By DAVE GUSSOW, Times Technology Editor
© St. Petersburg Times, published January 3, 2000
Paul Horn, who directs IBM's team of 3,000 researchers worldwide, has an insider's view of technology's evolution into the 21st century.
His company is focusing on two major areas for development. "Pervasive computing" will take technology off the desktop and into an array of gadgets, such as cell phones, that will link us to a network in our everyday lives. And "deep computing" increases the power of machines to unprecedented levels. Last month, IBM unveiled a project called Deep Gene that will create a computer 500 times faster than anything available today. Its mission: unlock the secrets of the human body, potentially assisting understanding of diseases.
Horn recently talked to Tech Times about the course of technology. Here are excerpts:
Q: You have predicted that the PC will lose its throne to handheld and embedded devices. Do you have an idea of how and when that might happen?
Horn: The PC isn't going to go away, so don't think of a death knell for PCs. We're beyond a lot of eras and that doesn't mean the technology doesn't still exist.
The point is that other devices are going to be rapidly playing bigger and bigger roles.
Have you seen the IBM micro-drive? It's this little disk drive that's a little bigger than a quarter, and it goes in digital cameras and potentially all sorts of pervasive devices. Now, that drive pushes this technology harder and faster than the drives that go into PCs because it's smaller, so you've got to get more bits down on a smaller space. It's got to be shock-resistant cause it's going into cameras and all sorts of things. It's got to have low power because it's in devices that aren't connected to some big power line all the time, so in many ways that device is pushing our disk technology faster than the devices that go into your PC. Lots of those things are today getting pushed by what we call the pervasive devices, these things that you carry around with you, at least as much if not more than by the PC.
Q: The reliability of technology is not particularly good. How did the '90s change user-friendliness, and what do you see coming up?
Horn: My hope for the next decade is that where we took little baby steps in the last 10 years we're going to take some very big ones. So you'll start to see new technologies like natural language being more important, making it easier for you to interact without . . . keyboards, without ALT-CONTROL-DELETE if you want to stop something, without having to know this morass of things to make the PC work the way it does. And I suspect it will be with much more appliancelike devices, simpler, more special-purpose, more special-function, and just very easy to use.
Q.: One of the things your lab is working on is interfaces that are almost personalized, that the computer would recognize a child user versus an adult user and respond to that.
Horn: Once you start to have the systems be a little bit smart, then they can look at what you're doing on the computer and it can determine whether you need more help or need less help, the level of sophistication of the operations you're doing, and it can provide you with a set of information that's tailored to your immediate usage characteristics. It can learn about you. Furthermore, it can keep a memory of that so the next time you log on, it knows this is a sophisticated user or this is a beginning user. And it will start to give you more advice or less advice, depending on your needs.
Q. Some of the issues involving pervasive technology may affect individual privacy rights. The industry, especially the online industry, has a spotty record on protecting personal information. How does the industry build trust with the public on protecting privacy rights, and how much of a concern are privacy rights for people such as yourself who are developing the technology?
Horn: I think it's a huge issue, and it's a bigger issue in different parts of the world. It's right now a very, very serious issue in Europe and has been. The public has been more willing to take chances with their privacy in the United States, but frankly, your statement that online companies haven't done a very good job of this is right on the money.
Now, what can be done? From a technology point of view, since we've been focused on this issue, we're providing things like ways for companies to control, or parents to control, access to Web sites, or to have control of what kind of information gets stored, let's say, about who has been on what Web site. So there's sort of a set of underlying technology that will enable a company to provide and guarantee to its customers that they are going to protect their privacy rights.
The other thing, from a policy point of view -- I think it's very important that the industry demonstrate that there are ways to protect people's privacy and there are a number of suggestions on how we can do that, like having a privacy mark on the Web site which will tell the consumer that certain privacy standards have been adhered to. . . . The more the consumer says, "I need to see a privacy mark, I need to see something that's going to convince me it's safe doing business with this organization, safe interacting with the site," the faster those standards are going to get accepted.
Q. Is there a danger of gadget overload as more information appliances come out with pervasive computing?
Horn: I think you will get people on different sides of this answer. It's sort of the question of: Is there going to be a Swiss Army knife? Are you going to get one appliance that's got all sorts of things in it and you get your corkscrew and your fingernail clipper and whatever else is in a Swiss Army knife, or are you going to have separate, individual devices? And my own personal belief -- and now, this is a guess about how the world is going to be -- is that it will be very specialized. It will be very much a model where appliances tend to be single or have a few functions rather than to have many, many functions. And we don't seem to mind that in the kitchen. You've got all sorts of different devices. They're specialized. You don't run into a kitchen appliance overload problem, I don't think, because each one is individually simple. If they're really simple, I think simple and specialized is going to win over the present PC model that it can do everything but it's too d-- complicated to use.
Q. What are some of the things being developed that you think are most promising for the next 10 years?
Horn: To me, natural language is one of the biggest and most exciting advances you're going to see. As it is, the technology's getting good enough so that you can dictate now reasonably well and the computer will convert what you say into (text) characters, sort of standard dictation: I talk, it types. But over the next few years it's really going to be much more: I talk, it understands. And that means it will do within various contexts, it'll not just type but understand what you say and be able to provide you with various levels of support. And I think that's going to be a real change in the way people interact with technology . . . much more ease of use, much more the feeling that you're getting assistance from the technology rather than having to deal with it.
Q. You have talked about a personal-area network. How would it work, and what kinds of information might be exchanged when, say, you and I met in person?
Horn: Let's imagine you could go to your car door. You don't have a key because when you touch the handle it knows it's you and it unlocks the door. And you can start it by pushing a button. Again, you don't need a key because it knows who's trying to start it. Or you pick up the telephone and the telephone knows who to bill the call to. So those are personal-area-network-style applications -- person to machine. Ultimately, I think you could well see more person-to-person utilization of personal-area networks.
If you extend the thought just a little bit and add global positioning, then you could imagine, for example, personalized advertisement when you look in the store at a certain thing because it'll know where you are because it's interacting with you on an ongoing basis. So think of this as part of a whole set of technologies that are going to allow for the information technology world, the Internet, to know where you are and feed you local and personal information based on where you are.
Then, the next step in that is, well, it knows where you are and another person knows where you are and then you can interact person to person. That might be, again, by shaking hands. Or, if you extend the thought again just slightly, (it) can communicate, tell you who's in the room, remind you that you've interacted with that person before. I'm having dating analogies come into my mind, I'm sure, but you can imagine how the potential value of being able to find out that someone that you really need to see and you haven't seen in 10 years happens to be in the same crowded room that you're in.