Louis V. Gerstner, Jr. Keynote Speech

COMDEX '95

Las Vegas, Nevada

November 13, 1995

Thank you, Masa. Good morning.

Every now and then, a technology or an idea comes along that is so profound,so powerful, so universal that its impact changes everything.

The printing press. The incandescent light. The automobile. Manned flight.

It doesn't happen often, but when it does, the world is changed forever.

I joined this industry and IBM because I believe that information technology has that potential.

Now that I've been through our laboratories, talked to thousands of customers and even a few competitors, I am even more convinced that I/T is the defining technology of this decade and will be well into the next century.

I think that's why so many of us are here at Comdex.

When you get past all the glitz, this event is really about two things: enthusiasm for our industry and a keen interest in its future.

We'll see it, hear it and feel it when the show opens in less than an hour. We'll be bombarded by extraordinary new products and by promises -- oh yes, lots of promises -- of dazzling innovations that are just around the corner. You just have to find the corner.

All of this makes our industry fascinating and exciting. It's all about discovery and creating new things ... new technologies, new products, new capabilities ... the future.

So what do I think after 30 months on the job?

My expectations have been exceeded. This industry does have an incredible future. I say that for a couple of reasons.

First -- and this isn't surprising to any of you -- IBM researchers see no limits for at least a decade to the underlying technologies that drive our industry. Processor power, memory, disk capacity, bandwidth -- the trajectories will continue upward without hitting any scientific limits for at least a decade or more.

I don't need to spend a lot of time on this because there are examples of hot technologies everywhere you turn -- and I know a lot of you work on them every day. But let me share with you a few of the things under development in IBM labs.

A few weeks ago, I visited one of our major research facilities. I was briefed on a new kind of optical microscope they've invented. It has the potential of imaging a single atom in visible light, which would be unprecedented, and I'm thinking: What do we do with this? Where do we fit it into our product line?

It turns out that this same technology may eventually enable incredibly dense storage devices that can pack information at densities 100 times higher than anything we can demonstrate today. That's the equivalent of putting the entire collection of the Library of Congress -- that's 16 million books -- on a diskette the size of a penny. Now, I'm not sure if anyone wants to carry around the Library of Congress on a penny, but it may be possible.

Next, they showed me two ThinkPad flat-panel displays. One was a prototype high-resolution display ... the other was a color photograph attached to a ThinkPad display. They challenged me to distinguish between the two. I honestly could not tell the difference between this color photo and a color panel displaying a digital image. They went on to tell me about pixels per square inch and all that. I said: "Forget it. I need it. Just get it to me today."

Let me tell you about one other thing they showed me. It was a speech-technology interface that didn't just recognize dictation (we have products today that already do that). This interface understood what I was saying, it extracted the meaning of what I was asking for and delivered some information in response (something some of my IBM managers can't do when I talk to them).

I could go on and on with examples from IBM and other companies about:

-- microprocessors with several billion or a trillion transistors

-- intelligent agents

-- transponders the size of postage stamps that will be attached to everything and everybody.

But even more impressive than these individual technologies is how they're being put to use today in new applications. Some are really "out there."

Like what they're doing at the MIT Media Lab. One of their projects is called "things that think." The idea is to put intelligence into everyday, inanimate objects: door knobs, furniture, glasses, even shoes.

A computer in a shoe?

Actually, it's not such a crazy idea. As I understand it, you power the computer with every step you take, and when you shake hands with someone who's also wearing a shoe computer, you establish a low-voltage electrical connection and you instantly exchange information through that handshake -- sort of a "personal area network." Maxwell Smart, eat your heart out!

What happens is, you download to your shoe the person's title, what he does, where he works, phone and fax numbers, maybe a digital photo or a video clip.

If you think that's nutty, think about it tonight before you hit the sack. Try to remember the names and faces and affiliations of all the people you meet today... empty your pockets of all the business cards you've collected... then look at your shoes. Not such a wacky thought, after all.

So, in terms of raw technology, our industry's future is extraordinary. There's no letup in sight.

Now, we tend to think about our industry in terms of faster, better, cheaper pieces of hardware and software -- this product, that application. And that leads me to the second reason I'm excited about our future.

We're at the threshold of the next major phase of computing. It draws upon many of the technologies we've been discussing, but one in particular -- high-speed, high-bandwidth networking -- which is why we refer to it as network-centric computing.

Now, this term doesn't exactly roll off your tongue, but what we call it isn't nearly as important as what it means to people: to individual consumers, to businesses and institutions around the world ... and all of us.

As you well know, this phase of computing follows two others. The first was mainframe-centric. Naturally, we at IBM can talk about this one at great length because that's where we really thrived for many, many years.

What did that model of computing look like? A large central processor, centralized data, centralized applications, centralized systems and network management -- all serving hundreds and hundreds of "dumb" terminals. This wasn't a great model for end users because they were chained to the central I/S department for computing power and applications.

But the technology at the time simply didn't support personal computing.

All of that changed in the 1980s with the microprocessor, which made possible the PC, which made possible shrink-wrapped software, local area networks and on and on.

So the model of computing went from centralized to decentralized and, not surprisingly, so did the composition of our industry. We went from a few large integrated companies to 60,000 niche players, and that nearly blew IBM up. I remember. I was there as a customer. I helped light some of the fuses.

So computing was democratized, and we placed the power of information technology and applications into the hands of millions of end-users and consumers. And there have been a lot of advantages.

But the promise of distributed computing has not been fully delivered. Our industry promised customers that they could buy all these individual pieces and put them together very easily. But that hasn't happened.

It's harder than a lot of people thought, mostly because of system incompatibility -- the pieces were never designed to interoperate or share applications. And when you try to scale these piece-part solutions up to support thousands and tens of thousands of users -- as well as critical functions like heavy transaction processing and data mining -- it gets very, very tough.

Client/server computing has also been far more expensive than anyone imagined. It's expensive because of the complexities I've just mentioned and also because a lot of customers -- especially large organizations -- are waking up to the fact that they have put the equivalent of a 1985 mainframe on the desks of every one of their clerical workers -- with all the maintenance, backup and service costs now multiplied by thousands, and in some cases tens of thousands.

I think consumers are also beginning to wonder if they're getting full value from all of the horsepower and software capability they buy -- and the industry expects them to keep buying -- every nine months or so.

No question: the PC solution in many situations is less expensive and more accessible to people than the mainframe era... but is it really the end-game?

As I see it, that's pretty much the situation we're in today... customers continuing to buy increasingly powerful PCs, and they grapple to integrate them in these very complex client/server systems.

But we've come to understand that client/server is, in fact, not a full-blown phase of computing. It's really the leading edge of what will be the next phase: network-centric computing.

There are a lot of forces propelling us to this phase.

An obvious one is the technology itself. If you look at microprocessors, memory, software, storage... the laggard of the technology family has been communications technology. PCs and servers have become enormously powerful, but they communicate through the equivalent of soda straws.

Well, all of that is changing. Very powerful networking technologies -- principally ATM -- will be to the next phase of computing what the microprocessor was to the current phase.

But I think the most profound implication of this new technology is that it will change the nature of computing itself.

If the communications link between the PC and the network is cheap enough, fast enough and has virtually unlimited bandwidth, why not migrate a lot of the functions that currently reside inside the PC to the network -- the applications, the data, the storage, and even some of the processing?

This is not a throwback to the mainframe era. End users will still have the user-friendly features they value in PCs. They'll still click on applications, and response time will be just as good, if not better. And because we won't have to stuff this new device with lots of technology, it can be much less expensive to purchase, maintain and upgrade.

Frankly, I think few end users -- particularly consumers -- care where the processing, storage, data movement and all that stuff takes place -- whether it happens inside their tin box or somewhere in a network -- just as long as what happens in front of the screen is compelling, simple to use and is the least-expensive solution available. And, really, that's what it's all about - the way it should be.

Sounds pretty good. But what will end users give up? They'll give up a lot.

Like having to upgrade their PCs every year to get the next fastest processor -- or have that sinking feeling they've fallen behind the times. Instead, they'll "upgrade" by subscribing to higher levels of computational power on the network -- effortlessly.

They'll have to give up figuring out those arcane things called operating systems... why applications run on some but not others... why applications that ran last year are slower or don't work at all on the new version of the operating system. Instead, the network will mask that complexity and incompatibility.

And they'll give up having to buy a shelf-full of applications, swapping dozens of diskettes in and out, and repeating this delightful task when version 2.5.4 comes out in nine months. Instead, they'll get the new version through the network -- effortlessly.

And, of course, these applications will support collaboration and the interactive sharing of information. We already see it happening in the Notes phenomenon.

Some people, of course, will worry about flexibility and choice. The way I see it, customers will have infinitely more choice in a networked world. You can move from application to application without throwing away hundreds of dollars of software investment.

You'll be able to access more storage, memory and processing power than you could ever economically justify in a PC -- or even put in a PC.

I assume that all of you have at least one PC. Most of you probably have several. Unless you're quietly tapping on your notebook while I'm over here talking, all of those PCs are sitting idle -- in your briefcase, back in your hotel room, office, in your car, or your home.

Think about all that latent computing power that's wasted, totally unused. But in a truly networked world, we can share computational power, combine it and leverage it.

So this world will reshape our notions of computing and, in particular, our notions of the Personal Computer.

For 15 years, the PC has been a wonderful device for individuals. But, ironically, the personal computer has not been well-suited for that most personal aspect of what people do: We communicate. We work together. We interact. So maybe we'll end up calling these new kinds of devices I-PCs -- Interpersonal Computers. We'll see.

Now, I have to be careful here.

One of the things I've discovered about this industry is that it absolutely thrives on hype. It just loves hype. It's constantly prowling for the "Next Big Thing" to promote, with the presumption that everyone will instantly move to the next wave, and all that preceded is dead. It's the most brilliant example of planned obsolescence I've ever seen in an industry.

The network-centric world will not replace the PC world entirely any more than the PC world replaced the mainframe world entirely. Yes, there will be fundamental changes in PCs. Many will be like the inexpensive appliances I've described, which draw their sustenance from the network.

As a matter of fact, we're building one today in one of our labs, and we expect to be shipping it to selected customers early next year.

But let's not lose our perspective. There will still be very powerful standalone PCs to perform applications that are not network-centric, just as mainframe demand today is the highest in history.

So rest assured, I'm not here at Comdex trying to win this year's industry Pied Piper Award.

When will network-centric computing arrive? Well, it's already arriving -- just look at the Internet, the most powerful and important of all networks. People, organizations and networks are connecting at a rate no one expected, and that pace will continue, if not accelerate.

I think we'll see network transactions -- electronic commerce -- take hold very quickly. The final step -- pushing the applications, operating system, storage and processing up into the network -- will take longer.

A lot of people are looking for network-centric computing to pay off in the consumer space first -- interactive services to the home, 500 TV stations -- but we are seeing it take off first in businesses and large institutions.

That's because while I've talked a lot about technology driving us toward network-centric computing, technology is only an enabler of a much more powerful force: a whole new way that institutions, and companies, have conceptualized their strategic priorities. They've discovered a powerful new form of leverage: the leverage of organizational knowledge as a means to compete more effectively and to differentiate themselves in the marketplace.

When I talk to people about powerful networks, the discussion inevitably turns to "content." Who owns it, how to deliver it, who pays for it. A lot of people think of "content" as movies, music, artwork, sports scores, weather.

The most important, most valuable, most sought- after content in the world belongs to corporations and large institutions. It's created and collected every minute, around the world: intellectual property, designs, market intelligence, supply and demand, customer trends.

The problem is, this knowledge is spread across incompatible computer systems, it's distributed across PCs in various departments, it's locked up in desks and filing cabinets and, yes, some of it is locked in the heads of employees.

Powerful networks can unlock this knowledge and move it to the people who can use it effectively. This drives shorter cycle times and allows teams of people to become more productive and creative.

These are the kinds of concrete, real-world benefits that are convincing companies to make the necessary investments in these powerful networks. So I think the consumer will be affected, but it will happen as companies and institutions push out and build the links to reach them.

Something very significant is happening here.

For the first time, the models of computing supporting businesses and institutions on the one hand and individual users on the other are coming together. They're becoming one and the same. They're converging in network-centric computing.

Frankly, most of the knowledge in businesses and institutions around the world resides on IBM legacy systems. That's why one of the major thrusts of our network-centric computing strategy is to help our customers get their valuable content to the right people: to employees in far-flung locations, suppliers, partners or -- and we're seeing more and more of this -- directly to the consumer. Let me give you an example.

I would argue that Federal Express today doesn't compete on the basis of its 500 airplanes, 35,000 vehicles and 110,000 employees -- although obviously those are very valuable resources. I think FedEx competes on the basis of making vital knowledge instantly available to its customers.

FedEx has put a front-end on its databases, which happen to run on IBM mainframes. That front-end enables customers via the Internet and online services to track their packages around the world, day or night, every step of the way. FedEx already has that information, of course, to run their worldwide enterprise. But now they're using networks to get that information in a secure way to customers and seek differentiation and competitive advantage.

As companies use networks to push out to their customers directly, we'll see dramatic changes in the nature of competition. Companies that stand between a supplier and a customer will be on dangerous ground.

Maybe you're familiar with what Levi Strauss is doing. They have a Notes-based system that goes in a store. It takes four measurements from you. Those measurements go directly through the network to a factory where a custom-tailored pair of jeans are cut for you and are expressed to your house. They charge you an extra ten bucks, and you're delighted to pay ten dollars for a custom-made set of jeans.

And guess what? Levi's has eliminated the need to have mega-retail stores stocked to the ceilings with pants of every conceivable waistline, inseam, cuffs, no-cuffs, relaxed fit, traditional fit. No huge inventories. No obsolescence, no waste. And the supply chain they've depended on for decades - and that has depended on them - is obsolete.

And the implications of network-centric computing go on and on. It will transform every business, organization and institution in the world. It will create winners and losers. It will change the way we do business, the way we teach our children, communicate and interact as individuals.

A while ago I met with Nelson Mandela and he was telling me about one of the greatest challenges facing South Africa. One of the legacies of apartheid is that the best education infrastructure is concentrated in a few major cities. What were formerly the "homelands" have very little infrastructure. It will cost millions and millions to build it and it will take years. We got into a discussion about distance learning, about bringing teachers in cities to students in the homelands via interactive networks. I saw a light bulb light up.

He's very interested in this as a possible solution to the problem.

As the network world develops, we're seeing combinations of networks being built and interconnected: Private networks to support secure mission-critical business applications... At IBM, we're building a lot of these for customers in industries like health care, pharmaceuticals, finance and travel. These applications are hosted on the IBM Global Network -- which we're rapidly upgrading to ATM capability -- and customers subscribe to applications on the network.

These private networks are interconnected with other corporate networks, with public-switched networks and, of course, the Internet.

We are making enormous investments in Internet products and services because we think the Internet represents the most powerful manifestation of what network-centric computing will be.

The Internet has been a wild and wooly electronic frontier. The settlers are arriving. Every day, the Internet grows, in size, in reach and importance. Every day, people add more function, more security and capability to the net.

You may have seen our announcement at Internet World a few weeks ago. We're putting together Lotus Notes, InterNotes, IBM network software and services... the idea is to help our customers reach their customers by securely linking their core business systems to the Internet.

Notes -- with its collaboration, security and replication features -- is the ideal way to do that. It's exactly what customers have been telling us they need in the Internet.

So I hope you get the idea that we're taking network-centric computing very seriously at IBM. I'd say we're betting much of our future on it.

We at IBM know all too well that those who dominate one phase of computing are not necessarily the ones to lead the way to the next phase. In fact, if you think about it, they are the least likely to accelerate the transition. It's in their best interests to maintain the status quo for as long as possible -- we certainly tried to do that in the mainframe era. And as we found out, when the industry moves to the next phase, the current leader doesn't always have the right skills to lead the next one.

The skills and technologies of the PC-centric, piece-part phase are not the important assets needed in network-centric computing. What is important? Complex network management, systems management, heavy transaction processing, massive databases, powerful scalable servers, systems integration...

Now, let me think... what company knows how to do all that?

Obviously, we think many of these requirements play to IBM's strengths... and they open doors for many other companies in our industry... and for an entirely new breed of competitors, some of whom have already burst onto the scene... and others are probably just getting started in a garage somewhere.

Now I know a lot of what I've been describing about this new networked world may not be new to your ears. In fact, many Comdex keynoters who have preceded me on this very stage -- and perhaps a few who will follow me -- have proclaimed or will proclaim the same.

Some have put on multimedia extravaganzas depicting future scenarios -- how we're all going to be connected, work differently, play differently. Everything will change. Everything will be better.

But at this very important juncture in our industry's history, I'm reminded of what Dickens wrote in a "A Christmas Carol" -- maybe because the holiday season is upon us.

Remember when Scrooge encounters that last ghostly spirit who reveals the future? Scrooge asks: "Are these the things that will be... or might be?"

I think we're looking at the same question in our industry. We can see the future. It's there before us. The question is: Will it happen?

I think it will, but only if we step up to certain responsibilities as an industry. If we don't, we will not grow as quickly as we'd like, and we will not realize the potential of our technologies -- no matter how impressive they may be.

All of these challenges fit under one banner: listening to the customer.

I don't think we pay nearly enough attention -- or spend nearly enough money -- being sensitive and responsive to customers. I think I can make that statement since I was a customer of this industry for 20 years. I've also led consumer-oriented companies, and I can tell you, we have a lot to learn.

I think we need to work on many things, but three in particular.

The first is ease of use.

This is a tough, multi-dimensional problem. It ranges from basic confusion over on-off switches and loading software to the way we integrate and manage complex systems.

Think about it: Entire segments of our industry have been spawned from this failure. Look at the explosion in computer services. Customers are spending billions for smart people to help them figure it out, pull it together, manage it for them.

Consumers, of course, run into problems all the time.

We all have our favorite examples of clunky products. But let me pick on my own company.

You may recall the 1983 debut of this exciting product...

[ Video of TV commercial for IBM PC Jr.: ]

[ Stops commercial ]

That's enough!

Customers -- lots of them -- told us what we could do with the PC Jr.

We got some things right with that product, but a lot of important things wrong -- like its infamous "chicklet" keyboard. It was really cute. Too bad no one could use it.

Or remember this baby?

[ Pulls out IBM PC Convertible ]

This is the IBM PC Convertible, circa 1986.

I think people called this the "lovable luggable."

At 14 pounds, I understand the "luggable" part.

It had no hard drive, but it did have dual 3 1/2-inch diskette drives.

Very advanced. Too advanced.

There were only a handful of applications on 3 1/2-inch diskettes at the time.

I'm happy -- very happy -- that our company has learned a lot from those early disappointments. Today we've got retractable keyboards and trackpoint and voice recognition. And I know a lot of other companies are investing in ease of use. But we can and must do more -- as an industry. If we expect the whole world to be connected, and if we expect to grow as an industry, we must make it easier to learn, to use, and to maintain.

The second issue we've got to focus on is standards -- open industry standards.

Ours is the only major industry I can think of that refuses to rally around standards. I really don't understand why customers have tolerated it.

I mean, would you buy a telephone that only dialed certain area codes... or a TV that received only odd-numbered stations? I wouldn't.

Think about the networked world that we see before us... every digital device connected to every other digital device in the world, all supporting seamless, easy access. How are we ever going to get around the problem of incompatible hardware and software systems?

I think we have two choices.

We can ask customers to set aside their freedom of choice and preferences in hardware, operating systems, applications and user interfaces... junk their trillions of dollars of investment in information technology... and all of us -- everyone, everywhere -- move to one architecture provided by, priced by and controlled by one company.

Or ...

We can embrace open industry standards.

Open means that software from one vendor can operate on or with hardware and software from any vendor -- not just one guy's. We need to work with standards organizations. We need to openly agree on APIs, interfaces, tools and protocols -- on anything the customer sees and touches in the journey to get something done.

Compliance with standards does not mean that we won't compete aggressively or that we can't distinguish our products. We will. But we'll compete on the basis of innovative implementation of industry-standard technologies and architectures, on performance, features, design, service and support.

Besides, in the long run, closed, proprietary architectures -- that's a losing strategy. I bet you thought you'd never hear that from IBM. But having had a near-death experience, we know what we're talking about.

Every time I meet with customers, I say the same thing. I urge them to demand compliance with open industry standards in the products they buy. And you know what? They're beginning to listen. They understand the need for the industry to move to this level.

The third and final issue I want to talk to you about is broader than these two other issues.

Our industry operates in a free space. It's part of the reason for its success. This freedom must be honored and protected -- but it also must be earned.

We've got to come to grips with the fact that the proliferation of our technology raises some very provocative and serious societal issues.

Some people feel that the Internet is getting out of control... that instead of being an information superhighway it's a dangerous pipeline for pornography, intellectual property theft, and a serious threat to commerce and privacy.

Governments are going to be increasingly concerned with the impact of our industry on the future of the nation-state -- which has been the basic form of political and societal organization for centuries.

People are concerned about universal access... about creating a society of technological "have's" and "have-not's." Just last week, I read a survey in Time magazine. Nearly half the respondents said that technological change was responsible for the growing economic inequality between people.

These are serious issues. There are many different and divergent views on what to do about them. Frankly, most of the solutions that are being proposed are equally troubling. They raise questions of censorship, over-regulation and, of course, the specter that the fires of innovation that drive our industry will be stamped out.

Yes, these are tough, tough questions. But we can't shrink back from them just because they're potentially explosive.

Let me tell you why. A few decades ago, there was another "defining" technology. You may remember it: nuclear power. It, too, was going to change the world. It was going to make energy as cheap as water... bring air conditioning to deserts, heat to frozen tundras.

Reactors would power cars and ships, airplanes and spacecraft. Inexhaustible energy.

But what happened? The nuclear power industry is a dormant industry today.

Those grand visions were never realized. Why? Was it because of bad science, disappointing technology? Not really.

While that industry was focused on megawatts and reactor cores, communities and governments were worried about the implications of that technology. They stopped that industry dead in its tracks.

Today, our industry has grand visions. I've shared some of that with you this morning. And they are grand. And very ambitious. We're talking about no less than changing the world in very fundamental ways.

Yet, too often, what are we focused on? What will you hear and see here at Comdex? Megahertz, gigabits and multitasking. There's a disconnect between our priorities and those of a lot of people and governments.

So, I'd like to leave you with the thought that the time has come to step up to these challenges. Even as we continue to innovate and create, we must now also think about the broader implications of the future we are creating.

We must do this as individual companies, but also as an industry. And, in fact, there's quite a bit of activity already underway to address some of these issues.

The Worldwide Web Consortium based at MIT is working with many companies, including IBM, on filtering technologies and content rating for parents to use.

The Center for Democracy and Technology, also with the support of many companies and advocacy groups, is working on free speech and child protection.

And trade associations like the Computer Systems Policy Project are working on privacy and cryptography.

I'm sure these organizations have a lot of good thinking under way, but we have to be careful not to confuse activity with results. We've got to be impatient, drive toward consensus. We've got to lead rather than be led.

And we need to keep doing our part as individual companies and put our money where our mouth is. I think what AT&T is doing to bring the Internet to schools across the country is an outstanding initiative. At IBM, we're investing millions on a program called "Reinventing Education," which is aimed at reforming public schools with the help of information technology.

The printing press. The incandescent lightbulb. Manned flight...

It is an extraordinary and rare -- and humbling -- opportunity to earn a place in history.

I, for one, think we will.

I think we will make our visions reality.

I think that as stewards of our industry, we will step up to the challenges of social responsibility.

We will improve the world, and the way we work, the way we communicate, live and learn as people.

That's why as much as I've talked to you today about various phases of computing, I also wanted to emphasize that we must now also pass into a new phase as an industry.

We have grown, we have innovated, and we have prospered at a rate unsurpassed by any other. It's been an amazing, breathtaking ride. It can continue -- and accelerate -- if we remember that our future rests on how well we respond to the total needs of society and of our customers all around the world.

Thank you, and I hope you have a great day at Comdex.

Copyright 1995