Remarks by Bill Gates
Microsoft Corporation
Cambridge University
Tuesday, October 7, 1997
Cambridge, England

MR. GATES: It's great to be here. It's a little odd in a way for me to be speaking before a university audience because I am a drop-out myself. I don't recommend it. I think it'sgreat to finish school and I wanted to give you a little bit of my background and talk a little bit about where I seethings going in the future. It was actually when I was inhigh school that a friend of mine, Paul Allen, came and showed me an article in Electronics Magazine about themicroprocessor. He and I had been doing a lot of software. In fact we found out we could make money doing software.

The greatest scam was we discovered that by getting the job doing high school scheduling, we could decide exactly what boys and girls were in our classes, and that was an incredible reward. It really motivated us to learn how to write interesting software. But the thing that Paul showed me about the microprocessor that was pretty amazing was that here was a computer on a chip, and if you took exponential extrapolation from that, you could say that the computer industry would change very dramatically. And we sat there and said, "Geez, does Digital know that their world has been upended? Does IBM know that things will never be the same?"

Well, of course, change takes a lot longer to happenthan you sometimes expect, and so it was almost a decade later before the big upheaval in the computer industry took place. I had decided to drop out of Harvard because I was worried that somebody else would start the first microcomputer software company, and our unique vision was that software was important. Up until the microprocessor came along, really hardware was at the center of everything. If somebody could make a new machine 30 or 40 percent faster, they had to just completely start over on the software, new operating system and new compilers, because hardware was very, very expensive. So, in fact, nice advances like symbolic debuggers or profilers sometimes got done on old computer systems and there was never time to carry them over to the new computer system.

But with the microprocessor making things inexpensive, several important things began to happen. First of all, thechip companies made microprocessors and they didn'tknow anything about software, so instead of just having thesemonolithic companies that did the instructions and operatingsystem on all the tools, my co-founder PaulAllen and I saw a chance to get in there and create a puresoftware company. In fact we had been very frustratedbecause we'd been calling up people like IBM and Digital andsaying, "Would you like our compiler or would you like oureditor?" and they said, "No, we do that stuff ourselves andyou guys are just high school kids anyway. Why do you keepcalling us up with these things?"

It was kind of ironic because we bugged them enough that when they had a project that got them into a lot oftrouble, for instance where they were taking all the dams in the Northwest United States and connecting them together, theyunderbid the project and were way behind, sofinally, when they were trying to figure out how to salvage the project, they said, "What aboutthose guys who call up up all the time? Why don't we go and hire them." That's how I managed to skip my senior year ofhigh school and meanwhile get at the programming. I reallygot the bug, so that I only ended up in college a couple ofyears before it was time to go out and really make sure thatwe got out in front of the software world.

Another key point in the separation of hardware andsoftware is that hiring and managing software people is verydifferent. You don't want to have a company that tries tomix those two things together, simply because software's tooimportant to mix it up with those other things. You want tobe able to be neutral to all the different hardware companiesthat are out there, and that had the incredible advantageof giving customers more choice. The structure of the PCindustry with that separation was really a novel thing andhad very profound effects that are still being felt today.The debate still rages on. Should a hardware company definethe operating system or should it be a company who'sindependent who goes out and does that? People like Sunstill disagree with us. They would like to see an integratedhardware/software approach.

Now, the original vision of Microsoft was that there would be a computer on every desk and in every home, and theworld is very far away from that today. Even in the UnitedStates, where about 40 per cent of homes have computers,that's less than halfway towards that goal. So I think maybewe're halfway towards achieving the original vision, and it'scertainly going to take a lot of improvements in hardware andsoftware to see this happen. I still believe very much inthat original vision. In fact, everything that's happened inthe last few years around the Internet just reinforces for mewhat a breakthrough it is to broadly give people a tool withthe power of the future personal computer.

Due to the wonderful benefits of Moore's Law, the microprocessor is now a million times faster than when Idropped out of school. In the next 20 years, it will be a million timesfaster again, fast enough to run the software of the future. We want to bevery ambitious in terms of what the software does and thatreally brings us to the whole idea of doing research, beingwilling to take risks and do things that have a very long term view.

It's exciting to be tied up in that partnership with Cambridge. This was a place a lot of the great original workon programming, sub-routines and microprogramming took place. Today, it's a leader in many aspects of computer science -- automated reasoning, natural language, and many other areas,so it was a very natural choice for us, partly because ofthe incredible strength of the university, partly because ofthe entrepreneurial spirit around here. That's why we're pleased that another aspect of whatwe're doing here is contributing to the venture fund thatHerman Hauser has and, and we hope that will allow some of the greatideas to move from the research environment out into themarket-place.

The explosion of the Internet is just an amazing thing. You know, I met with the Prime Minister today and he knew theword Internet. Five years ago, prime ministers didn't talkabout things like this. He knew to talk about technology ineducation being an incredible priority, and aboutfinding the way through that challenge in terms of who's goingto install the networks, what technology should they use, howdo you change the curriculum. There are a lot of things to belearned about how this can be done properly and it willprobably take more than a decade before all of that's figuredout. But I'm a very big believer about what this can mean.

One milestone for myself recently was I moved into a new house, and I had been waiting about seven years to move intothis house. It took a long time to build. It's got about 50Windows NT machines that help run this house. It'sfantastic. It turns the lights on and turns them off --really advanced stuff like that! You know, it's a littledifficult right now, first moving in. Sometimes it playsmusic I want, sometimes it plays something else, but it'skind of an environment of the future.

I've got all these high resolution screens. I'm a bigbeliever that there will be breakthroughs in screentechnology with machines that will change the way we thinkabout dealing with electronic information. Even the mostavid believer in technology today probably does the vastmajority of their reading off of paper. I read the Economistoff of paper every week. I will go up to the Web site to seeif there is an article about our industry, because it takesthree or four days before I get the delivery, but most of itI'll just wait until I get the paper copy. The reason isit's just not that comfortable reading long things off of thescreen. I recently got a 20 inch LCD screen. They're stillfairly expensive. It shifts the balance a little bit. I'llread say a three-page memo off the screen whereas before Imight have only read a one or two-page memo before I would goand print that out.

I do think in order to really get to the next level ofrelevance in society, to get people to live what we call theWeb lifestyle where they turn to it many times a day andeverybody is connected up on electronic e-mail, we need amuch better user interface. Computers are still prettycomplicated. Microsoft error messages get to be famousbecause they're so cryptic and there's a lot that can be doneto simplify that.

The big breakthroughs are the things we're tackling inresearch. A few years ago, there was a big fad of deliveringcomputers that understood handwriting. Well, the machinesweren't good enough but I still believe that can be done.The hardware and software just wasn't right. For the last 20years people have been demonstrating systems that understandspeech. In fact when I dropped out to start Microsoft, Ifelt kind of bad. I said I guess I will go and do all thesetools and meanwhile my friends will stay and solve the speechproblem.

Well, here we are 20 years later, and it turns out it'sa much, much tougher problem than anybody expected. Doingthe demos is fine but until we link up a system that hasessentially all of our common sense as well as our linguisticunderstanding as well as the acoustic models, we probablywon't get anything that you want to use as part of youreveryday interaction. I am optimistic enough to believe thatwe are certainly within five to 10 years of being able tosolve that problem, and I think people outside of computerscience underestimate what we're going to be able to do.

In the next one or two years, sure, it will all beincremental. Bigger storage, faster speed -- nothing verydramatic there -- but as you get the higher band width, asyou get satellite systems that connect every point on theglobe with high speed connectivity, as you get the ability tospeak and write into these systems or even to have a visualcapability, you are talking about something very, veryprofound.

People will understand why we call this the informationage. And the cool thing about this is that the magicalelement that will make it all work is software. It won't besimple software like the 4K Basic that was Microsoft's firstproduct. That was nice. One person could write it. I stillhave it memorized today because it was so intense trying tosqueeze it down into 4KB of memory. That was for theinstruction set-up and the data tables and the program thatthe person would put in, so the Basic itself was only about3KB. Programmers today come into my office and say, "Yeah,it's only 300K, it's small." I think, "Wow, this software'sgetting big." So, thank God for all the power that's beingdelivered.

The image of what a computer is 10 years from now iscertainly very, very different than what it is today. That'swhat makes this a very fun field. What is an operatingsystem in 10 years? Sure, it will have memory managementand file systems, but the bulk of the code will relate tothese natural interface systems. So we're following all thegreat research that's been done in those areas and we'refunding a lot of research in those areas, because we thinkthat is not only the future of Windows, but the future ofcomputing itself.

It's amazing to be able to say that after 22years working in the same field, pursuing the same basic ideathat we had when the company was started, this is still asfascinating a business as it was at the very beginning. It'salmost scary how much impact it's going to have. You know,scary that these heads of government want to meet with us andtalk about what should they do, and how will this all rollout. But as long as this is happening, being at the centerof it is a great place to b,e and that's why I think I havethe best job in the world. Thank you.

Read the Q&A with Bill