Copyright 1992 by John Perry Barlow. All rights reserved. NeXT in the Real World Part II of the MicroTimes interview with Steve Jobs By John Perry Barlow Everything is changing all the time. At that point in early December when I had the following conversation with Steve Jobs (the first half of which was published in MicroTimes last month), the world around NeXT looked different to me in a lot of subtle ways. For one thing, I was still somewhat unpersuaded that their bold new marketing plan-- selling NeXTstep as an environment for developing custom applications--wasn't yet another way to shovel smoke. After all, NeXT has had, depending on how you count 'em, at least five bold new marketing plans over its existence, each touted with equal fervor, and each more ineffectual than the last. Then there was Jobs himself, who, in his actions leading up to this conversation, seemed as compulsively elusive and manipulative as he is sometimes reputed to be. But I'm writing these words toward the end of January, in the immediate afterglow of NeXTworld Expo. During the course of this event, a couple of things became clear to me. NeXT is going to make it, and possibly make it big. Their perceived new market actually does exist and, more to the point, it knows that they exist. NeXTworld was crawling with sober, suited individuals who gave no appearance of being there on some kind of cyber-cultural lark. In evidence were all the warning signs of MIS. While my own personal enthusiasm for their arrival was restrained by my blind prejudice against suits, they're like the 7th Cavalry as far as NeXT is concerned. Furthermore, assuming that their port of NeXTstep to the Intel architecture really does work as seamlessly as it appears to, they may be able to grab a commanding lead in the Operating System of the Future contest before the systems heretofore regarded as the Real Contenders-- Big Pink, Windows NT, and Solaris--have condensed out of their current vapors. In general, NeXTworld, and the product announcements it showcased, made NeXT seem like a much more credible and mature company than the kind of company which might, for example, actually expect to sell computers without floppy disk drives. There is, once again, a New Steve Jobs. Over the course of his career, this shape-changer has gone through more metamorphic revisions than one generally achieves outside the insect kingdom, but he has finally entered one in which he is starting to look comfortable. His two-and-a-half-hour presentation at NeXTworld still contained some auto-Barnumizing. At one point, the respected Mac pundit in the seat next to mine muttered, "If this guy had been born in Alabama, he would have been bigger than Swaggart." And I didn't have to ask him what he meant by that. And yet... And yet what amounted to the third NeXT product introduction contained little of the Wagnernian theatricality which caused the first two to feel less like a computer intro than Albert Speer's staging for the Rally at Nuremburg. This time, there was a lot more steak than sizzle. Jobs was calm, relaxed, conversational. He didn't appear to be hiding much behind the glitz of his presentation. The products themselves were certainly meat and potatoes. There was version 3.0 of NeXTstep, which contained advanced abilities for transparent group endeavor over EtherTalk and Novell networks, a highly adaptable database tool kit, 3-D modeling support based on Pixar's Renderman, and system integrated ISDN so that one's NeXT at home could feel part of the office network, even though connected to it by nothing but standard copper phone cabling. There were many other incremental improvements too small to detail here, which will nonetheless solidly enhance the already well-integrated NeXT working environment. He also introduced a great and inexpensive ($3500) color printer using PostScript 2 and Pantone colors, a 40% increase in hardware performance (with a 33 MHz '040 CPU) at no additional cost, and a new CD-ROM drive to be used, among other purposes, for loading the new system software. All pretty prosaic stuff by NeXT standards, but, well, grown up. Which is pretty much how this new Steve Jobs comes across. He is, after all, a brand new father, a condition which usually stimulates the maturity glands. On the second day of NeXTworld, Jobs spent the afternoon wandering the show floor, openly available to anyone who sought his ordinarily hard-gained company. He was amiable and engaged. And obviously very happy with the way things were going. Though this conversation took place much earlier, some of that genuine confidence is in evidence here. With the greatest reluctance, I recently ordered a PowerBook 170. I hated to do that, since my Macintosh has been reduced to being sort of a dedicated QuickDex server, but my NeXT... But you need a portable. Exactly. Can you give me any sense of where we stand there? I can. Stay tuned. We're working on some stuff in the lab. There are a few directions we're going in. What most of our customers tell us is that they don't need it to run on batteries, they just need it to be really transportable. Just transportable. That's how I feel. That's pretty much our philosophy right now. There are a lot of things that are the enemy of running on batteries, like a big screen. Like a bigger hard disk that uses more power. Like a faster CPU that uses more power. Things like that. So what our customers are telling us is that they want to be mobile, but they can live tethered to a 110-volt outlet. Let me shift to connectivity and standards for connectivity. I'm concerned that the market doesn't realize the extent to which you're already meeting standards. I agree. I mean this looks like a good document [referring to a PR document about NeXT and standards], but I get skeptical questions about this every day. For example, I hear things like: "Objective C is a dead language, regardless of its merits. NeXT has bet on Objective C." Or they'll say, "NeXT Berkeley UNIX is not exactly parallel to Sun Berkeley UNIX." I haven't run across anything that won't compile yet, but I hear these things... Is this all just idle slander and gossip, or is there something to it? The two biggest things you hear in the industry are "which chip are you using?" and "which version of UNIX do you have?" Then there are minor ones beyond that. And my view is that these are the battlegrounds of five years ago. It really mattered what chip you used five years ago. It really mattered what version of UNIX you had five years ago. Those things are not the battlegrounds of today. When you write an application for NeXT, five percent of what you see is UNIX. Ninety-five percent of what you see is NeXTstep. Opeal processors right now. What matters is, what does the application developer see? What's your application development environment? This is my point of view. You couldn't even tell me what operating system Macintosh used, right? So I guess to me the people that are worried about these things... well, they don't seem to be the customers, because we don't run into these problems with customers at all. And if we do, we have them talk to a technical person for a few hours and their concerns evaporate. It seems to be people on the sidelines who have these five-year-old ideas stuck in their brain. Their light bulb hasn't gone on, in terms of what's really important today. I think you may be fundamentally right about the irrelevance of these distinctions, but as you have seen many times, there's an amazing willingness in the computer industry to adhere to blind fashion. But see, that's not who our customers are going to be. Our customers are going to be early adopter corporate and government workstation customers. And we shouldn't spend our time trying to persuade every last person that we're using the right chip or that we're using the right version of UNIX. What we should do is go with those early adopter customers who can grow our company to a half-a-billion to a billion-dollar company, and let our success demonstrate that those issues don't matter any more. That's our strategy. Our strategy shouldn't be to convince the world. It should be to convince the one percent of the world that can grow us to a half-a-billion to a billion-dollar company. And let our success illuminate our strategy for the rest of them. Corporate customers started buying workstations two or three years ago. They started buying Suns, as did the government. And they bought them to write their mission-critical custom apps on. You don't buy a Sun to run productivity software on, right? Wall Street was one of the first early adopters And we're walking in right now, and ninety percent of the time we compete against Sun we're winning. Why? Because what do you buy a Sun for? To develop custom apps on. NeXT can do it five to ten times better. So we're winning right and left. Not in the scientific and engineering world--we don't know anything about that. But in the commercial and government arena, we're winning right and left. I worry about things like this, though. [Waving a copy of Networking Today at him.] There's an article in here called something like "Will UNIX ever succeed on the desktop?" They are basically telling their readers in corporate networking to wait for whatever comes out of Big Pink and Solaris, because that's what you've got to have. And they're touting all the virtues that will be derived from Solaris and Big Pink someday, and not saying anything about NeXTstep, which is shipping with them now. Give me that magazine. We'll call them up. [Grabs magazine and reads the whole thing in about 15 seconds.] I was alarmed by this. Because this is really your market, in many respects, and they don't know that you exist. Well, this magazine doesn't know that we exist. But our market is starting to hear about us. But we'll call these guys up. This is wonderful. We'll fix it. But to get back to the chip issues, I was on a plane recently with one of Intel's main men. He said, "You ever see Steve Jobs?" I told him about this interview coming up and he said, "Well, I would be appreciative if he would come around and pay us a visit. We've got something that he really ought to be using." What is that? A chip under development. The 586. Yeah. Sure. Intel's over here once every month talking to us about that. We knew about the 586 a long time ago. Well, I knew you knew about it, but apparently he doesn't realize there's been this kind of conversation. [Intel CEO] Andy Grove and I have known each other for fifteen years. And we talk a lot. We're very aware of that stuff. I think a lot of people are concerned that CISC chips in general have reached a point where they can't go much further. Well, the 586 is a CISC chip, for the most part. It's sort of a combination. So's the 040. They're all combinations. The RISC/CISC stuff is ninety percent marketing and ten percent reality. I think we evaluate things on one simple benchmark--how fast is it? I don't care if it's a little hamster in there on a treadmill--how fast is it? So far you're running faster than most garden-variety SPARCstations. Well, and you'll see us very shortly catch up with SPARCstation 2's. Very shortly. So the 040 is a fine chip. It runs as fast as the SPARC stuff. Yes, there are faster chips, and you'll see us do things with other processors in the future. But they'll all run NeXTstep. How do you deal with something like the Indigo? I gather that that's now a sort of orthogonal consideration. We never see the Indigo out competing in the real world. We've never seen the Indigo in one corporate account, competing. I think the Indigo's selling to [SGI's] traditional base. Right. The SIGGRAPH types. The 3D modeling types. Yeah. Which is another question I've got. You've got a relationship with Pixar- - Right. And I keep thinking that there's an obvious tie-in at some point along the line. Not just for professional 3D graphics. It seems like the interfaces themselves are getting increasingly three-dimensional. I agree with you. Our NeXTstep release 3.0--which is going to ship in March or April--is going to come bundled with several things. One is our DBkit, our database kit. Another is Novell client support, so you can hook up to any Novell network and automatically map that right into the UNIX filesystem. Ethertalk, which we've written ourselves, so you can just hook it right up to an Ethertalk network and we map that all into UNIX-- nobody's ever done these things before, by the way. That all comes in the box. Another thing that comes in the box is RenderMan. Photorealistic Renderman is bundled. With an interface? Well, let me keep going. Photorealistic Renderman is bundled, so that you can use it to do photorealistic rendering, the same thing you pay a lot of money for on other platforms. We've got all the printing and everything working with it, so you don't have to think about that, which is something nobody else has done. But we did something else that's quite nice. As you recall, when we did the Mac we had QuickDraw for the screen and PostScript for the printer, and that was a real problem. It continues to be. It continues to be. So when we did NeXT, we adopted a unified imaging model for 2D geometric graphics in PostScript. That solved an enormous number of problems. One of the biggest problems it solved was that it saved developers months of time, because they don't have to translate between one and the other and read these two big thick manuals to get something to print. Well, the same sort of dichotomy exists in the 3D marketplace. You can have PHIGS for interactive graphics, or GL. But these are very primitive standards. You should be able to do better than those now, but in the future it gets worse. Because ten years from now we'll have the CPU performance to run photorealistic Renderman in real time. So during that ten years we want to incrementally get better. Well, PHIGS and GL and that stuff will not get better. The architectures are not going to get much better. So you're going to have to throw them out and get better 3D packages a few times during the '90s, and every time you throw the package out you throw all your apps out. So we thought, what would be the right solution? The right solution would be to have an interactive Renderman, and then as the '90s progressed, leave a few more of the switches on that can run in real time, until eventually all of the switches are on. So we went to Pixar just like we went to Adobe, and we worked with them to create an interactive Renderman, and it's bundled with 3.0. It's wireframe on the 040s, and it gets more interesting on other products. That's not modelers. We're not going to write a modeler, because we want don't want to compete with the third parties. But all the underpinnings are there to do some wonderful things in 3D in the next few years. Let me talk to you a little about multimedia. I just interviewed John Sculley. He talked about Kaleida and their intention to develop a set of authoring and architectural standards that anybody could plug into. Uh-huh. But see, standards are actually developed. Well, I understand that... Like fifty million CD players shipped. That's a pretty compelling standard. MPEG is a pretty compelling standard. But there are also 24 different formats that a multimedia author could choose to publish in at the moment, which divides up a pretty small market. Oh, I agree with that. But the actual data representations are chosen more by the consumer products industry than they are by the computer industry. MPEG is an example. It's going to be a pretty compelling video format. This gets us to something I wanted to ask you about-- I'd still like to see that compression chip on the NeXTdimension board. Well, as you know, C-Cube conked out. We now have the boards-- the hardware boards are done, and they've given them to our software group, and as soon as they get some software up on them, you'll see it as a product. But we have some much more exciting stuff than that in the lab. We'll do all of it. I actually think the QuickTime stuff is not a bad idea, but the implementation is very poor. It's trying to do too much in software on a machine that doesn't have the juice to do it. Or they took the wrong software algorithmic approach on it entirely. That may be. Well, would you would be willing to support something like QuickTime? QuickTime will never reach to the kind of quality that people are going to look for. The algorithms of implementation are fundamentally flawed. The concept of having an open system like that is something we support entirely--as a matter of fact many people have come to us asking us to license some of our existing multimedia technology, and in almost every case we do. Sun's sound compression formats are the same as our current ones, although we're rolling in a new one for 3.0. We found a way to compress CD sound to data rates that are Codec data rate. Imperceptibly. You can take stereo CD sound, compress it down to Codec data rates, and you can't tell the difference. It's a wonderful thing. It's never been done before, all in software. We unified all our sound formats into one format. We don't need all those different quality level ones now, because our best quality now has the data rate of our lowest quality one. So we're unifying all that into one format, which is going to simplify our lives enormously. We'll be willing to license that, as we have a lot of others. I think you're going to see video over the Net fairly soon, and it's probably going to be in a QuickTime format, to start out with. I mean, if all you're trying to do is get a video box to pop out of your e-mail, you're willing to suffer some crudity. Unless you can have something quite a bit better. Yeah. If you can have something quite a bit better that you can actually communicate with somebody else on... Well, right now QuickTime only runs on the Mac, so it's the same problem. Right. But there are a lot of Macs out there. Sure there are. Of course, if your vision is achieved, there'll be a lot of NeXTs out there. We're going to be willing to license our stuff in the same way, and my guess is it'll be substantially higher quality. Well, it is now, for whatever that's worth . What I can do with a video camera and a NeXTdimension board, even without the compression chip, is pretty extraordinary. Just wait. This relates to a problem I have with NeXTmail. I love NeXTmail, but I constantly feel like I've got a great operatic voice in the country of the deaf. I want to be able to communicate with everybody in NeXTmail, but I can't. For example, I have no ability to communicate with people on Sun workstations... Well, you do with text. I do with text, but I want more. But see, Sun can't play the music. And Sun can't open the Improv files. So what else do you hope to do? If we could change the file format around, let's say we could instantly change it to whatever you want, what would you do with the information once you got it to your Sun? Well, I would hope that there would be some joint effort between companies like Sun and NeXT, to find a lingua franca. We are open to that. Perhaps something like Carousel, Adobe's new document description language... Oh, of course we will support that. That's just a subset of PostScript, so we'll be the first company to support that, because we already do. We've held out the olive branch to Sun and said we'd love to work together with them on this stuff. So go ask them. You're a brand-new father. Yeah. When that happened to me, it redefined my whole sense of purpose in life. What is your mission now, in the broad sense of the word? What is my mission? What kind of world do you want to help create for your kid? I have felt for most of my adult life that I'm a tool-builder. The things I have chosen to build are not things that can be done by a sole artist. That's a wonderful thing--many times I am envious of people that have chosen those paths, because one can be a perfectionist and have complete control of something yourself. But what I've chosen to do requires the efforts of hundred, if not thousands, of people. So I consider myself the tool-builder that works with a large team of people to build tools. I've been through a few generations of tools in my adult life--the Apple II, the Macintosh, and some in between that didn't work out so well. This is our third generation of tool. And unlike a lot of people who think the computer industry's maturing, I think it's in its infancy. I think there are technological breakthroughs that happen once every ten years, maybe. Sometimes a little more frequent, but not too much. And that those technological breakthroughs have the force to reshape the tools that we build, and to reshape the industry along with them, as certain companies pay attention to them earlier and certain companies wake up fairly late. So my mission right now is to try to help seize these new technological breakthroughs that we've had, and to steer them right, into some wonderful tools for people to use. Sort of the third generation of tools out there, the first being the character mode PCs, the Apple II and then the IBM PC. The second being the graphical, pretty much standalone computers, like Macintosh and now Windows. And the third being these highly networked environments where expressing your desire in a custom app is going to be a lot less painful than it's ever been before. And these things interoperating together. I don't even fully understand it yet, just like we didn't fully understand Mac in 1984. But you get the feeling, and you recognize it after it's happened a few times, that you're on to something really big. And that's how I feel right now. So our goal is to try to fashion this stuff into the most wonderful tools we can. And then to listen to the people using them--maybe more than we have in the past--to let them help us. That's what we're trying to do. We'll see where it goes.z .