Bill Gates |
October 21, 1993
|
|
Bill Gates
Chairman and Chief Software Architect, Microsoft Corporation
It's a very exciting time in the information age; it's the very beginning. I want to share with you today my enthusiasm for the developments to come, in particular, this "Information Superhighway." But in order to get some context - and perhaps even to establish some credibility - let me go back to the early days of the microprocessor and the PC.
When we say that it's the "Information Age," what we mean is that people should have easy access to information of any type, whether that's for business use, information, entertainment, or collaboration. Anywhere they go, they should find that easy to do.
What are the key elements? What are the key tools here? Well, of course, the microprocessor is one, and it is subject to a rather magical rate of improvement - a doubling in power every two years. That's a rate that we see, if anything, accelerating in the decade ahead. The easiest way to think of it is simply to think of computing power being almost free - infinite power at no cost.
When the microprocessor was created, it had a very humble beginning. It was developed to run elevators, to control calculators. Even the very people closest to that invention didn't appreciate the impact that it would have. At that time, myself and Paul Allen were students, very young, and we were shocked that the microprocessor didn't seem to be widely talked about. This first phase of creating these new tools was one that took place really not being much talked about. It was a really surprising thing to the computer companies that software and software standards would completely reshape their business. So it's very different than what we have today, where I can hardly pick up a magazine without saying, "Oh, no. Another article about the Information Superhighway."
That means that the development will take place in a different way than it did up until now. With the microprocessor, one of the greatest insights was understanding what the business structure would be. The computer industry, up until the advent of these advances, had been highly integrated. Companies did everything. IBM had its own sales force, its own instruction set, its own operating systems, and they were able to generate, throughout the '80s, four and even $5 billion a year of profitability by having that structure. And everyone else in the industry followed exactly that same approach; they were modelled after IBM.
With the pace of technology taking place in the late 1970s and early 1980s, the supply of computing was going up so fast that there was really a qualitative change in the way it was being used. This necessitated the creation of a very large software industry. But the economics of software are very simple: You can't afford to sell software at low prices unless you sell it in very high volumes.
So even the early personal computers - where every company came up with their own model and we had over 50 different types - it was not going to be sustainable. We weren't going to get a large market. What we needed were standards that spanned all of these different machines, and really separated the market for building the best hardware from the market for building the best software, and allowed each to proceed at exactly their own pace. That was the thing that allowed for the creation of standards like MS-DOS, and now, Microsoft Windows.
It's really amazing, even to me, to see how far the software industry has come. Of all of the major industrial categories, it is the fastest growing in the United States. It also is the industry that is most dominated by the United States of any major industry. All of the large software companies, and the majority of new, exciting small software companies, are found in the United States; in fact, a very high percentage of them, not too far from where we are right now.
The PC has rapidly become the tool that's used for the creation of documents. Most word processing, spreadsheets, presentations start sitting down at your PC and building that document. But, in fact, we haven't really realized the potential of electronic documents whatsoever. Because once you create the document, you usually print it out in paper form, you usually send it across a fax machine that moves it into a very low level, you usually go to a meeting with a lot of people and shuffle paper around. So there's a lot that can still be done to deal with information flow and the kind of efficiency that the future will require.
One element of this is revolutionizing all office equipment. When you send that fax, it should go in high-level form - it should be secure, it should be an editable document, it should print out at high quality; that's very straightforward.
The only thing that has stood in the way of that is that in the past, the companies that tried to re-architect the flow of information in the office, they used the same kind of approach that the computer industry had used in the past. That is, people like IBM, and Xerox, and even Exxon for a period of time, tried to come out with a family of their own hardware equipment. The standards only worked between their various pieces of equipment.
Now the approach that has worked so well in the software industry - of having standards that span all of this equipment - will be brought to office equipment. For the first time, you will be able to move documents across companies, track where they are, allow people to do annotations, and bring them back together, using the installed base of fax, network connectivity, and PC networks, and it will all be very straightforward.
The best way to take advantage of these new technologies is not simply to go out and buy the latest and greatest and hand it to all of your employees. The best approach involves stepping back and thinking about: How does your company work? What are the key steps? How should information move? What are the key databases?
For example, when you have a contact with a customer, is all the information about how you dealt with that customer - what the status of their account is, what complaints they've had, who else has been trying to help them in your company - is all that gathered in one place, so that the minute they call in, it's easy to call up that information? That's very straightforward; it's something that people are more and more demanding that their suppliers have. And yet it is only by consciously thinking, "OK, we're going to take this technology and apply it in that way" - that it can be achieved.
Likewise, looking at internal processes, employee reviews, business planning, product development: How could these networks and these tools be used to make them more effective? Perhaps Microsoft, in its zeal for the technology, takes it to an extreme. But something like deciding whether you want to make a contribution to United Way, the way it works is: you come to your office, you sit down at your desk and there, in your mailbox, is this nice little message that says, "United Way contribution." Well, it's not just a simple message with some nice words from me - although that is part of it. It's actually a little program we've written and sent around as a mail message, so you can click on it, and indicate what type of agencies you're interested in, get data about those agencies, see the history of your own giving, see how that compares to other people in our company, and in high-technology, and by going through that very interactive process, at the very end you make your choices, click the "done" button, and that's sent back, electronically, into a database. So that's immediately entered into our payroll process. There's no manual overhead. Literally, day by day, I can say, "Are we getting good participation or not? Do we need to have a few more rallies to get the really important message out about how we think that fits in?" So that's one example where electronic contact makes a big difference.
I think it's fair to say that Microsoft would not be organized the way it is today if we didn't have electronic mail. As I'm out traveling the country, every night, I'm connecting back to electronic mail, sending off the messages that are sitting on my machine that I've created as I've been flying around - those are all going back into the company. And I can just, as I'm connected, click on a single icon and see, "How are sales doing? Are we behind forecasts? Are we ahead of forecasts?" Hopefully, as soon I see it's all OK, I can close that icon and move on, confidently, to other things. So there's a lot that can be done with the tools as they exist today. In fact, they are continuing to improve at a very rapid rate. That's why we hear so much talk about this idea of a "highway." The key technologies are not just the microprocessors; there's two others that I'd really like to highlight.
One is: the way that you interact with the machine. What is the quality of the screen? Can you have a "personality interface," where the machine appears to have learned from your previous interactionsand is kind of talking to you, and suggesting things? Is that possible? The answer is: with the magic of software, it is.
We will move from the graphical interface to what we think will be termed the personality interface. It requires a lot of very complex software, but it doesn't need to be slow or expensive because of the incredible power that the hardware will provide.
A final piece of the technology is networking technology. How does data move around? There is a new approach to very high-speed networking called "asynchronous transfer mode." The beauty of this technology is that it actually brings together for the first time the requirements of real-time networking, like voice networks - where you can't afford to have little gaps in the conversation - and data packet-type networking - which is the way computers talk to each other.
So now we can have all the existing networks - whether it's broadcast TV, or phone lines, or local area networks, or these cable TV networks - they're all talking about adapting exactly the same technology, which is a two-way, very high speed, so-called "broadband" network. This technology is coming down in price very rapidly. That's why it's very appropriate that today we find a great interest in this highway opportunity.
As I have said, I am on the extreme; I am on the lunatic fringe of believing in this highway opportunity. Many times in the past, I have believed in technical opportunities, and in terms of what would happen, I have a very good track record, but in terms of predicting when it would happen, I have often been over-optimistic. A good example of this is that over a decade ago, Microsoft had its first conference on software being distributed on CD - CD ROM-based multimedia software - and at that time I said that, very rapidly, medical information, sports information, encyclopedias, buying catalogues, phone books, information of all types would be put onto these very cheap disks, and a high percentage of personal computers would actually contain the drives that would let you use this kind of software. Well, in fact, it's taken five years longer than I expected, but I can say with some confidence that in the next several years that really will catch on.
We often see a phenomenon in technology that when you don't have enough users that are working in a certain way, it doesn't make sense for people to come along with a software to help out those users. But once you reach a certain critical point where software developers see an opportunity, then they start to write more and more software, and because they write more and more software, more and more people buy machines of that type. That's called a positive feedback loop. Once you get to the critical point you see this very exponential, dramatic, effect. Some examples of that are that we went from having those 50 types of personal computers, down to having only two types of personal computers, within a period of 12 months. As all the software developers figured out that the DOS-compatible machines and the Macintosh were the only two that would really get to critical mass. Now we see those same seeds have been sown in terms of people writing software on CD, and so we're seeing lots and lots of great multimedia software come out.
The information highway is at a much earlier stage than that. It's something that people talk about, but the types of applications that will be used, the people that will win in this world are still uncertain. What will be the key business assets? In the PC world, it turned out systems software and applications software were very valuable. What will it be here? Will it be owning the network connections? Will it be making this hardware? Will it be taking branded content, like the L.L.Bean sales catalogue, and putting that out on this network? Nobody is quite sure, including the people that are most involved. It is amazing the way that, even at this early stage though, it's so covered in the press. I talked on a fairly regular basis with people like John Malone and Ray Smith, and I was complaining to them that I have to spend a half hour each day just reading articles about them just trying to figure out what's going on.
The term "mania" is very appropriate, and like any mania, there will be a lot of excess here. There will be companies that are funded, and years later we'll look back and say, "Who funded that company? What was it; was that just mania at work?" There will be investments that are made that turn out not to be effective. It's easy to do things that, in the long-run, don't pay out. For example, when the microprocessor was first invented, the people who appeared to have the deepest understanding of how it should be used were companies that built dedicated word processors.
Wang soared in size, doubled in size, because they did the best job of taking the microprocessor and word processing and putting those two things together. Well they missed the really fundamental idea that the PC would be a very general purpose device, and that the software and hardware would be separated. And so their corporate fortunes waned, even though they were in such a strong position.
When I think of this highway, I think of an incredible breadth of applications. Medical diagnosis, if somebody prescribes a drug to you, it should be easy to get up and find out about it. Or if you want urgent help, it should be easy to get somebody to come onto the screen and talk to you. Take education; if a teacher thinks they are doing a great presentation for their class, how would they let other teachers know? How would they find out if somebody is doing a similar thing? Well, with this system, it will be very straightforward to get up on that system and say, "Who else is presenting the Civil War? What did you use?" and take that down, bring it to your system, edit it, in certain very simple ways, and literally have the best practices spread very efficiently throughout that system. The kind of bulletin boards of information that will be easily accessible will be very diverse. The Parent Teachers Association, the Canoe Club, the local merchant. In a certain sense, you won't think of your community anymore as necessarily being the people who happen to live physically next to you. Rather, you'll think of communities that are formed very easily across this network. People who want to go hiking with you, people who share a political point of view that you want to get together and promote that point of view.
The applications people keep coming back to because they're concrete, and in fact they'll fund most of this build-out, are video conferencing, and video-on-demand. And those are important applications. We can make the user interface of picking a movie a lot of fun. You can see the reviews, you can see who else has been watching this movie and what they said, this personality can come up and say, "Based on these movies you liked in the past, we recommend the following movie for your enjoyment." In fact, if you're willing to put out on this network your response to these movies, some people might be interested enough in your response to pay even a few cents, just to call that up, and so your account on the network would be credited if you are actually an interesting person.
There will also be incredible disintermediation with a system like this. Let's say somebody is a very popular newspaper columnist. There's absolutely no reason in this system - that will be totally open - that they can't simply go out on the network by themselves, with their own name name as essentially their brand name, and say, "Here is my column," on a daily basis, "every time you look at it, one cent is deposited into my account." Something like an HBO or a Showtime - or even the very TV networks themselves - no longer are necessary as the vehicle to get information out.
Of course, all of this presupposes a major investment in infrastructure, between the new box and the home, some optic fiber going up to the head-end and these head-end servers. It's around $100 billion that has be spent to put this all into place.
That's just the beginning, because then we need to get lots and lots of applications on top of it. I think the most optimistic time frame is that some really good pilots will get done in '95 or '96. Personally, I am disappointed with all the pilots to date, because if you take these pilots, where they just have crummy interfaces and a narrow set of services, in a way they disprove the proposition that's important, which is: Do customers with these systems generate enough activity to justify the fixed-cost of the build-out? So I think people should have held back a lot longer and had richer systems in place, because so far, all we've proved is that crummy systems generate crummy results. In other words, I won't let any of those systems deter me in my belief that this is a great thing. Microsoft is today investing $100 million in building the software platforms for these systems. We're just one of the many companies that believe in this very much.
One of the things that is a little bit scary is that as computer technology has moved to new generations, there has never been a case where the leader in one generation was the leader in the next generation. Certainly, in retrospect, we can say in the PC generation, that Microsoft, Intel, Apple, and a handful of other companies are the leaders of this generation. So at least from a historical perspective, we are disqualified from leading the new generation. Now partly, people say that almost by definition, because they start a new generation when the leaders have fallen. But we are attempting to defy past results by building a great platform, making it very easy to use, making it available to phone companies and telephone companies to allow it to get out there on a very rapid basis. So that's really something that I find very exciting, the kind of empowerment, the ability to make markets work better than they are today, to allow every buyer, every seller, every person who wants access to something, to go out onto this network and find it very easily.
I think that it has profound, positive effects for every area of work and home life. That's really why when people say to me, "After all the things that I've done, why I don't retire?" I tell them that I have the best job in the world and that this whole era is going to be a lot of fun. Thank you.
Generously funded by:
Tom Campbell
Dee Dee Myers