Doug Engelbart's    Colloquium at Stanford

An In-Depth Look at "The Unfinished Revolution"

Thank you, Doug. I want to talk about the value proposition and how I think about Doug's ideas and how we apply some of them at SRI. I am going to make some very simple concepts and I hope  it doesn't offend anyone. I want to do it to make a contrast so that when I can cross some of those  things with Doug's thoughts we will see a difference. I hope that you will bear with me on this. 
 

Einstein was once asked what is the most powerful force in the universe, and he said compound interest. He may or may not have said that but compound interest is still an amazing thing.  This is the central theme and concept behind my talk today. Here is a simple characture of what it  means to build a product. Some one has a tool, and they build a product. That is good, and  historically people have worried about quality at the same time. This is pretty much the mindset.  Get the product out and you are done. 

What Doug is talking about is a knowledge accumulation model. 

You have a product and in some sense it become a tool, Tool 1, that essentially becomes a second product, Product 2. That product can be used as a tool to make a third product. Keep this  process going. You will recognize that as being an equation that says that the next thing that I  produce is equal to what I have now, times fractions 1+k, and if k is bigger than one it basically  derives a difference equation for an exponential. So the goal of all this is to in fact create  exponential performance improvements. Some well known exponentials, Mores Law is a classic version  of this where the power of computers doubles every two months. Medcalf's law is a different kind of  law but it is also an exponential that says the value of a network goes up as the square of the  number of connections. That was not true before we had the Internet back when we had radio and  television; it was a linear model. Every time you added another consumer that value proposition went  up by one. This case when you add another proposition it goes up by n squared. So, it is  fundamentally different from what we had. It is not the radio, it is not television, and it is  something different. These are powerful forces, as Einstein suggested we pay attention to. Now these  ideas have been around for a long time. 

Here is my simple characture of what the job was a long time ago. You pick berries, raise kids, hunt, you pick berries, raise kids, hunt. That is a different kind of recursive process, but  pretty fundamental. Never the less, even though people were doing this they were still on this kind  of exponential improvement curve, even though I say we were on the tail of series part of the curve  that was linear. At least it seemed linear at the time. 

So, you can put Doug into this context. An A level task would be to grow and gather food. B is an ax so that you can do it better, and C developing new material and processes as a way to improve B so  you can improve A. A NIC was the campfire everybody got together. That was the network improvement  community. The Dynamic Knowledge Repository and the Co-Diak was old folks telling stories. It is  hard to improve on that. The most compounding form of communication is through a story. Just that  one example shows us how far away we are scientifically in having the kind of capability that Doug  dreams about. The Meta-NIC was the annual pow-wow where all of the different communities got  together and shared stories about how they can improve the processes about how they improve how they  do there business. There are lots of examples. Language is one that evolved this way. You go from  the stone ax, to the bronze ax, to the iron ax. There is a nice set of metaphors here. What is  different with Doug? These ideas are resolved as mankind, so I tried to write down what I think is  different about Doug and the way he views the world. I have emphasized a few things here. Network  Improvement Community. Network is the key word here. 
 

It is often said that point of view is worth eighty IQ points. The objective of the NIC is to collect as many plus eighty IQ points as you can. That is collective intelligence. We are a lot  smarter as a group or as the magnitude than we are individually. It is recursive, it builds on  itself. That is bootstrapping idea. Doug is recursive in his model in every way that you can  imagine. It really is a very through analysis of how many recursive ways you can build value to make  an improvement community. It is comprehensive. There are many NICs. Quality circles have been around  for a long time, but Doug says that is not enough. Quality is important. It is fundamental. That is  one of the key things we ought to pay attention to. It is not enough. It applies to every aspect of  the job. Innovation, new product development, ethics, and all kinds of issues that intercept what  makes progress possible. It is inclusive. People often talk about computers and machines, but they  forget to talk about people. I am going to argue shortly that the people part of this equation is  extremely fundamental particularly in the world that we are about to enter. Doug's approach is  systematic. A task, B task which improves A, C task which improves B, so that you can improve A.  That kind of thinking is systematic and every time you have a situation which you don't have a B and  a C, a little red flag should go up and say that maybe I have an opportunity to do an better job. It  is intelligent and persistent. I made a joke about old folks telling stories. It shows to you the  value of that. Doug is a great storyteller and we are all here because of his ability to tell  stories. I'd actually say that he is a perfect example of the point that I am trying to make here.  It has to scale. There are many MetaNICs. A MetaNIC is one to improve the NICs, but there are many,  it is not just one kind and we can learn from everyone. 

We all know about some exponential doubling times, computers, memory, fiber optics, the human genome, the web, there are lots of them these days. Eventually, they are going to impact  things that are really matter to us like shelter, food, medicine, entertainment, and education.  Education hasn't really been touched yet, it is an important problem. Entertainment. We are  beginning to see that there is one so called standard, we move on to another level of technology. It  has gotten to be really interesting. Back in the fifties a standard like radio or television would  stick around for fifty years or more before another one would come up. It is not true anymore.  Medicine is a profound one. We are being seen now as basically big strings of software. In treating  us as information. Once we start treating ourselves as information then we go on the equivalent of  More's Law in terms of how we're going to be treated and the way we are going to interact with our  doctors. So obviously this isn't going to last forever. 

Just to make this point on other way, thirty years ago most stuff was on a thirty to fifty doubling time scale. More's Law was down at the bottom of the economy. Even though it was chunking  along at two times every eighteen months, frankly it didn't effect very much. That is different  today, More's Law intercepts just about everything. That is basically the motivation behind the  bubble that we are going through now. Of course, Medcalf's Law came along in the form of the  Internet as another driver to push the activity that is going on. Those are not going to stop. They  are going to affect most of that other stuff too, as I just tried to argue. So most of the things  that are important to us are going to be on some kind of exponential in the area that we are going  into. 

The question comes up that if you want to be on an exponential, what are the key ingredients that you must satisfy to be on it. I think that there are four things. I believe if one  of these is missing, it doesn't work. I believe that there are other things too. Doug is always  arguing the whole family of things. I am going to argue these four. First, it has to be an important  problem. It has to be something where there is no limit. You double the performance, it is like  raising a kid. You teach them something you want to double it and double it. There is basically no  limit to how much you can teach your child. Education. Communication, fiber optic communication I  put this one down right now because for every having of the cost of the fiber optic communication,  the market potential goes up by fifty percent. You can't satisfy it. So, the companies that are  after that market are doing everything that they can to grow as fast as they can. Ideas. In order  for you to say on an exponential there needs to be new ideas. You can't grow unless there are new  ideas. That means concepts, representations, new technologies, invention systems, and business  innovations. Ultimately business innovations have historically been more important than the  technology. The technology opens the door, but it is the business innovations that really makes it  work. Strategies as Doug has mentions. You also need resources. People and dollars. Today dollars  don't necessarily matter. You can raise lots of money, the only thing that is really important is do  you have a keen idea, and do you have the people. You need a recursive process. Build it and  iterate. If you have these things you can be on an exponential.

At SRI, we take these ideas quite seriously, and we apply them to a couple of different activities. We apply these when there is an important problem. If it is a trivial problem, then we  don't think about it we just get on it and then go on to the next thing. If there is a really  important problem, that society needs solved, we will go after that one. That requires a vision and  a strategy to think about those. When you think of Doug's classic work in 1968, he had a very  powerful vision that motivated Jeff and others to join him. To make computers and humans more  intelligent. To create a more productive and creative environment. Champions and teams. I like  passion. I like people with passion because I think that they change the world. Doug is a passionate  champion of his ideas. You need ideas. You need new ideas, and lots of smart people in many  disciplines because most of the problems that are interesting today are multidisciplinary. You need  multiple representations and mathematics but if you don't have at least two representations, you  can't solve a hard problem. Format's Theorem was solved with something like forty to fifty  transformations to be able to look that problem in enough ways to solve it. That is a metaphor that  is useful for all of us in our work today. There are business models, processes, and human  considerations, which I come to. And the recursive process. What this means is if you are putting  together a proposal you put it together quickly, you put it in front of people, collect ideas, you  do another version, you do it again. If you are doing a demonstration, it is not enough to talk  about it. Talking about it doesn't do it, it doesn't capture the idea. You have got to build the  demo, put it in front of people, collect their ideas, and implement them again. You need to do that  really fast with a lot of enthusiasm. If you do those they you get on... there is no such thing as  an exponential for this kind of activity but there is something that looks an awful lot like it. It  is fun when you do this. I want to get back to another thing that Doug emphasizes. Technology may be  a driver for this but the thing that makes it work is people. It is perhaps the most important thing  in the world today, the technology is going to come but you need to think about how you are going to  treat and work with people. People make exponentials work or not work. If you have a company that  has a doubling time of two every six months, and you have a personnel crisis in your company, you  can be out of business. It is very non-linear when these kinds of effects come together and you get  involved in a business that is so dynamic that is doubling every six or twelve months, anything a  financial or personal crisis, it can basically crack the bubble and put you in jeopardy. 

So these funny words that are in this view graph are actually things that we take very seriously. The three commandments. Respect, integrity, generosity is spirit. At SRI we have at least  six NICs that I can count that are up and running. I went to the person who runs one of them the  other day. I said, when you put together a NIC, what is the most important thing? I thought she  might say the application or something. She said Kurt, it is respect and trust. I said I am sure  that is true, but what else comes to mind. She said Kurt, it is respect and trust. I said well, I  understand that, but what else do you think about. She said Kurt, it is respect and trust. How many  times does she have to tell me that if you want to have a network community, and you don't have  respect and trust, nothing can happen. It is really fundamental. 

So these funny words, you know Gilbert Stock and Trade, cynicism, distrust, and misunderstanding it is natural when you are changing an organization. When you are on an exponential where you are  doubling every six to twelve months, the amount of cynicism you can build up is enormous. So we have  to name it, address it, and deal with it. When I am going through all of these things I really think  that these things are important. More important than ever when you buy into this kind of a vision.  The reason I pick staff and it's partners is that it is the only way that you can treat your staff  in this kind of a world, they have to be in partners in this kind of a relationship with them. I  think that is a really big transformation in the way we work. It is a really nice work. There is  nothing on this page that anyone would disagree with. They are good and positive values. They are  values that help people grow and be bigger and better than they are. 

The DNA of change has to do with there has to be a demonstrated need for change. The N stands for a new vision and the A is an action plan. If you don't have a need, a vision, and an action plan,  people don't change. That is not enough. If you ask me if a company has a need, a vision, and a  plan, and I don't fit into it, I am not going to buy into it. So, this personal transformation on  the left hand side says vision A to vision Z. I have to help you see how the skills that you have  can be morphed into this new job. When you get there, not only are you going to be valued, you are  going to be more valued and have more potential and more returns than you have today. That is my job  as a manage, if I can't do that, you are probably not going to buy in. And so it goes. 

FUD: Fear, uncertainty, and doubt, that comes up all the time, you have to address that. The motivation part. Freedom to do your job, you have to respect people. A is for achievement. A lot of  us are scientists in this room. The thing why we get into the business is for achievement. We want  to make a difference. I stand for involvement. If you don't involve us in decisions that are  important to us that can affect our freedom to do our jobs, or our ability to achieve, it is  predictable that there will be a crisis. C is community. We want to be part of a community. I  believe the companies that can address these four things are well on their way to keeping people.  Just talking about this. Even if they are not getting the kind of stock options that they are  getting someplace else. I do know that if you do not satisfy these four things no matter how much  you pay people they will not stay. They will go someplace else where they can get those done. You  can see I have a little passion on this topic. 
 

I went thought this because I wanted to make a point about Doug's work. If you accept the idea that we are moving into an exponential world, and the business models, the way we do business  needs to be exponential, then you can look at other organizations and quantitatively say are they on  that path or aren't they? DOD research, corporate R&D, and incubators. I pick four things with Jeff  and Doug the other night. DOD research. It should go 6,1,6,2,6,3 they have a linear process for  innovation. How can they be successful in this world? They are not and they can't be. You can just  look at and the way that they do business. Corporate R&D. Put a lot of really smart people in a  building out in Montana some place, and don't let them interact with the market and all of the other  people in the company. You have to get permission from some vice-president before you can go and  talk to somebody, how can this be useful? It can't. This is basically going away in America.  Incubators. That is another one we came up with. It is a really curious idea. On the face of it, it  is a really good idea. Let's take a couple of smart people who want to form a company. We will put  them in a place, take care of heating, and give them janitorial services, but there are no ideas in  an incubator. It is a sterile environment, it doesn't feel like an exponential environment. 

You look at these and say what is missing? It become easy to look at them and say what is missing and what is added. This diagram is an attempt to begin to think about some projects that  would satisfy some of these requirements. Where you could go to people and challenge them about why  they would like to participate and ways that they may want to participate. I mentioned DOD on the  left hand side. That is a great example. They are so fundamentally broken that almost any  improvement would make a huge change about how they do business. So, what this says is go to DOD.  I'll do that. What they ought to do if form a NIC. To pick a problem that is important to them, and  use it as a benchmark for their thinking. Use it as a bench mark for how you are doing your R&D, and  the other parts of your business. Then borrow Jeff's idea from the last session. While you are at  that, why don't you take ten percent of your money that you want to spend on that and put it into a  MetaNIC. The tools and team box. Work on some hard fundamental problems, and learn from your  experiences that you collect from that. I put some things that need to be worked on. 

Collaborative environmental space, don't work today. In spite of what people say, they don't work. There is no security, they are not interoperatable, and you can't have the kinds of collaborative  tools that you want. They basically don't work. You can't do it. It is a good thing to work on. By  the way, if it is out of the context of an application, it obviously doesn't make sense to do that.  DOD here is something that you could make a big impact on, but putting them into the right  environment. Data bases, the CoDiak. I told the story about old folks telling stories. That shows  you how far away data base technology is from the kind of goal that is required for this. That is a  grand challenge and a good problem to work on. Human performance. I emphasize the human properties,  because they ought to be kept track of. When that lady I mentioned, Judy, said respect and trust,  that should be understood. What does she mean by that? How do we capture that and share it with  other people so that they don't have the same personal crisis the next time that they form a NIC and  basically do the same thing. Doug mentioned world help, I think that is another great application. I  am going to be giving a talk to the head of the National Cancer Institute and I am going to propose  that they form one. Education. We already have a bunch of ones going in education. So, if we could  get them to be taxed according to Jeff's rule, maybe we could begin to put together a MetaNIC. Maybe  other groups would form other NICs and they would have their MetaNIC, and you could get people to  share and they we would begin to increment recursively. If I could steal a word. Move toward this  vision that Doug has been laying out. 

Inclusion. I think that the goal of this needs to be exponential performance improvement. I think that is a grand challenge, I think it is inevitable. Frankly, we don't have any other choice. We  have to move in that direction. It requires important problems, champions and teams, lots of ideas,  recursive processes and more. 
 

Great human values and improvement based on respect and trust
 

I kind of like that and would like to acknowledge any ideas that I have picked up from a  lot of really smart people. Just mentioned some here Bill Welmont, who is a wonderful guy, we  actually have a draft of a book on exponential teams. Romanovski, a colleague from Saranov. My wife,  and of course Doug, so thank you. 

Audience: I'd like to pick up on the DOD idea. I want to share with share with you right away DOD came to us as a darba initiative, it essential said the same thing, the system is broken. DOD takes  seven to fifteen years to put the research into any kind of product that might be useful. As a  result they started an initiative called ACTION. We have been given part of that grant and have  already started working on it. We are planning a workshop in dealing with those issues. What we are  going to be dealing with is not the DKR or CoDiak part, because there are other people who are  dealing with that. We are talking about some of the strategies and issues that have to do with  business and economics, I would like to compare notes because we have some money to do that. 

Great, I would love to. The thing I think would be useful for DOD is to have an exemplar. They need something that they can constantly look at and compare what they are doing today with what  they ought to be doing. I think that is a great contribution from Doug. If we could work together to  create that it would make a great difference because basically they try things. When you look at the  trials you can find the things that don't work. It is broken in some really fundamental way. Even  thought it is interesting it is not going to be successful. 

Audience: I have two questions. The first is about benchmarking. It has been apparent from previous sessions at SRI that benchmarking is something important that is going on with in SRI. Do you have  any benchmarks for measuring your own exponential improvement. No offense. The second question I had  was regarding your comment on collaborative environments. What is your opinion on augment NLS in  terms of how it addresses needs we have for collaborative environments. 

I am not qualified to answer the last one. I will answer the first on however. We have very poor benchmarks. The way I see it is successful, is when you can watch it happen. When a team goes  from nothing to doing something that changes the world in six months. In retrospect you can see what  went on and it certainly was not linear. Whether it is exponential or not, it is certainly not  linear. When you see people who don't do this. In my experience, when all of those ingredients are  in place, the percentage of success is about 95%. When one of those ingredients is not in place, the  percentage goes down to about 5%. When we put together teams, if I don't see all of the ingredients,  we don't do it. That is one kind of benchmark. Now the actual rate depends on all kinds of things.  The web stuff that is different from biotech stuff which can take many years because it is a lot  harder to process systems and come up with the technology then it is to come up with the concept. 

Audience: Do you see the possibility to do multiple intelligence research. The ability to do far left brain, far right brain to tap into different forms of intelligence as tying directly into  exponential performance improvement. 

I do. I think that is one of the beautiful things about putting people together in a room, they think differently and so they develop different representations of the same problem. That is  one of the things that you want to tap into. I mention that the services do very simple stuff. They  give people a task and they have them solve them individually. Then they give them the same kind of  a task, and have them solve it as a team. Then the compare the scores. It is not even close. They  don't have to tell them the implication of that when they are done. 

Audience: Does Howard Gardeners work and some of the people that are around that have any direct influence? The interesting part about that is, take any of the practical implication of say Michael  Jordan's physical intelligence, to get more specifically into this area. 

Audience: Don't know. Jeff, what is the most important thing to you? 

Jeff: In what respect? 

Audience: What do you think about these things at SUNN? How do you think about what is important and what goes on in your company? 

Jeff: That is such an open-ended question. In some ways, I think SUNN has stumbled into some of these things in an interesting way with out really understanding it. I want to go back to the DOD  thing a little bit, and maybe abstractly tie these things together. SUNN unlike all the other  computer companies, SUNN hardly makes anything in the hardware that it ships. It is an interesting  phenomenon. So we don't have a FAB plant, we design ships, but we don't actually make them. We don't  make CRTs, we don't make power supplies. We don't make disks. We actually don't make any of the  components that we put together. Yet some how we have managed to do quite well, both in terms of  market growth and stock shares and stuff like that. So what is going on there, I can relate it back  to the DOD example in a way, we basically take components from all of the other companies in the  world and piece them together in a way that allows us to massive system integration value. Do that  in such a way that our component suppliers, we can choose and stay ahead of the competition in each  individual component. Yet all of our competition has availability to all of those components but we  have some how mastered a way of putting that together so that our system, total system integration  in what we deliver stays out in front. It is an interesting model for a company. Which I believe I'm  going to see a lot more of. It had a lot to do with what Allen talked about, about  interoperatability of the components and how we manage that and deal with that. It allows us to do  scalability. It is really interesting, cause we can scale up enormously to do things. I am not  taking about the research part that I am in, but SUNN as a company as a different kind of model.  Makes it an interesting primitive example of the way some of these things will go. 

Audience: I would just like to build on what you were saying about trust and teams. How an organization has got staff in Europe, both sides of the U.S., and Japan. Our membership is similarly  distributed. It is about the same. Either when we are working internally, just with the staff,  trying to develop product, services, or what we are presenting ourselves to....members or customers.  Or when we are grouping customers and suppliers, to try and build some consensus around a  requirement, an idea, or a vision, or a standard. The overriding important thing, in all of that is  trust. In a virtual organization, trust is incredibly difficult to build. If you are distributed  around the world, it is always the other guys fault. Communication is not the whole thing. It is  about knowing that you have clear understanding about what you expect people to do. About knowing  that they have a clear understanding about what they are going to do, and they are competent. It is  about other things like, I remember WWC, World Wide Web Consortion, once upon a time there is no  such thing as a virtual beer. If you are going to bond and build a team that is going to work  effectively together, occasionally you have got to do the beer and pizza thing. Then there is the  task of the senior people to make sure that it propagates through the organization in a consistent  way. Trust is a really huge thing. Underneath all of that, trust isn't just about honesty and  integrity. Are they able and capable, willing? At the end of the day, if people don't fit, it is so  hugely disruptive. 

Engelbart: Thank you very much, Kurt.

[  principal lecture]

 ---