Love with eniac. He knew what it could do. Then it was this very wonderful thing where he taught clarie how to program it. There was a debate of who so what first, but somebody saw that they were designing the institute machine and writing code for it. But the machine was not ready. Someone said, we can rewire eniac and run programs on the eniac. They spent six weeks rewiring it and getting it to run. That was the origin of modern software. And that was such a fruitful time. So it was a mixture of their marriage, the coding. And she had no training whatsoever . She did have training. During the war, when johnny went off to england to do what, we do not know . She was left behind in princeton. Women could apply and get a wartime job. Hers was at the center for Population Research at princeton university. Her job was modeling populations of people. What would happen if you created a new state in the middle east or something . They model growth of population. That is what you needed to solve these early bomb problems. It was all about studying the population of neutrons. Escape is like immigrating. She taught herself the mathematics of population. That ended up being what they needed at los alamos. There is a point of controversy in history about von neumanns relationship to the intellectual work that eckhart and mo9ckley were doing. And whether von neumanns wondered to circulate that was sort of an opensource guy because he understood that the power should truly be secret. Or whether he played by the rules, so to speak. He certainly broke the rules. This is very controversial. There is no doubt that von neumann did not write the whole paper. He put them together. There is some debate what parts of the paper goldstein wrote and what part von neumann wrote. There is no question a lot of the ideas came from eckart. They should have been coauthors in the modern standard. The paper was released under john von neumanns name. I try to find out what the truth is, and the truth is that it was released and considered a publication. So it avoided the chance of patents. All i can tell you in terms of a smoking gun is that in 1945 von neumann signed a consulting agreement with ibm. So this result was highly favorable for ibm. That there were no patent restrictions on these ideas. From von neumanns1 point of view, he just thought this would be for the good of everybody. I think you have to remember they are coming out of world war ii, all these groups, they were all collaborating. They were all collaborating together. People from manchester coming to princeton, people from princeton going to manchester. Sort of arguing about who should get credit came later. They all come out of world war ii. The war ends, von neumann wants to transplant the entire eniac team and continue the work. And eckhart and monthly decline. Originally, eckhart was supposed to be bigelow. People went back and forth. There is a great letter from johnny saying they shanghaied two of our guys. Always stealing each others engineers. So in the end, it became very bitter. Again, you can read some documents in this book that are pretty incriminating that Accor Markley was doing quite well. They had a contract for three machines with the government that would put them firmly on the path that ibm took. And they have these security clearance question, and they lost the contract. There is a sort of disturbing memo describing what had happened. I do not think they were a security risk, but it put the company off the path forward. Von neumann was bitter about the decision, critical of this . He was annoyed. He wanted to go full speed ahead and felt they were Holding Things up. And they would feel differently. I try not to take sides, but i think there is truth on both sides. In 1946, he put that aside goes back to ias and begins working on this highly improved binary Storage Program successor to the eniac called the maniac. The maniac. That name was adopted by los alamos. The era of the acs. Was very practical. He said, we are not going to originate anything. We are going to work with the stateoftheart as it existed at that moment. And that meant vacuum tubes and crt memory as it existed at the time and 15 tons of air conditioning. Yes, magnetic wire, inputoutput. They made a lot of things. At that time, you could not modify ibm equipment. It was like the old telephones. They did modify ibm they got some punch card machines that still read 12 bits. And they converted the machines to read on the 80bit side. And that is the whole world we have. That is why we have 80 character lines. Is that innovation . That was a guy named hewitt crane, who moved to stanford research. He just died a few years ago. He did that alone and got in real trouble. Then ibm people said, we can probably sell this. They did very well. And put them in data processing. Francis bufford had a great review of the book in the guardian of london today. He says no other book brings to light anything so vividly or appreciatively like the immense engineering difficulty of creating electronic logic for the first time. Talk about the design, von neumanns drive to do this. He was tireless in bringing these theories to life. He was trying to do the almost impossible. What they did was really crazy. They are making a 40 bit parallel machine, where one bit is in a different to. Tubes. And if you walk past wearing a sweater, you might throw 300 bits off of them and nothing works. But they got it to work in new jersey, the least sustainable environment for delicate electronics. Any time a car went out the memory. Thunderstorms, they were so persistent. The fact they got this thing working is quite amazing. You used a phrase, a deal with the devil, talking about los alamos and the atomic bomb. There was a another deal with the devil made in the development of the computer that von neumann wanted to build. I feel, and i think this is still a fable for the future, the deal was made that the devil could have this weapon that could destroy all life on earth. And the scientists would get this computer that would reveal all knowledge. It was this incredible trade. Give me everlasting life, and i will give you my firstborn child sort of thing. And we think that we kind of won the deal. It is income rental today how real the threat of Worldwide Global nuclear war was at that time in the 1950s. There was a 20 minute launch window to destroy the world. We survived that. We do not worry about those conflict like we used to. I think what you have to remember is that computers could be equally threatening. So maybe the devil is out there. I do not want the bomb, i want the computers. That is what we need to be watchful for, that we do not let this Global Computing Network that is so beautiful and is a cathedral, make sure it does not become the tool of some totalitarian maniac. I want to get to that in a second. It was interesting incredibly compelling, that von neumann could see both moving and parallel. Having worked at los alamos, he could see what the net result of the Hydrogen Bomb was likely to be. And at the same time, had this premonition of what computing taken to its far extent, could turn out to be. He did not perceive the internet. We give von neumann credit for everything, but not the internet. Didnt they devise what you call a real problem and perfect cover . They are working on thermonuclear explosion models but they are also working on strategic weather. Meteorology. Von neumann was just an opportunist. A genius. He did sort of need a cover for this sort of work. Meteorology was the perfect work. They brought in real meteorologists. Anytime you check your iphone and get a five day forecast, it is the same code they developed. Just better input data and infinitely more processing power. Two applications that seem to catch on. Weapons and whether. Both hydrodynamics, is really what it is. You talked about the code for the eniac, and this is a great revelation of the book, which is that they are developing the monte carlo algorithm. Von neumann is instructing clarie coaching her, it getting involved in writing code for monte carlo. Had that been known that she was writing code . Some of these codes there is one in an envelope. That you can mail with stamps. Is what we call the source code for the Hydrogen Bomb. This code would run today. It would run on the eniac for six weeks to get a yes or no answer. Now, in microseconds, your screen is refreshed. You use up a zillion bits. Very different kinds of code but incredibly important. Monte carlo, it is the perfect example. Ulam invented it while he was recovering from a brain virus. He started playing solitaire and realized, we could do computing this way. [laughter] how did monte carlo, which is an incredibly sophisticated way of doing that, writing software, how did that happen so early in the evolution of computing when machines were so primitive and memory so small . They needed it, is why it happened. They needed to follow populations of neutrons and did not have the horsepower to do it in an analytical way, so they had to do it in a statistical way. The beauty of monte carlo is that it is not an approximation. It is closer to the way the world of physics works. Physics, at its essence is a physical process. Can you explain a little bit about that . There are people that know about monte carlo, but there may be people watching that do not. Instead of trying to get an exact answer you sort of developed a game of chance and run that game of chance. The more you play, the better your answer gets. If you are gambling at a casino, if you gambled a long time, you would get an accurate estimate of how what the take of the other side is. 3 or 2 . What is so beautiful, you could not imagine it was true, but johnny and clarie meet in monte carlo. He has a system for roulette and has lost all his money. [laughter] with his first marriage, he is still married to marinas mother. But he goes over to the bar, and there is clairie, whose husband is a compulsive gambler, and he knows her from childhood. She was an attractive figure skater. She buys him a drink. She had the money. So they meet in monte carlo. You could not make that up. It is a great story absolutely is. Work begins in 1946 and in 1951 the team from los alamos comes to princeton. They load a very large thermonuclear calculation into the maniac. It runs for 24 hours without interruption. For six weeks. 60 days. More like eight weeks. With a flabbergasted . Confident . What was the reaction . Nobody was supposed to talk about it. We are not even supposed to know the maniac was working. They got it working early and ran it. We were desperate to know. That was when we were building the first big Hydrogen Bomb. So there are a few people left. Harris is still alive. Marcia rosenberg died not too long ago. That people were not supposed to talk about what they were doing. They were testing machine and running a real problem. You can tell by the dates at which nick metropolous shows up. [inaudible] you should give her a mike. [inaudible] we will have to get you a mike. Nick and stan came with the los alamos problems. It all came from po box 1663. Because nick had a girlfriend in new mexico. [laughter] she said he was so handsome. And the Hydrogen Bomb is detonated 60 years ago this year, november 1, 1952 in the south pacific. Von neumann writes about knowing, and he uses the phrase creating a monster, but he also said he felt it would be unethical for scientists not to see through to the end what they knew they were capable of. I think it is interesting that he juxtaposed ethics in that way. The monster being created, but the scientific obligation to see it through to the end. Was that characteristic of von neumann . And very character a stick of stan too. Stan was married to france wall, who was again the francoise who was against the bomb. He said, we have to know what happens. If there is a way to know what happens with energy density, it is our job to find out. It is not our job to say if it is good or bad. The next year, the soviets detonate their Hydrogen Bomb. It was not very successful. Like with the germans, they were not as far along as we were afraid of. Von neumann had los alamos. They were working on a Hydrogen Bomb design that was not successful, but we did not know that at the time. Then we find out that one of the researchers was a russians by. Russian spy. Less than four years later von neumann dies of cancer. Tragically. And the team scatters. They pull the plug. The computing team, because the plug gets pulled, it does not last. It essentially collapses. It setback Computer Science almost a decade. They had the script does doing computing for scientific purposes, which ibm picked up on. The Research Center started doing that, but there was a gap in between that was lost. It is understandable why the institute did not want to become a computing center. It was von neumann who kept it going. When he was gone, it was over. We have plenty of questions. We do. There are a lot of really good questions. Lets talk about the implications of all this. You talk a lot about the implications of where computing is and where it is going in the book. You said about a week ago that the last time you checked, the digital universe lets make sure that these numbers right is expanding by 2 trillion transistors per second in boston power and 5 trillion bits per second in storage. That is hard disk storage. Von neumann predicted a universe of 10,000 switches, i think. He said that was all you needed for a computer, 10,000 transistors. 10,000 switching units would be enough. With this unleashing of computing power, there are things you talked about. One is Artificial Intelligence. When von neumann spoke of computers, he never talked about Artificial Intelligence. Alan turing talk about little else. Talk about that dichotomy. The two of them, and where you personally leave Artificial Intelligence is headed. I am on the alan turing side. Turing never publish anything until it was perfectly proved. He spoke in perfect, complete sentences. Very much the other way. Stuttered, and said what he thought. Very different characters. That is the tragedy of von neumanns death. He did not want to bullish anything until he had a complete theory. You never got there. Turing that at 41, von neumann at 53. Is what we are seeing the approximation of Artificial Intelligence as they might have thought of it . I think it is awfully close to what turing was looking at. People remember his 1950 paper the imitation game, and his 1936 paper, the universal machine. The one that is equally important but less remembered is his 1938 phd dissertation at princeton that was on nondeterminate machines that he called oracle machines. Every once in a while, they take a jump like we do, in thinking. We put it together, and that is intelligence. Turing believed he had proved a machine that never makes mistakes could never be intelligent. Gerdle proved that as well. If it never makes mistakes, it is not going to be intelligent. But if you look at what google is doing, this indoor miss deterministic machine, a million servers all perfectly predictable. Yet they are connected by nondeterministic links, which are the people. You are given 10 search results and click on one. That is a nondeterministic process. And the deterministic machine incorporates the state of that nondeterministic leap into the deterministic machine. That is why google gets you result in a millisecond. It knows where other people have found meaning. You cannot imagine a more perfect blueprint for an oracle machine than what google is doing right here. That is not scary or anything else. They are just doing it and we love them. We could not live without it. The second thing i wanted to talk about was computers as an organism. You talk a lot in the book again, it is the juxtaposition of turing and von neumann. You say many decades later, we still face the same question. Turings question was what it will take for machines began to think. Von neumanns question was what it will take for machines to reproduce. The notion of a replicating computer is in your book. You talk about that logically and practically. What do you think the implications are . That is why we ended up with silicon valley. These machines became effectively self replicating. They are replicating themselves. I think that is why this von neumann machine is so important. Even though that were there were other machines, it is bigelows machine, the chip factories that used to be here making millions everyday. Is organism the word . I am more interested in codes than organisms. You have to pick the great character in this book. It is von neumann it is Julian Bigelow. Of araceli came in with the idea that codes can be viewed as organisms, because they self replicate, they crossbreed. And he looked at that in 1953. In a way, more stuff is happening on the coding side. The chips are sort of a soup out of which interesting stuff happens. The third one is Big Computers. Von neumann envisioned a world in which there was no network to speak of. A few Big Computers would perform all of the worlds commutations. Computations. You see that being realized in some respects. Strange how were going back to that. His vision was it would be three or four Big Computers that you would dial in your computation over a network. Then we went to a vast distributed network and going more to things like google and facebook, which is essentially are large computers in a broad sense. In a way and of course people in the industry, it has gone back and forth many times. Lets get to some audience questions. Heres one that says it is not unreasonable to say that theoretical Computer Science is still dominated by turings concept. You think that is possible to change in the near future . Yes. I think the way it will change is not from the bottom up. I do not think we are ever going to escape the turing machine running on the von neumann matrix. Turing had a onedimensional model. I think we are stuck with that. It works so well. At a higher level, i think we are now free to build all sorts of different models. I think the answer is yes, it is going to change. There is a question from someone who went to the ias in 1955 and saw the computer they had had vacuum tubes hanging out. Was that a temporary situation or did it look like that . They might have been looking at i do not know. I would like to know exactly when that was. They had monitors. That could have been