Watch videos with subtitles in your language, upload your videos, create your own subtitles! Click here to learn more on "how to Dotsub"

Bob Metcalfe presents the Enernet

1 (1 Likes / 0 Dislikes)
[♪classical music♪]

Singularity University - Preparing Humanity for Accelerating Technological Change

Tonight's speaker is Bob Metcalfe. His talk is "Enernet: Internet Lessons for Solving Energy" Bob is the founder of 3Com, he is partner at Polaris Ventures, he is a co-chair for the Networks and Computing Systems Track at Singularity University. His company created PowerPoint, but this is the first PowerPoint presentation Bob has ever given. [laughter] and, oh yeah, he invented the Ethernet protocol.

Bob. >>Okay. Thank you. [applause]

I was born and raised in New York and that's my explanation for why I am so annoying. And my talk will be annoying. Then I spent 22 years clawing my way to the top here in Silicon Valley. That didn't help at all. So I'd like to apologize in advance for being annoying. I mean well, and I'm hoping that you'll get something out of this. My job is to finish at 8:30 sharp, so you may rely on the fact that we'll be done at 8:30.

That's my job.

Does anyone have any questions? [laughter] Well, what I do not want to be accused of is monopolizing all this time so you don't have time for questions. So I thought we'd get that out of the way. Would someone please ask a question, just to get warmed up.

Please. >>[audience member] What is Enernet?

What is Enernet. I have a stunning number of slides [laughter] to answer that. Enernet, since you asked, is a rubric for a set of lessons that I'm collecting from the Internet history. So this is a history talk. It's an applied history talk. And so Enernet is the rubric I've given for this study of the history of the Internet with the intent of applying that so that we can end up with cheap and clean energy faster than we would otherwise by, of course, mimicking the successful features of the building of the Internet. That's what Enernet is.

Should I get started, or is there another question?

Please. >>[audience member] If you were an entrepreneur in a garage, what sort of energy start-up would you build? If I were in a garage and-- Limited to energy? >>[audience member] Or whatever. [laughter] If I had more time and I were an entrepreneur in a garage, I would start a nuclear reactor company. Okay? >>[audience member] Why? We'll get to that later, but [laughter] because I view that as a very interesting problem, both from physics and sociology, communications, government policy. It's all wrapped up in that little mess called nuclear energy. So I'd do that just because it's really hard. It would be really hard. I saw the movie recently, "The China Syndrome." Actually, the first movie I watched was--my wife and I watched "Chinatown." And halfway through it, we realized that was about water in Los Angeles, and then we switched to "The China Syndrome," and that was the one. Jack Lemmon dies in that movie. Michael Douglas and Jane Fonda and Wilford Brimley watch Jack die. But he doesn't die from a meltdown, and he doesn't die from radiation poisoning. He is shot and killed by police officers. But no one really remembers any of that. A very unfortunate movie with good timing.

So I've already answered with Enernet is, and it's been mentioned that I was on the board of the company that developed PowerPoint. These slides are my first ever PowerPoint presentation. There was a 20-year gap between Forethought, which was founded in Mountain View, and then we sold PowerPoint to Microsoft for $14 million in 1987. And then Bill Gates and I had a falling out because I'm so annoying. So this is actually a Keynote presentation. [laughter]

The way I got into energy was the president of MIT, Susan Hockfield, several years ago gave a talk in which she decided that energy was one of the topics that MIT ought to focus on, and she would lead that effort as our president and being what we call a team player and an alumnus and a trustee, I decided to get into energy. That's how that happened.

But, of course, my day job is a venture capitalist, so I had to figure out how to be on the energy team, and the way to do that was to make a series of energy investments because that's what we do as venture capitalists. This is our portfolio, and being a venture capitalist, I should now speak for the next hour about this portfolio. That's what we do; we tout our portfolio. After we make the investment, then we tout the portfolios. But I don't plan to do that at all. This is my day job. This talk, however, is my hobby.

So one thing I've noticed about the Internet is that we use the Internet to build the Internet. So I think we should use the Internet to build the Enernet-- that is, to solve energy-- that is, to provide the world with an abundance of cheap and clean energy. And I'll return to that controversial point in a moment. I said abundant energy. And that's not the mindset. The first annoying thing I'm going to say is if the Internet is any guide, we're going to have abundant energy, not less of it. So we can use the Internet to solve energy because of the gift of collaborative intelligence that it's given us. We now have this huge collaborative intelligence tool we never had before, and we can now apply all these brains--6 billion brains-- to the solving of energy. We can use the Internet as a substitution for transportation and carry our bits around instead of our atoms and save all that energy there. That would help. We can--and I'll return to this point later--use the actual architecture of the Internet as a model for the architecture of how we're going to solve energy. Another way is we can use the innovation infrastructure-- that is, the network of people and institutions that have been used to build the Internet. We can use those same institutions to solve energy. So I've entered this history project: Lessons from 63 years of Internet history about meeting world needs for cheap and clean energy. So I'm looking for lessons, and this project is ongoing. I'm mining the Internet history for lessons. You'll see some of them today.

This analogy between the Internet and energy is not a new one--it's an old one. And so I'm just the latest person to try to analogize energy and information. By the way, I'm going to go through these slides very quickly, and you can have them later. You could have had them in advance. They're all in a PowerPoint with all of my speaking notes. Did you know in Keynote you can put notes with your slides? I just discovered that recently. [laughter]

Is there a question or a comment? Please, someone. It would be good to have a comment or a question at this time.

Yes. >>[audience member] Why is the Internet the right model? It may not be, but it's the only model I have. That is, I'm a hammer and everything looks like a nail. So since I'm an Internet tycoon, that's my basis of knowledge, so I'm analogizing. But it may be a broken analogy, and that point has come up many times. My favorite version of that is that some energy hacker will say, "Well, Bob, energy has to obey the laws of thermodynamics." And my answer to that is so did the Internet. There's nothing new there.

Anyway, history. Maybe you've heard the expression "Those who don't study history" "are doomed to repeat it." Have you all heard that? Yep. The problem with that is the Internet history went so well, the temptation is to ignore it and then just be doomed to repeat it. But there seemed to be something wrong with that, so I went and I found the guy who said it: George Santayana. He said the original paraphrased version of this, but he also said something else about history: "History is a pack of lies about events that never happened" "told by people who weren't there." And my only defense is, in the history of the Internet I was there, so that should be an improvement. While googling for other things George Santayana said, I mistyped and I found [laughter] another guy who had something wise to say about the Internet. My wife's a historian, and she said, "Well, you've got to have an era." "If you're going to study the history of the Internet, " "you've got to know when the Internet era began and ended." And most people around here--and just looking out at you-- believe that the Internet probably started with the founding of Google in 1998. [laughter] That's a respectable point of view, but some of us think it happened before that. So you can go back to 1991, which is the year in which Al Gore-- and I'm not going to say he invented the Internet because he's denied that repeatedly. He never said that he invented the Internet. What he actually said is he initiated the creation of the Internet, which is completely different from that. [laughter] Ironically, that law was signed by George Bush in 1991. But that seems too recent also. You can go back to the invention of the World Wide Web, you could go back to packet switching, or maybe '57--the launching of Sputnik, which scared the hell out of America. And so we invested a lot of money in defense R&D, which, in the fullness of time, led to the Internet. But I like going back to 1946 because, arguably, the transistor is the invention that really fueled the word Internet, broadly construed. And what I also like about '46 is that was the year I was born. So the transistor, the Internet, and I would all, then, be born in the same year, and I really like the sound of that.

Whether it's 63 years or 60 years or 50 years or 40 years, I claim, by analogy, the solution of energy is not a this year thing or a next year thing. It's going to take decades to solve energy. That is, if the Internet is any guide, the proper mindset is that it's going to take decades. A consequence of that is, for example, that there's time for science, so we better be doing some science about energy. The energy problem is not shovel ready. We do not have the solutions to energy at hand, so we can't just run off and put billions of dollars into shovel ready solutions. That's the wrong mindset. The correct one, if the Internet is any guide-- going back to your objection--if the Internet is any guide, we should be planning to spend decades on solving energy. We will.

One of the consequences of that mindset is hardening of the categories. That is, when you have that long to innovate, the categories--the taxonomies that you have at this point in time-- are very likely to be eroded. The big example of that is in the early days of the Internet, computing and communication were two categories, and computing was done by IBM and regulated by the Antitrust Division of the Department of Justice, and communication was done by AT&T and was regulated by the FCC. Those were separate categories. The FCC held hearings to draw the line between computing and communications because they wanted to be careful to only regulate communications. And about three weeks later, the Internet completely blew that distinction away because the communications system became computers, and the computers became communicating, and there is no way to separate those two words. That was a misleading dichotomy. A more enduring one was voice, video, and data--three categories. Voice was synonymous with telephony, video was synonymous with television, data was an ASR-33 Teletype plugged into an interactive time-sharing system in the '60s. And those were three different worlds. They had different networks, they had different regulatory agencies. And the Internet destroyed that categorization. You probably heard the distinction voice, video, and data, but, for example, on the desktop today, that's voice, video, and data all merged together on the Internet. Voice, video, and data became data. The same is going to happen in energy, and it's already happened in energy.

How many of you are really enthusiastic about corn ethanol? Ooh. That is the first time in two years I've gotten someone to raise their hand in favor of corn ethanol. Thank you very much. Corn ethanol is a terrible idea. [laughter] It's the wrong feed stock, it's the wrong fuel, and the folly in it-- I claim this as an example-- Excuse me. I didn't mean to embarrass you, but I did. Yes, I did. I claim it's a categorization error because the well-intended people who ran to Washington to get subsidies that would make the production of renewable fuels with corn ethanol forgot that feed, food, and fuel are not actual hard categories.

Yes. >>[audience member] I'd like to relate to that for a second because you're making a point that I think is actually reflecting sort of an impending crisis of finance because if you would have looked three years ago, the common sense and every smart person around in the newspapers, outpass, people in VCs would have invested in corn ethanol. You saw so many articles discussing it, etc. So we're in an age where a paradigm rises so quickly and then it just declines so quickly. So, for instance, if you're a company that wanted to devise a smarter way to harness ethanol for energy, two years ago you could've gotten funding. But today people saw what happened with ethanol and they're telling you, "Two years from now" "someone is going to come up with a better idea from yours," "and by the time you get to the market in three to five years," "I'm just not going to fund that" "because in three to five years there's going to be something smarter," "and they're going to figure out what's wrong with your idea."

I'd like to ask you to stop your question >>[audience member] Okay. because you're giving half my talk. [audience member laughs] I'm going to return to this point. Corn ethanol did seem like a good idea to a lot of people because, for example, the United States Congress passed laws. That means there's at least 51 senators and 200 and some odd congressmen who thought this was a good idea or got cajoled into thinking it was a good idea. There are sort of two upsides here. One is that maybe it's transitional technology, and it's paving the way for what's to come later. You've just made the opposite argument-- that it actually poisoned the well. The other advantage is the money is now being wasted in the Middle West instead of the Middle East, which is progress. [laughter]

Moving along, the Internet was--may still be--a movement, a social phenomenon. The technical term is movement. Vint Cerf, who addressed y'all a few days ago-- I don't know if he actually said this this time when he was here, but every other time he walks around saying, "IP on everything." [laughter] Internet Protocol on everything. And the Internet was a movement; in fact, it was several times a movement, and you can remember that it was a movement. For example, there was a time before the bursting of the Internet Bubble when the seven most feared words on the Internet were, "You just don't get it, do you?" [laughter] So if you raised an objection to some stupid plan like pets.com, the little Silicon Valley smart asses would turn on you and say, "You just don't get it, do you?" [laughter] So that was a movement, and it was inquisitorial and repressive and obnoxious and arrogant at times, and I was right in the middle of it, so I'm not shedding blame.

Energy is a movement. In fact, I sat through most of today, and I could hear it. I could hear how much of a movement it is. Movements--and I'm going to really annoy you on this point-- have slogans, like Vint Cerf's famous slogan, which I happen to like a lot. But they also have colors. For example, this is an early Ethernet diagram where the coax cable was yellow, so yellow became Ethernet's color. We had all the coaxes in the-- They've changed colors now, but they used to all be yellow. That was our color, our movement. The Ethernet submovement of the Internet movement had a color. Energy. Has the energy movement adopted some sort of color? >>[audience member] Green. Green? Would you say there's consensus on that, that the energy movement's color is green? >>[audience member] Red. I think green is the new red, actually. [laughter] It was Tom Friedman in the "New York Times" who first raised this question. He said, "Is green really a good color?" "It's got all this baggage." Green is environmentalism, sure, but it's also against all the ideas we need to solve energy: trade, capitalism, America, technology, nuclear, hydro. Green is a terrible color if you consider all the baggage that it carries with it. So I, as a hobby, have decided to find another color for the energy movement. So I considered black. [laughter] I noticed that oil, coal, silicon, and carbon were all black, and then I realized silicon actually isn't black. Silicon is shiny. It's like a mirror because it rejects light. It's a mirror. But black is depressing. So white. There is white. We could use more white because if we could raise the albedo of Earth, we wouldn't have a global warming problem, so white is a good candidate. But white isn't really a color. So I then looked at photosynthesis and the absorption spectrum, so this makes it a scientific talk. And we looked at the absorption of light as a function of wavelength, and you'll notice that plants absorb light in the blue and the red. Even plants don't like green. [laughter] So that leaves blue and red, and there's a NASA photograph, which kind of gives you a hint as to where I'm going on this: The Blue Marble. And then I became a submarine captain and saw a lot of blue. So I think blue is probably the color that I'm suggesting for energy. I know. Let's let the environmentalists have green, and then those of us who are trying to solve energy will take blue. [laughter] Now, implicit in that statement is some sort of separation between energy and the environment. That's annoying, isn't it? Every university in America has a department now called Energy and the Environment or Environment and Energy, and they're like the same thing. Well, I'm arguing that this is a false categorization. They're not the same thing. There is this little Co2 link between them and much is being made of that, but I claim these are separate problems-- that if we solve global--I'm sorry. We call it climate change now. If we solve climate change, we still have an energy problem. If we solved energy, we would still have a climate change problem. So I think I would argue they are separate. Let's leave green for the environment and blue for energy. Are you annoyed yet? [laughter]

Speaking of colors, here's a color. Jose had this slide up earlier today. You've seen this slide. This is several pictures of Earth at night. The people showing you this slide, especially if you're American, want you to feel bad, and they like to point at all this wasted energy and how these people really know how to live because they don't waste a lot of energy. [laughter] So let's go back to the beginning of the Internet. In the early days of the Internet we had a thing called the Information Explosion that was occurring. This is in sort of the '60s. And we were stuck with a huge amount of infrastructure that had been built over a hundred years by AT&T, so our first impulse was to go for the low hanging fruit and begin to learn how to compress information so we could send it. All this information--compress it, conserve it, get efficient about it so that we could squeeze it through the pitiful infrastructure that we had left over from the telephone network. That is, the first impulse in the early days of the Internet was efficiency and conservation. Does that sound familiar? And so here's the big question: We've just built the Internet. Do we use less bandwidth now than we did before we built the Internet? Like, half? No. Do we use 50% more bandwidth than before the Internet? Do we use double? Do we use ten times? Who knows what the number is? It's probably a million times more. So, going out on the limb again, if the Internet is any guide, when we solve energy, we're going to be using more of it, not less. In fact, the term I like is we will have a squanderable abundance of energy when we're done. If you weren't annoyed until now, you should really be annoyed because the entire mindset, I find, of the green energy movement is that somehow we're going to use less energy. But I claim, by analogy with the Internet, that when we're done, we're going to be using more--a lot more. We will have a squanderable abundance. Now, does this mean I think efficiency is a bad thing? No.

This is my new car. I traded a 12-cylinder Mercedes for this 3-cylinder Mercedes. That's not my wife. My wife is better looking than that. [laughter] But that is exactly my car. It's my midlife crisis car. It's a red convertible. [laughter] I am already dissatisfied with the mileage. That car gets about 38 miles per gallon, and it's not enough for me, so I'm looking for the next car to get. Of course, the car I'm interested in getting might look like one of these. This is a bit of a digression, but the Internet is going to help us do robotransport on demand. Since I'm in the Valley, you know how we complain about the user interface of Windows? It's terrible. Or UNIX--terrible. Here's a user interface for you. You put your hands on it like this and you get up to about 60 miles an hour, and then if you just go like this, you kill your entire family. [laughter] That is a user interface that needs attention. [laughter] [applause] I predict that in my lifetime--and mind you, I'm only at my midlife crisis, so that gives me-- In the next 63 years we will no longer have driver's licenses. We will no longer have to go to the DMV over there on California Street because we will have robotransport.

Over the decades, another thing that's happened-- and this is akin to hardening of the categories-- is that the conception of the Internet has changed. When I first joined the Internet, the dominant mode of computing was punch card computing through mainframe batch processing, which I refer to as cellulosic computing. [laughter] But then quickly I became a technology crusader for a thing called time-sharing, and the dominant mode became build ever larger computers and then share them among ever larger numbers of people through dumb terminals, and then that got blown away and the personal computer came along, and we could do stuff like that. And now we're getting disrupted again by these little computers. So over that time our conception of what we were building has changed dramatically, and that's only in 30 or 40 years. So we can expect a lot more of that in energy. We can expect surprises, and this is a particularly embarrassing surprise to me. In the original ARPANET we had to report to our sponsor where the traffic was going as we learned to use those packet switches. This is the guy who actually built the Internet--Frank Hart from Bolt, Beranek, and Newman. This is the first packet switch called the IMP, and it had four ports on it. Just really overkill--they thought they'd put four ports on the packet switch so that no matter how many computers you had, this could handle it, up to four. And then we had to report to ARPA the traffic that was going through this packet switch. And there was this embarrassing thing that happened every month. The amount of traffic going from MIT to Harvard was ... The amount of traffic going from MIT to Stanford ... You go down the list--I was at MIT. The biggest row in that table was all the traffic going from MIT to MIT. And we called that incestuous traffic, [laughter] which is a pejorative because we weren't building the ARPANET in order to connect the computers in the same building. We were intending to connect them across the United States. So we never reported the incestuous traffic, even though it was the dominant traffic. And only a few years later, Ethernet came--my baby came along, and suddenly, I was in the business of carrying packets 100 yards, and, in fact, most traffic today goes 100 yards through a dozen Ethernets. That was a surprise but not the only surprise. Here's an even bigger surprise. When I joined the Internet project, we thought that we would use wireless for long haul, microwave, and satellite. And then for local distribution we would use--to the last mile--AT&T's copper network. And that was how it was going to go. It was completely backwards. Today long haul is on optical fibers buried in the ground, and short haul is wireless. This is known as the Negroponte Reversal, but it was a big surprise. I'm just making the point that there are going to be surprises in energy. Please do not think that you have a 92% idea of where energy is going to go in the next few decades because things like this are going to happen. These are three surprises worth mentioning because I think we need to be careful not to make these mistakes again when we're solving energy. Since the Internet was built by grad students, we thought everybody else was a grad student so we didn't need security. We thought everything was free, so we forgot to put economics in, and that's still playing out. And the goal of the early Internet was to carry a 7-bit character across the United States in under a half a second. So you can imagine what video has done to all that plumbing. It just wasn't expected. These are the three major bugs of the Internet. And if we're smart, we will learn from history and we will not recreate these three problems again in our solutions to energy. And then there was this surprise: The Internet Bubble, which burst March 10, 2000, as I recall. It took everyone by surprise, except me. I wrote a book predicting the collapse of the Internet. Since the Mortgage Bubble has burst, no one remembers this bubble anymore because it was nothing compared to what just happened. Global warming is-- >>[audience member] What are you charting here? Stock prices. And they went up and up and up and then starting March 10 they went way down and now they're recovering. Sorry. Global warming is a bubble just like this one. How do I know? Al Gore and I inflated the Internet Bubble together, arm in arm. He's back. [laughter] When the Internet Bubble was inflating, was the Internet growing? Yes, of course, it was. Is the world getting warmer? Yes, of course, it is. But is global warming a bubble? Yes, it is. In fact, it may already have burst. I notice in today's rhetoric the word climate change was the actual preferred phrase, so we've gotten off global warming because that's not worked out quite as we had hoped, and now we're into climate change. There's a bubble here, so you should be careful. Once again, I am not a denier. I'm just saying there's a bubble here, and you should be careful. In particular, venture capitalists need to be careful. The point was just being made. You can get all excited about something like corn ethanol and throw a lot of money into it--biofuels in general-- and lose a lot of money because of the bubble aspects. The Internet had many, many bubbles.

Incidentally, I have been recently to the top of Mount Kilimanjaro. These are our tents in the crater there. This is the glacier at the top. It's melting. There's me standing. Do you see this dark area here? That's a puddle, and that glacier is now-- It's only 11,000 years old. The mountain is 100,000 years old. The ice which is melting is only 11,000, and it is equatorial Africa, after all. But I just wanted to include this so I could brag that I had been to the top of Mount Kilimanjaro.

So bubbles are good, though. You shouldn't take my comments to mean that I'm against bubbles. My business is technological innovation, and the status quo is a resourceful and nasty foe, and you need all the tools you need, and bubbles are one of those. Speculative bubbles help innovators break through the status quo, so I'm in favor of bubbles, even though I've just complained about one. Of course, to mix metaphors you need to have a seat when the music stops. One of the things I remember about the Internet, especially in 1970 when I started on it-- joined the effort--I went to a lot of meetings, and at no time in those days were we planning to build the Internet for YouTube. It never came up. Nothing like YouTube ever came up. And yet YouTube and its variations and descendants is most of the Internet today-- the fun parts anyway. So what are going to be energy's YouTubes? Earlier, I posited that we're going to have a squanderable abundance of energy. What might that energy be used for? That is, what will energy's YouTubes be? This is a hobby, and you can join in this. Think of how we would use energy if we had an abundance of it-- a squanderable abundance like we now have a squanderable abundance of computing and communication. What if we had a squanderable abundance of energy? One use of energy is a factor of production in our economy, and Jose was talking--where are you, Jose? You were talking about those models: the Club of Rome type models. Anyway, there's a bunch of models that simulate economic activity. It would be fun to see--I'm sure someone must be doing this-- if you take the cost of energy and double it, what happens to the economy? And if you take the cost of energy, say, over a period of 20 years and you reduce the price of it by a factor of 10, what happens to the economy? I would love to know the answer to that question. Actually, I think we just performed an experiment. We increased the cost of energy over the last few years and now we are in a deep recession. My theory is that that recession was not caused by what they call greed or securitization of mortgage debt. The cause was housing prices stopped going up because the economy started drooping, and the economy was drooping because it was carrying the burden of all these increased energy costs. So I submit that we have performed the experiment of increasing energy costs, and we don't like the result. I was up El Camino today looking at all those buildings that are empty and for lease. My theory is, which I'm working on, is that that was caused by energy. So a good use for energy--abundant squanderable energy-- is just general economic growth, which leads to freedom and prosperity.

Another use would be space travel. My former Xerox colleague, Charles Simonyi, has now been to the International Space Station twice at increasing cost. He spent $35 million last trip, so it's not the sort of thing a lot of people can do right now. But in the year 2070--how about 63 years or so from now-- could we use energy for space travel? So I went to Space Adventures, and I met with the CEO there, Eric Anderson, and I asked him to compute the amount of energy that we would need for space travel in the year 2070. And we did the math, and the assumption was there were 10 billion people and 5% of them are either going to go to the moon and back or just migrate to Mars; it's about the same. He did the math, and it turns out to be 500 exajoules per year, which happens to be approximately the amount of energy the entire human race uses right now every year. So that would be a really big use of energy is to support space travel. I mention space travel with some reluctance because that borders on kooky to talk about going to outer space. But 70 years is a really long time, and 70 years ago no one flew to New York and now we do. So maybe in 70 years that's a good use of that squanderable abundance of energy.

Another one would be water. Water and energy are very closely linked. We need water. A lot of people are dying with dirty water. What if we had enough energy to just boil all the water we needed-- make salt water into fresh water, dirty water into clean water? And I've assumed boiling--it would take 7.6 times the amount of energy we now use total to boil enough water to give everybody the American standard: 400 liters of fresh, clean water every day. 7.6. Geniuses have told me that we don't have to boil the water to clean it and desalinize it. We can use osmosis. But still, that would be an awfully good use of abundant energy. And here's another one. We're so concerned about Co2 in the atmosphere. Why don't we just use the energy to remove the Co2 from the atmosphere? I've run across at least one professor who has a machine that you just put energy in and it just takes atmospheric Co2 at 338 parts per million and just takes it out and removes it using all that. So that would be a use of energy.

This is a fun exercise, and I'm not done. You should play with it. If you had squanderable abundant energy, what would you do with it? I've just given you four possible answers to that question.

One of the things that the Internet had in its history is a number of laws. The most famous, of course, is Moore's Law--Gordon Moore, just down the street here. There was a law before Moore's Law called Grosch's Law. Herb Grosch, an IBM executive, had a lot of data that showed that the cost of a computer grew as the square root of its power, so what's the implication of that? Bigger and bigger computers because the cost only grew as the square root. About three days after that law got promulgated, Gordon Moore came along in 1965. Moore's Law kind of upended that law, and Moore's Law has been very useful. But there are other laws. There's my law, which I'm pleased to show you. Energy needs some laws too. Moore's Law, in particular, has been enormously useful as an agenda setter and a pacer for the development of the Internet and its associated semiconductors. Maybe we need some Enernet laws as agenda setters. Nemet was a Berkeley guy. He didn't do this slide, but I'm calling it Nemet's Law for the moment. This is showing the decreasing cost of solar energy as the years go by. There's a law in here somewhere. Solar energy has been getting cheaper for 30 years. At some point it's going to be cheaper than coal, and when that happens, hallelujah, because then we won't have all the complications with solar. We'll just buy it and end up ahead of the game. So I'm suggesting--and this could be a homework assignment-- that you all try to think of some laws that would help us guide the development of solutions to energy.

Another lesson of the Internet has been what happens when you go to Washington. Have you noticed there is now a strong impulse to go to Washington these days for a lot of-- General Motors has just been mentioned. One of the things the Internet has taught is that when you go to Washington for stuff, you often get the wrong stuff. You get outmaneuvered by the professionals there. My favorite example of this is the Department of Energy. The Department of Energy and the Nuclear Regulatory Commission were both created in the late '70s during the Carter regime. The purpose of their creation was to reduce U.S. dependence on foreign oil. In 1974 and other years there had been the Arab oil crisis in '74, which some of us remember vividly. A consequence of that is someone had the idea of going to Washington and creating a Department of Energy and a Nuclear Regulatory Commission for the purpose of reducing our dependence on foreign oil. So it's been 30 years. The current budget of the Department of Energy is approaching $25 billion a year. How are we doing on our dependence on foreign oil? How's that been going? >>[audience member] Not so good. Not so good. So I submit that as an example of going to Washington for stuff and getting the wrong stuff. One of the wrongest stuffs, I think, was that since that time there has not been another nuclear reactor built in the United States. Thirty years of no nukes. They're building them in France. Eighty percent of France's electricity is nuclear. Ours is only a mere 20%. This is a terrible mistake that's been made. We went to Washington, and we ended up with a Department of Energy and a Nuclear Regulatory Commission who, in the fullness of time, with all the political pressures and everything on it, ended up killing nuclear energy in the United States. Now, the good news is nuclear is coming back. As I'm a venture capitalist, I've looked now at five nuclear start-ups, and one of them some day is going to make it. There are formidable barriers to these start-ups, which we're working to fight. I would like to point out, however, when you take a camera and you stick it into a nuclear reactor, that's the color you get. [laughter]

Yes. >>[audience member] If solar power is going to reach the hallelujah point of cheapness in 2012, why are you investing in nuclear power? Because it would be stupid to invest in only one thing. We learned that in the Internet. There was a time when, for example, semiconductors were all going to go to gallium arsenide because that was going to be the new hot--literally-- semiconductor, and so this major portion of the Silicon Valley started going after gallium arsenide, and what happened? Surprise occurred! And gallium arsenide really hasn't panned out and CMOS won, so the preponderance of our semiconductors are CMOS So if we had had this little debate then, someone would say, "Why are you investing in CMOS?" And someone would say, "In case gallium arsenide doesn't work out." So that would be my blunt answer, which is I don't know whether we're going to have a hallelujah moment in solar. It sure looks like it, and, in fact, my company is investing in that progress. But I think we need to also pursue nuclear. The energy densities of nuclear are phenomenal. We should be after nuclear energy, and we're not because we went to Washington and got the wrong stuff, and we have Jane Fonda, Michael Douglas, and Wilford Brimley and Jack Lemmon to blame for this, if we want. Washington hasn't been entirely useless. It's done some really important things, so there's a lesson here-- lessons from Internet history that we might keep in mind for energy. For example, in the early '80s there was a historic reduction of capital gains taxes-- dramatic--which led to some capital formation in the venture capital community, which led to the backing of almost every company you can name that built the Internet, including my own, Cisco--just make a list of all the important Internet companies. They all came from the venture capital that was produced-- that is, extracted as an investment instrument or type to the reduction of capital gains. So the reduction of capital gains taxes worked. Hmm. Maybe we should do that again and see how that goes.

Demonopolization. Until 1968 you were not allowed to connect anything to the telephone network that was not made by AT&T-- actually, a subsidiary of AT&T called Western Electric. It was illegal to connect things. Answering machines, cell phones, wireless home phone. You couldn't do that unless it was made by Western Electric. Modems, for example. Western Electric was making these very expensive kluge 1,200 baud modems and stuff, and you weren't allowed to improve on them because the argument went-- First of all, the AT&T network was built by ratepayers and therefore, it would be unfair for people to leverage off that investment by connecting their own stuff to it. But the big argument was safety. It would be unsafe to connect these fly by night equipments to the telephone network. In '68 the Federal Communications Commission found in the Carterfone decision that that was probably a bad idea, and so you should be allowed to connect things like modems to the network, making the Internet possible. So demonopolization. The Justice Department of the United States went after IBM for decades for anti-competitive behavior, and that created opportunities for companies like Intel and Microsoft and Cisco to come up. And had there not been antitrust activity, we might still have an all IBM world, which would not be good. We can thank the government for that. The federal government was also a lead customer in the Internet. They bought the first packet switches, and I showed you one earlier. And by buying them, and then the Defense Communications Agency came in and bought a bunch more, and it primed the pump for packet switching and what eventually became the Internet.

Certainly, there's another question by now. [audience member] What would energy deregulation do? Do you have an example? Yes, I just happen to have one. [laughter] There is the Nuclear Regulatory Commission, which is busily hiring 2,000 people now to consider designs of nuclear reactors. [audience member] So all nuclear reactors have to be approved? All commercial reactors in the United States have to be approved. By the way, there's been considerable progress at the NRC, and I don't want to discourage that. But in order to build a reactor that operates in the U.S. for commercial purposes, you need your design to be certified, and there are these five start-ups, so I looked at one. The best one I found was called New Scale in Corvallis, Oregon. They want to raise $100 million, and venture capitalists of my ilk find it hard to produce $100 million in any one project. But then you look inside and you see 50 of the 100 is to do a design over the next four years, and then another $50 million is to be paid to the Nuclear Regulatory Commission to review that design and to certify it. That makes it not venture backable proposition. By the way, there are some venture capitalists looking at this, and some have even invested, but as a general matter that kind of capital requirement is killer for your standard Silicon Valley investor. By the way, there's another factor. The Nuclear Regulatory Commission hasn't successfully approved a reactor that got built in the United States for 30 years. The FDA is in the habit of approving drugs. The NRC is in the habit of not approving new reactor designs. So you'd be stupid to invest in that kind of proposition. There's an agency that needs some work, some regulatory reform. But it goes way beyond that. There are 50 PUCs in the United States--51, probably. Those Public Utilities Commissions were very difficult. They were an obstacle to the development of the Internet. We saw that. The remedy was to make the FCC stronger, not weaker, because then it could get things done nationwide instead of having every little modem maker chewed to pieces by Public Utility Commissions all over the U.S., wringing their hands over safety-- safety always comes up-- safety and ratepayer fairness and so on. And so the Internet would have been killed if the FCC hadn't accumulated a little power that it could use. The PUCs are standing, of course, in the way of the installation of new transmission lines for the so-called Smart Grid. It's just horrible, all the permits upon permits. There's regulatory reform that needs to be done there.

What was the most important thing that the federal government did for the Internet? Research. That is the best spent tax dollars that I saw in my many years of watching the Internet grow were monies spent by the feds on research that led to technologies that turned useful in the Internet.

So we should, as I mentioned earlier, have some research in energy. Where should that research be done? That question arises. I'm going to annoy the rest of you now, those of you who were not yet annoyed. Should private companies do research? The evidence is that the only companies that can afford to do research are monopolies. So AT&T with Bell Labs based on the telephone monopoly, and then there would be Watson Labs based on IBM's computer monopoly, then it would be Xerox PARC based on the copier monopoly, and now there's Microsoft Research based on the desktop software monopoly, and Google Research is building up, based on some monopoly I don't understand quite yet. So I don't think monopolies are the best way to do research, and there are two reasons. One is, of course, they get the money by overcharging their customers, so that's not good. But that's not the killer reason. The killer reason is monopolies are part of the aforementioned status quo. These are the least motivated companies in the world to promote innovation. So they have their labs, which they can afford, they do their research, and then they just don't bring it to market effectively because they're clueless about doing it and they're not motivated. I would rule out relying on monopolies to do our research. So how about government laboratories? Well, in my experience with the Internet, for example, the DOE labs are places where research gets done, but the marginal return of investment in government laboratories seems to be small and declining. They seem to be geographic jobs programs. They all have 2 senators attached and 14 congressmen, and the whole story is about job preservation. Effectiveness of research ... So where should research be done then, Bob, if you're negative on monopolies and government laboratories? The United States and other countries have research universities. This is where this research should be done. MIT is my favorite one, but I understand there are some down in the Valley here that are pretty good too and up the Valley, come to think of it. Berkeley is that way, and Stanford is that way. Boston area has 10 majorish research universities. There are probably 100 or 200 in the United States. These are jewels in the crown. This is where research should be done, and why should it be done there? Looking at the history of the Internet, it shouldn't be done because research universities are well-managed. [laughter] They are not well-managed, and I'm constantly fighting with MIT about all the stupidities in that management system. But it's common among the research-- No, no. It's not because of the ace management. It's because universities graduate students. That's why the research needs to be done there because the best vehicle for innovation is not patents, it's people. And so the graduating students of these research universities are the best vehicles for getting these new innovations into the market. That's why research universities are where the money should go. So the DOE, which I'm vilifying today--and I'm sure I'll regret it-- when new money goes into energy research, most of that money at the DOE gets captured and spent in the DOE national labs, and very little of it goes through DOE into the research universities, and that's wrong and we should fix that.

What we'll get for it is silver bullets. The history of the Internet indicates there are silver bullets. Very often when you have an energy speaker--and I've been through this talk 50 times-- the wizened, gray haired, former physicist who's involved in the energy will tell you in energy there are no silver bullets. This is defeatist behavior. We need to shove those people aside because we need silver bullets. We want to solve energy, and if the Internet is any guide, we'll do it with silver bullets. I recently, two years ago, trekked to San Diego to attend a conference I had never been to before called the Optical Fiber Communications Conference. I went with only one purpose: to thank those people for a silver bullet that they gave us. The silver bullet emerged in the mid '90s. It was called dense wave division multiplexing. It's how all our bandwidth is carried long haul. When I went off to college, my mom would say, "When you get to Boston, call me so I know you're safe." "But just let it ring three times and then hang up." [laughter] She told me this every time I left for school. Thanks to dense wave division multiplexing, we don't think that way anymore. We're making phone calls all the time, and that is a silver bullet. There were some other silver bullets. The IBM punch card was a silver bullet, the transistor was a silver bullet, TCP/IP and Ethernet, invented in 1973, were silver bullets, the World Wide Web in 1989 was a silver bullet. So we need and we should expect--we should demand silver bullets. We shouldn't take this garbage about "there are no silver bullets." Of course there are silver bullets, and we should demand it from our researchers and our engineers and so on. If the Internet is any guide, this is how we're going to bring the fruits of that research and those silver bullets to market.

We're going to do it in this model. The model goes like this: research professor, graduating students, typically postdocs but not always, as Google evidences, some scaling entrepreneurs, of which there's a huge abundance in Silicon Valley, and venture capitalists. And that seems to be, given the experience in the Internet, if the Internet is any guide, the winning innovation model. You get the research university producing new knowledge, you get the professor who you don't want to remove from the university-- you want to leave him to do something new, but his or her graduating students then.

These are the Stanford people. This is the scaling entrepreneur, this is the venture capitalist who made Cisco. But there are lots and lots of examples.

This is an East Coast example. This is Akamai. That's the professor, Tom Leighton, here's the two graduating students. This young man was killed on 9/11 by Jihadists, I might add. And this is my partner, George Conrades, who ran IBM. He went to school at IBM, he learned how to run IBM USA, and then he became the scaling entrepreneur that teamed up with this professor and his grad students to make Akamai into a company.

There are a hundred other examples of this. This, I think, is the innovation model that we should pursue in energy as well.

Question, comment? Yes. [audience member] At the moment, there's a lack of IPOs. I understand it has to do with the number of regulatory changes. What's going to happen in the future to the VC industry as a result? You began by saying at the moment, so there's a partial answer there. IPOs have historically come and gone. Our partnership had an IPO only last week--a company called LogMeIn. It was such a classy cash generator that it just went public very successfully. So IPOs are going to come back. But you should know that in the venture--if you care-- in the venture community, there's a lot of talk about how we need a new model for venture capital. But I think most of that talk is coming from the 500 venture capital firms that are not going to exist anymore when they run out of their current funds. There seem to be too many venture capitalists-- say, about 1,000, and about 500 of those are going down the tubes. The trouble is, in venture capital the tubes are really long. They're, like, 10 years. So it'll be a while. And those people, who are all going out of business, are the ones who say, "We need a new model of venture capital." Of course, that's a model in which they get money again. [laughter] We'll see, but I don't think venture capital needs to evolve very much. Now, in energy the current model of energy is not working for venture capitalists. That is, the capital intensivity of energy projects is noticeably large, but the same could be said of drug discovery. It costs half a billion dollars or a billion dollars to develop a new drug, and yet venture capitalists have found fertile ground for participating in that innovation realm--that is, drug discovery. How has that occurred? It is capital intensive like energy, but there's an ecology among the big players. The big pharma drug companies know how to partner with biotech, how to partner with, put cash into, and eventually buy biotech companies. So there is a whole innovation ecology in which venture capitalists have found a role to play in drug discovery.

This is from a National Academy report, and I only show it because it's complicated-- big and complicated-- much bigger than the Internet. The Internet is only hundreds of millions and billions. Energy is a serious number of trillions of dollars. It's a big, complicated system. What does the Internet teach us about how to deal with this complexity? One of the early lessons was the process of innovation. That is, the Internet culture brought with it a culture of innovation. So this is Jon Postel, who died early, but he is emblematic of that process involving the IETF, for example, and the RFCs, and it all took place within five miles of here--a lot of it. It is this process for experimentation and prototyping and competition and the introduction of architecture. So this is a culture that might be replicated in energy. We would be advised to do that. In the architecture of the Internet there was this layering model. Are you all familiar with the layering model? This is called the seven layers of the ISO-- Do they still teach that? If they don't, they should because this is an architectural idea that really worked. Basically, it said we have this complicated problem which is networking. Let's divide it into these levels--the physical down here and so on. The way this worked is by dividing the world this way I, for example, could live a rich, full life at Level 1 and 2, while other people like Tim Berners-Lee were having fun up here at Level 7. And we could live specialized and effective in our own worlds. Not only that, technology could evolve at each of these levels independently. It wasn't lockstep by the least common denominator. So we could be making things faster here while these guys were making standards and developing new applications up here. And, of course, to make this architecture work we had to design interfaces, and those interfaces were an exercise in the art of design, generality, division of function.

The most dramatic benefit of this design of these layers was serendipity. Of course, when you build something general purpose, then there are purposes that you're not anticipating. For example, this stuff got in place on the Internet circa 1973. The World Wide Web came along in 1989, and that worked. And then Google didn't come along until 1998. And all that stuff is still working, and none of it was anticipated. There's a lot of serendipity, as there will be in energy. This architecture will also be, by analogy, distributed. We went from the big mainframes with cellulosic computing down to microprocessors. Let's go from the big centralized power plants to distributed energy: Distributed solar, distributed wind, distributed nuclear, distributed everything. If the Internet is any guide, we're going to pursue distributed. There are sort of two benefits there. One is the intrinsic benefits of distribution technologically, but then there's also the go to market benefits of distribution. Ethernet, for example. IBM was the entrenched monopoly in computing. Can you imagine--I don't have to imagine--I tried it-- taking Ethernet and bringing it to the people who buy IBM products and saying, "You should buy this instead." Nonstarter. The way Ethernet got started was to be distributed. We sold to people who didn't buy IBM products, people who had personal computers, for example. And the IBM salesman wasn't selling, and IBM wasn't competing at that level, so we came in through work groups and personal computers through distribution. We were distributed. The energy industry uses the word distribution. They say, "Let's build a big coal plant and then distribute the energy out." That's not what I'm talking about. I'm talking about distributed generation of energy and its networking. I've begun to hear this familiar thing now in the Smart Grid. Have you all heard of the Smart Grid? It's a buzzword now. Everybody's doing Smart Grid now. Whatever you were doing before, it's now called the Smart Grid. [laughter] There's this thing I'm hearing now from utilities: "We think we need to build a Smart Grid and it needs to get to every home." "But that Internet thing is insecure and unreliable," "and it doesn't go to enough places." "So we have to build an entirely new network." Uh-uh (negative). I've heard that before. That's an Internet lesson.

If there's something wrong with the Internet for applications in energy, we should fix the Internet instead of starting over with a whole other network. So that's a really easy lesson to learn. Of course, now there's the nature of the network, the nature of the Smart Grid. If the Internet is any guide, of course it'll be distributed. But it's also likely to be meshed because we're not just distributing from a central point. We're not distributing energy. We're going to be exchanging energy. We have these uncertainties of demand, uncertainties of supply thanks to renewables, and how is all this uncertainty going to be handled with a meshed network for the exchange of energy, not merely its distribution? It will be much more symmetrical. The energy won't be going in one direction. It'll be going in many directions.

Another lesson from the Internet had to do with synchronization. The old telecom network prior to the Internet ran a central global clock. You wanted to put a telephone conversation, you had to get in lockstep with the TDM hierarchy and get synchronized. Does that sound familiar? Today's grid, which I guess we'd call the Dumb Grid, is synchronized. If you want to put energy onto the grid, you have to come up to speed, you have to establish frequency lock and even phase lock to get into the grid. If the Internet is any guide, we're going to desynchronize the grid.

Another thing we did is the telecommunication network prior to the Internet had no storage. The bits entered the edge and went flying through and came out the other side because they were in a rush. They were carrying telephone conversations. We're going to have to storify the grid just like we storified the Internet. That was said today. Maybe Jose said it. Jose said almost everything that I'm saying already. [laughter] But adding storage-- [Jose interrupts] What's that? [Jose's comments inaudible] [audience laughter] Oh, is that how you did it? [laughter] Many kinds of storage: utility scale storage. Just like in the Internet, there's storage in the servers, there's storage in the switches, there's storage in the local servers in the building, there's storage in your cell phone, there's storage everywhere. So just like we have different kinds of storage for the Internet: mag tape, core memories, discs, there is storage in the grid. When there's excess electricity, they pump water from here to here and then when they need the electricity, they run it back through the turbine. This is a really bad idea. So we're now developing new batteries and even hydrogen. There are many kinds of storage. We're investing in compressed air storage of energy, in tiny, little lithium ion batteries for the edges of the Smart Grid, and water splitting for storing solar and wind energy as hydrogen to recreate into electricity later. So if the Internet is any guide, we will solve energy by storifying-- adding storage to the grid.

And then another lesson has to do with LANs. For a while there in the early days of the ARPANET, networking was a wide area networking activity, as is the grid. When we're finished with the Smart Grid, it's going to have to have LANs. It's going to have to reach inside of buildings where all the loads are for improving energy efficiency.

I'm about to finish ahead of schedule, I might add, and hopefully, you'll have some questions. Going back to the idea of doing multiple things, a thousand flowers bloom. These are some of the collected thoughts on where I'm thinking that we-- In the long list of things we need to be pursuing over decades in order to solve energy, this would be a list of things we need to pursue. For example, when I tout nuclear, I'm often asked, "Well, would you want to do solar?" And my answer is, "Yeah, let's do both." Solar and nuclear belong--both of them--toward the top--not exclusively, but toward the top of the list. So this is my list, and this is sort of what the top looks like in my list based on lessons from the Internet. You'll notice the word distributed is repeated a lot. Distributed geothermal. A geothermal system in your house, not in Iceland, and so on.

Any questions about this? Yes, sir. [audience member] You've raised questions of the new VC model requirement. In order to invest you said that there's a convergence. Investing in energy is becoming like investing in drugs. What can you say about the incentives today in the market for that, and perhaps there is another law here that if the state makes more money than the private companies dealing with oil, it's not going to work. America as a government is making more money from tax money. Actually, the data is online and I can share it with you. It is making more money than the energy companies through taxation. So there's actually almost zero incentive for the government to encourage any other type of renewable as long as they're making more money from the current energy market. The number is, I think, a trillion and a half dollars over the last 20 years in government earning. Can you relate a little bit more to that? Where are the incentives, what can the government do, and does that even indicate anything-- the ratio between government earnings and private sector earnings on oil energy. You may have already detected, but if not, I'm going to reinforce the idea that I am not an expert on anything relating to the government. Here's a debate. Because Co2 is such a big problem and the world will end shortly unless we diminish Co2--I'm not exaggerating. I see presentations. The world is going to **** end unless we solve Co2 tomorrow. There's new urgency, by the way. Co2 used to be urgent because the world is going to end because oh, my God, Kilimanjaro is melting and nonlinear effects, positive feedback loops-- It's all going to go boom and you get a Nobel Prize for that. There's a new source of urgency in energy now, which is the whole economy has gone into the tank, so we have to stimulate it, so let's hurry up and don't think too much about where the money is going because we're in a rush now to solve energy. So I think there should be a debate about whether it's wise to increase the cost of energy to solve the Co2 problem. Right? We need to raise the cost of energy to capture the externalities, which all boils down to tax energy and raise its price because we want to discourage energy to cut down on all that Co2. But while we're doing that, we are also diminishing economic activity, so there's a big debate there, and I am the least qualified person to sort it out for you. But I'm glad you noticed.

That's probably a good way to end this session because it is 8:30, and end it with a question that I can't answer. [laughter] So thanks for your attention. [applause] [♪classical music♪] www.singularityu.org

Video Details

Duration: 1 hour, 14 minutes and 24 seconds
Country: United States
Language: English
Views: 251
Posted by: davidorban on Sep 30, 2009

At GSP-09, Bob Metcalfe talks about the Enernet

Caption and Translate

    Sign In/Register for Dotsub to translate this video.