The Connected PC

Andrew S. Grove
Hong Kong
May 2, 1996

Sean Maloney: Thank you very much for coming along this morning to our event, the Connected PC. This morning's speaker is no stranger to Hong Kong, having been here something like twelve times. But in the nature of the evolving role, ever expanding role of the PC, each time he comes here, he brings more equipment. I had the pleasure of working very closely with this morning's speaker for a number of years, in the last capacity helped organizing Fall Comdex in 1994. And that seemed to us then to be a fairly high-tech, complicated show, and I think we had something like 10 Pentium® processor-based PCs in all interconnected. We've got something like 14 here, and probably twice the bandwidth, so this is a fairly major show we've flown in, I guess you could call it Las Vegas transported to Hong Kong with all the risks of high-wired demos that entails. This morning's speaker has been with Intel for all of our 27 years, he's been president for the last eighteen and chief executive officer for the last seven. Please join me in welcoming Andy Grove.

Andy Grove: Thank you very much and good morning. Our subject today consists of just two subjects. I'm going to talk about the connected PC, which, you might have guessed, will somehow come up, and I'm going to talk about a phenomenon called hybrid applications, which you probably didn't guess because they didn't put it on the banners.

Let me start right off by talking about what the connected PC means for us. And to do that, a bit of a review of the history of the PC. The personal computer started as an office productivity tool 15 years ago or so, the home use of personal computer was implemented by people taking home or purchasing a computer just like the one that they had at work, and doing similar things at home as they did at work. They were doing wordprocessing, bankbook balancing and the like. Then as time went on, multimedia came out as a defining application for the personal computer, we now are talking the early 90's, the introduction of audio, of CD-rom, the introduction of built-in modems became part of this phenomenon.

Last time I was here 18 months ago, I demonstrated what such a multimedium, multimedia PC could do. It was a Pentium PC doing some games and some education software that was quite stunning, at least by the standards of 18 months ago. At that time, we, at Intel, along with our various partners in the industry, embarked on helping the Asian countries leapfrog in their use of personal computers. And instead of following in the way some people thought they might, kind of a year, two years, 18 months behind their western counterparts, realizing that they had an opportunity to leapfrog ahead of that development and start with modern technology.

Eighteen months later we look back and we look at a report card on that. What you find is, first of all, multimedia PC has become a reality, multimedia PC has become far richer than we even thought would be possible 18 months ago. We'll demonstrate that just in a moment. And the leapfrog concept in our Asia marketplace has happened far more than anything that we ever believed would be possible. Example being the whole region has skyrocketed for the year as a whole for 1995, from 1994 4% to 1995 30% Pentium processor as a portion of total shipments. The Chinese market itself, by the end of the year, has crossed over something like 60% of personal computers being built on Pentium processors, much of them high performance Pentium processors, so both of these actually happened. But before going any further, I would like to demonstrate what 18 months meant in terms of the same application in a multimedia PC. Jeff, come on and show us.

The application I demonstrated to those of you that may have been here at the meeting, had to do with applications that depicted the human body in three dimensions.

Jeff Lo: This is what you showed 18 months ago, 3D body. You were able to look at a skeleton and rotate it around in 3D. But in 18 months, the processors have gotten so much faster, and the software vendors have been able to take advantage of that to the point now in addition to showing the skeleton, we can start adding in the other body parts, the internal organs, the skin, the muscle, and still rotate it around in 3D.

Let's take a look at it in a little more detail. Another thing we can do, here we can see a much closer view of the organs. We just click on the heart, just simple point and click, and now we have a much more detailed view of the heart that we can rotate around, also to look at, see more text on the screen describing the heart - all with just a click of the mouse.

Andy Grove: Let me interrupt Jeff for a minute. Some things happened here that I would like to call attention to because you will see them appearing in various other demonstrations throughout the presentation. The way Jeff called up that heart is basically by clicking on the representation of the heart in the body, and all of a sudden, in that instant, a full representation of the heart showed up elsewhere. This technique, which is a graphical version of the hyperlink technique, has become an absolute defining feature of modern PC applications. And you'll see them throughout all of the demonstrations. Now proceed.

Jeff Lo: So we can look at the heart in even more detail than this, we go back to our main menu. There's a 3D theater here, I'll go ahead and click on this, and what we'll do now is we'll see on the monitor here as we fly into the body in a 3D animation. Move on in and then we'll be able to select from a number of different body parts that we want to look at. Once again, we've got the heart, let's go ahead and look at this in more detail. So here you can see a 3D rendered animation, along with audio which is going to describe everything we see in the heart as we go.

Andy Grove: Again by way of explanation of what is taking place. An enormous amount of computation is taking place in this machine in order to bring you this very simplistic looking presentation. Somebody rendered all these bodies, turned them into ones and zeros in vectors, into coordinates, and this machine is fast enough to bring all of that into real life in the form of a 3-dimensional, virtual reality type ride to body organs, fairly complex and fairly well rendered. All of that was practically inconceivable at the time we demonstrated the 3D body a year and a half ago. So that's one example.

Now the beauty of this industry, and the curse of this industry, depending on which side you are on, is that nothing ever stops, the same way as in 18 months we went from this very simplistic version in retrospect, a 3D body, that yet, 18 months ago, I was proud enough to show here to a much more elaborate version, eighteen months from now we're going to look back and say, "Gosh, how pathetic was the demo that I showed in Hong Kong in 96."

What is the direction that the development is going to be taking place? Excuse me, but before I go on, before I go in the direction, I want to show one more thing, which is one of my favorite games. I almost forgot to show you. This is a pinball mac hine, this is the one and only demo on this stage that I know how to use.

What you have here is actually an object lesson in physics. You know a pinball machine consists of levers and balls ricocheting back and forth in there. It is not random, and it is not controlled by a cursor. It is controlled by the laws of physics, exactly as a real life pinball machine. As the laws of physics are being computed and applied to this pinball machine, a pinball bouncing around, it is also rendered so it looks 3-dimensional on the real time, again, a tour de force of computing power.

So let me see if I can launch a ball. I'm going to play this until I lose the ball, or until you get tired depending.

Now we are ready to go out to the future. The same capability will be the foundation for developments going forward, and the direction where we are going forward is going to be the connected PC.

The connected PC, that is kind of a compendium of technology the same way as the multimedia PC was a compendium of connected technologies.

First of all is going to use a new generation of microprocessors which introduce the first instruction set architecture of change in Intel architecture microprocessors in the 386. It brings with it a set of instructions called multimedia extensions,or short we call them MMX technologies, that were particularly designed and conceived to accelerate multimedia and communications type applications even beyond the basic native speed of the processor.

We are going to combine this with built-in Internet access, built-in conferencing capability, all of which you will see in some form in the future. 3D graphics that will make the kind of 3D graphics that we have seen here look elementary in retrospect.

And lastly, we hope to incorporate a development that has been in the works at various consumer electronics companies for the last ten years called digital video disk or DVD technology.

It is going to be a veritable compendium, candystore of technologies and will render the experience of today's PC user, basically, unexcited by comparison.

Now, what is driving beyond the capabilities of the technologies, or what is driving these technologies is the confluence of three different forces. The first of these is high performance microprocessors. The second one is the presence and ubiquity of the Internet. And the third is this category of applications that I call hybrid applications. We will talk about each of these in some detail as I go along.

Now first let me talk a little about microprocessors. And the way I want to do that is by introducing to you a picture that I call Intel in one "fall," or one slide, or Intel on a postage stamp, which is the generation of microprocessors top to bottom and the generation of silicon technologies with increasingly fine geometries left to right. The way we do our work, and the way we make our microprocessors better is like this - We introduce a microprocessor, start from the top left hand corner, introduced the 386, at the 1.5µ technology, this was ten years ago. We introduced it in the marketplace, but even as we do that, we developed a new version of it on a tighter technology, 1µ technology, which would go into higher volume production, which would be lower cost because the "die" is smaller, the chip is smaller, and higher performance. We used that technology to introduce the next generation of microprocessor. Then we shrink the next generation of microprocessor and the next generation of silicon technology and we keep repeating this cycle, like a perpetual learning machine, until we work our way down, diagonally down on that slide where we end up with today's 0.35µ technology running the Pentium® Pro processor. Now the benefit of all of this as seen by the user is continually higher performance at continually lower costs.

When you add that all together and look at the cost per measure of performance, million instructions per second, the cost per MIPS, and look back over on an almost 15 year period of time, on a semi-logarithmic scale, where the cost is shown on a logarithmic scale in order to show the range of reduction, you see a straight line dropping down, running at approximately 30% per 18 months reduction. This is the direct outflow of the Moore's law phenomenon that every 18 months we can double the number of transistors that we can practically put on a silicon chip. When we apply that to performance and to cost, this is what comes out of it. Now this is good, it's a little theoretical, all the semi-logarithmic stuff like that, so I would like to demonstrate it, what this curve means in terms of permitting capabilities, that were expensive before, almost prohibitively expensive, and turning them into completely, generally available pervasive applications. And the application I want to do this with is through Proshare Videoconferencing.

Proshare is a technology we introduced early 1994, just two years ago. And we brought to Hong Kong actually, introduced it a year and a half ago at my last visit. When we started designing Proshare, we thought we would use an accelerator in conjunction with a 386. By the time we finished the design, we realized the market and the "installed piece" of computers went to 486, so we made the accelerator cheaper and we required 486 as a base. Even with the 486 and even with that accelerator, we required a high performance digital telephone line called ISDN, in order to provide reasonable video images on the two ends of the connection. What has happened since then is that the perpetual cost reduction machine has continued its work. We went from 486 toPentium processor, from early Pentium processor that operated at 60 MHz, to 75 MHz, to 100 MHz and 133 MHz. We discovered that with Pentium processors that operate on 120 MHz or higher, we were able to put the processor to work and do the compression job without any add-in boards, without anything added on. And on top of it, we were able to do the compression sufficiently well that we could do the Proshare work, the Proshare conferencing on plain old telephone lines. And that's very important because there are a lot more of those available at a lot lower cost anywhere in the world, than ISDN lines. So this permits us to take Proshare, which is meant for business application, and bring it to consumer application, letting ordinary people communicate with each other.

And I would like to demonstrate how that works. The phone is ringing here, so let's see who's calling us, and I should say this not a simulation, this is Hong Kong Tel at work.

Hello! Who's there? Well, there are a lot of husbands in this audience so I... Oh Sean Maloney. Is that you Sarah? I'm not sure Sean is here, he is probably traveling. No, he is here! Wait a minute, he is coming. This is Sean's wife, looking for Sean. And in order to authenticate that she has got the genuine object, we will have to use Proshare for that purpose. Sean, here you are.

Sean Maloney: OK. Can you hear me? Yes this is actually, definitely sounds like my wife. The telephone there is actually a telephone that's routed through the PC. This call is being made entirely in software over the PC. But it would be a lot more fun if we could see her so I could find out what's happening. So let me just turn to this control here, and we're going to send video.

Now Sarah, if you could just hold on there a second.

Now this is a software only video phone, using a standard Pentium PC, and this is a type of modem called DSVD, the modem line, which is really done in software by Pentium, is capable of sending voice and data at the same time, and the data we're sending is going to be a video image. So this is a software only video phone.

Well, and it is my wife, good morning!

Sarah Maloney: -....

Sean Maloney: No! Just five hundred odd folks sitting here. So....go ahead!

Sarah Maloney:-....

Sean Maloney: Well, like everybody else in Hong Kong, I tend to travel a lot, so I'm sorry about that.

Sarah Maloney: Oh by the way, I've sold some stocks... And I've been shopping.

Sean Maloney: Ah. Like everybody else in Hong Kong.

Sarah Maloney:-....

Sean Maloney: Yeah, the fung shui master told me not to wear red, that's very true. So you've sold some stock, have you fixed this?

Sarah Maloney:Yea, I've bought you a new tie.

Sean Maloney: Aha, let's have a look.

Yeah, there's no red in there, that looks pretty blue to me. OK.

Sarah Maloney: Alright, I hope you'll like it.

Sean Maloney: Yeah, it's great, thank you very much, and I promise I'll see you soon, as long as you don't sell anymore stock.

Sarah Maloney:-....

Sean Maloney: OK, good bye.

Andy Grove: Thank you Sean; thanks Sarah also. This is what high performance microprocessors can mean, features absolutely not available, then available, and then available as a built-in part of all computers because basically software and microprocessors can do what before add-in equipment was required for. So this is our first driving force.

The second driving force is the Internet. And, just so we all work from the same page, I'll offer you a working definition of what the Internet means to me. It's connection, a collection rather, of interconnected networks, that allow PC users to access data any time, anywhere.

That's good. And you're going to say, the Internet is not new, so, why is it a factor now?

Well, the Internet has been around actually around ten years longer than personal computers. It is an old phenomenon. In its first, approximately speaking 22-23 years, it has amassed a total of approximately, and these are approximate numbers, please don't attach any significance to them except to illustrate an order of magnitude, point 100,000 users in the first 20 odd years. Then in the last 2 or 3 three years, depending on whose numbers you believe the number of users skyrocketed to many millions, possibly several times ten millions worldwide.

What happened to the Internet today is these things: The access to that has been simplified by Internet service providers, who by and large, give you consumer usable access whereas before it was quite complicated to become part of this connected network of networks.

The second is a new piece of software, modern browsers, allows very easy navigation on the Internet, again, replacing arcane commands that took place before.

And lastly, high performance PCs that were largely deployed over the strength of the multimedia wave, allow these PCs to support high speed local processing. And this is an extremely important part of the phenomenon of why the Internet has become today, because it addresses one of the key elements, that is involved in the Internet, which I call the first myth of the Internet, and that myth is that browsers don't require processing power.

Browsers are complex software. Browsers, just as an aside, I don't know how many of you might have realized, Netscape Navigator 2.2, which is the most popular, most numerous browser software available today, requires 11 megabytes of disk space - that is approximately one third of Windows 95. It is not a small piece of software. And Netscape Navigator is an evolving species, and as with other cases of software, as it evolves, people load more features and more capabilities on it. The newly announced Netscape Navigator 3.0 in fact comes to 15 to 17 megabytes of diskspace, amounting to approximately half of Windows 95. This is complex software.

Complex software requires decent amount of storage. But more important, it requires high performance processor to provide a lively and crisp user performance. And that's what it does. Today's Internet, with high performance processor, provides exactly that. I'd like to demonstrate a couple of examples of that.

Jeff, come up again. One of the uses that this lively performance can do is get you interested in otherwise deadly dull corporate information, like the Interlinal report. Let's see the Interlinal report.

Jeff Lo: What we have here is a Pentium PC with Netscape 2.0, and this is an Intel page leading up to our annual report. What we have got is a shockwave version, using a shockwave plug-in of an animated version of the annual report.

Andy Grove: These are the most deadly words of the Internet. Better get used to them.

Jeff Lo: Here you can see it's not just a static page. It's not just text and a couple of pictures. It's this animation with motion video, things flying across the screen. This is not something you could do on a simple PC, you need a high performance processor to do it this with this kind of speed.

Andy Grove: And it requires very modern software, like Macromedia Shockwave. Pull up some information for us, something from the facts and figures. This enables us to go anywhere on that, and provides lively looking information, even makes financial numbers that otherwise would be dull, 50% annual growth that kind of stuff, and renders them far more interestingly.

But this capability goes far beyond, this is the first generation of Internet use, Worldwide Web use is in fact bringing corporate information, marketing information, online and jazzing it up with the technologies of the Internet.

Then immediately behind that comes a set of applications that will develop with this medium, the medium of connected PCs on the Internet in mind, specifically for that. And one of the better examples of that is seen after, which we can pull up, is our daily Internet newspaper, ..., newspaper without paper, I mean, we need a new set of terminology. Give us some demonstrations of that, first of all, the scrolling headlines are exactly the same kind of technology that what we have seen on the annual report, right?

Jeff Lo: A little different but similar. This is using the Java programming language here to display these different headlines, updating them all the time as we're looking at it, we've also got ads that appear on this, they're different every time you pull up the pages you've got different ad.

But they've also got like news. They have the C-Net radio which includes a bunch of the top stories that they've been covering. In fact one that just showed up yesterday was a story here on C-Net about the new Proshare products that we announced back in the US. So again, here it's got more animation, wherever we go these things just keep getting more and more complex motion on the screen and even audio, they've got a related story here on the Internet, again, just point and click, and now we're bringing up an audio stream.

Andy Grove: OK, what is driving all of this application in the consumer's space is the evolution of the connected PC in the Worldwide Web as a medium. As a medium, as a new medium if you wish, it is competing with all the old media, newspapers, magazines, static, outdated, but visually enticing. Television, one way broadcast, the word broadcast suggests one to many transmission fit for the lowest common denominator of the many, not personalized. These old media have these disadvantages, but they can be made visually compelling. And for the new medium to be able to lure viewers or readers from the old media, it has to be visually compelling. Consequently, multimedia enrichment, 3D graphics, video, sound, are all flocking in there as a means of making information, ranging from news to annual reports, visually compelling so that you can compete with people who are accustomed to that level of excitement, visual excitement, on the old media.

Now that's not the only place the Internet is present. In the same kind of technology, Internet technology, is penetrating corporate and business use. And it has been given the name Intranet. Again by why of definition, Intranet is where we use Internet technology that fortunately the modern organization of many corporate networks enable to be widely adopted in corporations. Once adopted, it can be the basis for the adoption of very inexpensive groupware and collaborative applications, which have been coming kind of slowly over many years, but whose adoption has been limited by simply the fact that they were each proprietary and consequently each very expensive. And in the same way as personal computers being low cost, mass-produced computing equipment and popularized computing all over the world, much faster than the proprietary computer technology that they replaced, groupware application based on Internet technology will do the same thing to groupware. And lastly, this is a space and an application where high performance microprocessors, such as our latest generation of Pentium Pro, have very ready place of providing high performance and cost effective use.

The way this is going to happen in our view is by a movement of business users to our new concept of business computing that will embrace 32 bit applications. Applications that are sufficiently powerful and rugged for reliable corporate use by the built-in implementation of the desktop management & instrumentation, which take advantage of the fact that all desktop computers will be connected back to the single connected PC. Therefore, once the desktop machine is instrument, administrators and network managers can do all the necessary observation, troubleshooting and maintenance job from one place as they used to do in the old days, when they used to sit in the glass house rather than going crawling under your desk. Lastly by providing performance that is sufficiently strong, it has headroom for all of these applications, and gives leeway for this investment to pay off in the next several years.

So I would like to demonstrate one application of what this Intranet, corporate Intranet applications, will look like as it gets deployed. Jeff, explain what we have here.

Jeff Lo: What we have here is a simulation of a departmental network. We've actually brought in a dual 200 MHz Pentium Pro server running Windows NT and an Illustra relational database and that's connected over an Ethernet to this Pentium PC here running Windows 95 with Netscape, the very same Netscape browser we are running to surf the Internet. What we can do is use the same point and click interface to access all the information in our relational database, instead of having to use a custom application that you would have to learn each time you used a new database. So, I'm simulating that I'm a customer of Intel's, and I just got an order for notebook computers. I need to make sure we have enough components to build those, so I'm going to search the system for the components we use for mobile computers. We'll just type in, in a standard Netscape search form, the word mobile, and tell it to search the database. It makes a request to the server, to the database, that's hidden from the user, just happens in the background as I click, and it comes up with the 75 MHz mobile Pentium part and the TCP-package.

Andy Grove: Obsolete stuff.

Jeff Lo: Yeah, that's one of the older ones, but we can look at the information here. But, like you say, it's one of our older components, so let's look at something newer. I can use the database to search based on the appearance here of this component. Tell it to find all the parts that look the same since all the mobile processors are in TCP-packages. I do the search and what comes up, the newer ones, the 120 MHz and the 133. Go ahead and again, just point and click and up comes that record out of the database. Here we can see the part.

Andy Grove: A couple of things are going on here, and one of them is obvious from Jeff's explanation. We are accessing a relational database for a commercial purpose. That's the large, basic picture that's going on. But as I said earlier when we were starting to look at the 3-D body demonstration a little while ago. And since then, in all of these demonstration, something more profound is taking place.

The point and click routine, that literally a 3 year old child can learn and does learn by the thousands everyday when playing the storybooks, is all you need to do to run a relational database search on this corporate application.

So what is taking place here? It's the use of Internet technologies, high performance multimedia running modern browser software, simplifies the most complex operations that are taking place underneath this layer. And any action that a child can learn is becoming increasingly sufficient to run these complex applications all the way through to relational databases. What this means is technology is being used, and performance is being used to hide its own complexity, and make the use of this immensely complex technology trivially simple to millions of people with zero training. Because what Jeff did today, anyone of you or anybody that you are associated with, could do in five minutes of training. And that is one of the most subtle, and most important developments that Internet technology, browser technology, combined with high performance capabilities brings to both the corporate and the consumer's space. So thank you Jeff.

Now, the next drive in here that is going to take us forward is the emergence of network multimedia in place of "distendable" multimedia. This is the kind of thing that you saw with the pinball machine and the 3D body. Network multimedia is multimedia delivered on networks. It is the delivery of voice, video, 3D animation over an Internet style network inside a company or outside a company from one PC to the other. As I mentioned today, network multimedia is galloping relative to the speed with which stand alone multimedia took to become pervasive. And it is doing so because of the media-like characteristic; and therefore the media-trained expectations of the user-community.

And this brings us to what I call the second myth of the Internet, which is that there is plenty of bandwidth, or if there isn't, there will be real soon.

Let me deal with bandwidth-reality. Bandwidth that is most commonly available where there is any, is plain old telephone service pots. This represents, again, I am using the same kind of semi-logarithmic plot because I have to deal with a large range of bandwidth values. This is 28.8. We are talking about ISDN, something that telephone companies all over the world have been developing for 15 years, and probably worldwide no more than a million lines are deployed today. And that moves it up about fourfold. If you are in a commercial enterprise and you can access a T1 line and can afford to pay for that, this is typical in the US, it's a couple of thousand dollars or so a month kind of fee, you move up that far. And if you are really rich and youreally want high bandwidth, you can go up to a T3 line, that moves up even further.

Now there are two other alternative delivery mechanism for bandwidths, one is direct broadcast satellite which is sort of in the same range as T3, and the other one is cable modem which is a little bit lower. Both of these are reasonably high bandwidth, one is deployed fairly broadly, the other is also a hope like ISDN, but both of those are also only one-dimensional so that obviously limits the way we can use them. Now, let me explain the significance of the semi-logarithmic plot.

Another way to look at this: How long would it take to deliver one minute of video on each of these networks, and when you take these bandwidths and compute, calculate how long one minute of video would take, on a pots line it would be eight hours, eight hours of connection time to deliver one minute of video. And as you move up to things like direct broadcast satellite and cable, you get to the vicinity of 10, 15, 20 seconds. The real improvement in delivering bandwidth, video capability, is right in the computer. When you put on what an ordinary available 8X CD-rom can do, or you put on what the disk access can do from your local disk, it is in the general range of T3s and DBS and cable, able, for instance, your harddisk can deliver video in ten seconds time, one minute of video in ten seconds time. Now there is something magic about one minute by the way. When it takes less than one minute to deliver one minute video, you can have real-time video. If it takes you more than one minute, then all you can do is download video and deal with it later. So getting over that one minute barrier is very key, and the reality is that is going to take us an enormous long time, if ever, before a large number of population will have the kind of access time, the kind of transfer time that you can have out of every computer today. So what this suggests is that there is a major dissonance, discrepancy between bandwidth myth and bandwidth reality. And what I would like to propose is that the answer to resolve this dissonance is a category of applications that we call hybrid applications.

What are hybrid applications? Hybrid applications are applications that recognize that there are two sources of bits in a personal computer. One from the local resources and another one from whatever that computer is connected to. And it judiciously uses these mixtures to allow the user to enjoy the benefits of both, covering the complexity of where the bits come from, shielding the user from that complexity, the same way as the point and click routine shielded the user from the complexities of arcane things like relational databases.

So what we are talking about in this picture is a computer where the complex content is delivered from the CD-rom and resides in the harddisk. And the updated, or interactive content, is delivered via a connection scheme from the Internet. This is the answer to the bandwidth dissonance. This is the answer that is going to be available to every computer user in the world in one form or another because the combination can be changed and adjusted to the reality of the bandwidth scene, today, tomorrow,and five years from now. And what I would like to do is demonstrate two very ..., this is a new concept, a relatively new concept, as a concept, although software designers are already intuitively using it. The applications I'm going to use are very early stage applications to demonstrate this; and I think before the year is out you're going to see far more exciting applications than this going ahead.

So the first application ..., Jeff, come back please and help me. Let's do an Internet application called Pointcast.

Jeff Lo: This is Pointcast here. It's an application that lets you tell it what you're interested in, examples like news stories, companies. And when you're not using it, it goes off and downloads it so you have it available. Here I've told it I'm interested in a number of companies including Intel, and so, it's gone off while I've been out and it's pulled down the current stock quotes (it will know how many ties Sean will be able to look forward to when he gets home) But it has also got stories relating to Intel, including the Proshare story that we looked at on C-Net, only I didn't have to wait for it to download. I just clicked here and it was already there because it has already been downloaded onto the harddisk.

Andy Grove: And let me emphasize this because this is a critical thing. Everything Jeff is doing there is taking place between interaction of his local memory and his harddisk. And it got to the harddisk while you were not looking, or while the computer wasn't busy, or in the "bag" run, at its own speed. If he had more bandwidth it would have gotten there faster, if he has little bandwidth it would take longer to get there. But once it's there, he will experience the same access time and the same response time out of this Pointcast as, no matter how the information got there.

That is one way to do that. The other way is, one of the problems with CD-roms, is that they are obsolete. By the time you purchase a CD-rom, particular CD-rom containing applications, information, by the time you put them into your computer, they are obsolete. They are as obsolete as an encyclopedia. One of my frustrations as a kid was reading things about an encyclopedia that I already knew were not true because I've learned in school the opposite of what ten years ago was proper chemistry dogma. So let's take a look at one of the most popular encyclopedias of any media of all kinds, Encarta.

Jeff Lo: Yes, this Encarta ‘96, the latest version available of it from Microsoft. It allows us to go and look at a number of things. The quickest way to find some things is to click on find, and it will bring up a box where we can type in a subject that we want to look at. Anything you want to see today?

Andy Grove: Let's pick something like "chess".

Jeff Lo: OK, I type in "chess", and it has got a subject here for chess, and it's brought up off the CD-rom, diagram of the board, some instructions...

Andy Grove: And note how fast this is happening. It is happening immediately because it's coming from the local resource. And note all Jeff needs to know is point and click, which he learned when he was working on the relational database, so he knows how to use Encarta.

Jeff Lo: But you know a lot of things have happened. There's some stories here about previous grandmasters and what not, but you know it wasn't too long ago that there was a big match with a computer and a grandmaster.

So what we can do is we can tell it to go to its yearbook builder. Here it let's us go out to the Internet and update what we purchase on a CD-rom. So I'll go ahead and tell it I want to add new months, and it gives us all the instructions here. It tells us: we need to update through April; we have only got it through August now; it gives us all the instructions on how to add the months; we can get it from Microsoft's Webpage on the Internet. So all I got to do is bring out Netscape again, and go to Microsoft's page on the network. It's pulling everything down. We can click on the file list, get the file, so we'll go there. Again it's got a list of all the different months that are available for this, September through April. So I've already pulled those down, and all I need to do is click on the integrate button. Now it's taking.....

Andy Grove: That integrates the new information, the little bit of new information to the masses of the old information that we have, rendering the old information up to date.

Jeff Lo: Now we're up to date through May. The next update is available later this month. I click on close, we're back to our chess page but you'll notice this is new. There is a new article available that has now been stored on the harddisk. And what's it about? It's about Kasparov playing the computer, IBM's supercomputer. He stumbled at first, but in the end, he beat Big Blue's computer.

Andy Grove: That may have something to do with the choice of microprocessors in this computer.

Jeff Lo: Should have had a Pentium® Pro processor.

Andy Grove: Thank you Jeff.

So what you see here is basically the very early stage implementations of what I believe is going to be a new software paradigm. The paradigm of hybrid applications. All personal computers will draw their applications from three different sources of bits: the material that is in the personal applications, the personal data, and the local storage that is available at a very high bandwidth no matter where you are; information that is available on the Internet in the fashion that Jeff demonstrated to us in these last two examples. And lastly, information that is available in private networks or switch networks, telephony, like Sean and Sarah did in Proshare.

All of these are capable of integration together in such a fashion that when you use the computer, you don't know where your data came from. They all look alike, they all operate, you all operate with them in the same point and click method that you've seen throughout this presentation. Frankly, you don't care where your information comes from because technology takes care of it all.

So that kind of wraps it up. And just in very very simple terms, I think this industry is guaranteed to be extremely exciting as these technologies get delivered in a pervasive fashion. And I believe the story for the next decade, for the remainder of this decade and beyond, will be the story of Connected PCs running hybrid applications, and achieving the best of all possible worlds.

Thank you very much.

* Other names and brands may be claimed as the property of others.