.
X: So the new offices will be in five countries in Asia, starting in late September. Where exactly are they, and how big will they be?
NM: They’re in China (Beijing), Japan, India, [South] Korea, and Singapore. That’s a good start—it’s enough for this year! We’ll find out how big, we’re doing 7 to 10-person offices right now. You always have this tension when you start up a new thing. We’re still feeling our way and figuring out what to do. I don’t want to pretend that we’ve got it all figured out. If we started with 100 in each office, I think our probability of failure would be close to 1. We’ll do an announcement at the end of September.
X: Why Asia, and what did you see in the invention market there?
NM: If you look at the biggest change in technology, the level of technology education and development in Asia is growing incredibly rapidly. One approach to that is to go start a venture firm in China. Everyone and his brother is doing that—including my brother [Cameron Myhrvold] at Ignition Partners, they have a Chinese fund that they’re affiliated with. If I was a venture firm, I’d probably do that too. I’m not a venture firm. The simplest, best way for us to participate in the growth of technology in Asia is to say, how can we help inventors there? The single best way for us to do that right now is to try to partner with inventing organizations, as opposed to trying to partner with individual inventors.
The model where we bring people into this room and brainstorm is great, but that’s for a relatively small number of elite inventors we can bring here. That isn’t going to help 1,000 professors of engineering in China. The way to help with that issue is to find a way they can get more traction and we can help. Because we understand the American legal system and patents and licensing way better than they do. On the other hand, they’re pretty smart people in their own technical domains.
X: You’ll be traveling to Asia to give some big keynote talks. How receptive have governments and universities been to the idea of partnering with you? Are they really interested?
NM: It’s a work in progress, but so far, yes, they’ve been very interested. They understand they’re not part of the technology economy. It’s unclear why they’d want to do it themselves. In terms of tech transfer offices here, some are very successful…but it’s like, do you do your own ads, or do you use an agency? Do you have your own lawyers, or use a law firm?
X: And will you eventually be nurturing individual inventors in Asia, and around the world?
NM: Not as intensively as this [in Seattle]. I’m here more than half the time; we fly them all here, we all speak the same language. It’s very intensive work with an elite group of inventors. You couldn’t generalize that to 1,000 engineering professors in China. So you want to nurture them, you want to support them in a bunch of ways, but you have to create something far more scaleable.
We’ll encourage people to come up with ideas. When you have an idea, you bring it to [a technology licensing office] and they say “that’s a great idea but someone already had it,” or “that’s a good idea but we don’t think it’s commercially interesting,” or “that’s a really interesting one, we’ll pay to get that all funded and get the patent written, and we’ll license it and we’ll send some money back.” A technology licensing office effectively vets ideas and handles the hard part—so that’s what we’ll wind up doing.