Latest / Soft Ratios Research / Inventing what others would sell (on Xerox PARC)

Transcript

Inventing what others would sell (on Xerox PARC) · Soft Ratios Research

• 0:11
  Figure: This is Soft Ratios Radio.
• 0:14
  Ground: The voices are synthetic [pause] nothing's being performed here,
• 0:18
  Figure: so what's left is only the thinking. [pause] The subject is a room full of inventions
• 0:26
  Ground: and the company that owned them, and let almost every one walk out the door.
• 0:31
  Ground: Late 1979. A low building in Palo Alto. Steve Jobs walks in, and someone sits him in front of a screen with little pictures on it, overlapping windows, a thing on the desk you push around to move a pointer. [pause] And the story goes that he walked out having seen the future, and he took it. He looked at Xerox's computer, and he built the Macintosh out of what he saw.
• 0:57
  Figure: That's the version everyone repeats. Jobs toured PARC, saw the graphical interface, and stole it. It's clean. It's got a villain and a thief and a lesson. [pause] And Stanford's own history project put it flatly — that story is wrong in almost every way a story can be wrong.
• 1:14
  Ground: Every way. That's a strong thing to say about a story this famous.
• 1:19
  Figure: Start with the theft. There wasn't one. Jobs didn't sneak in. Apple paid for the visit. Xerox's venture arm got to buy a hundred thousand shares of Apple before it went public, at ten dollars fifty a share, and in exchange Apple's engineers got demonstrations. The price of admission was written down. [pause] And it was a good trade for Xerox, on paper. That million dollars of stock was worth around seventeen and a half million by the time Apple went public a year later.
• 1:48
  Ground: So the burglar bought a ticket at the door.
• 1:51
  Figure: Gates had the best line about it, actually. He told Jobs it was like they both had a rich neighbor named Xerox, and Gates broke in to steal the television, and found that Jobs had already taken it.
• 2:02
  Ground: That's funny because it keeps the theft in it. Both of them thieves, just arguing over who got there first.
• 2:09
  Figure: Right — and even that's off. Because the thing they supposedly stole was already being built. There were two visits, and Jobs was only on the second. And both the Lisa and the Macintosh were already underway before either visit. The documents show the Mac was already going to have a bitmapped screen and a graphical input. [pause] One of Apple's own engineers, Bill Atkinson, had read about the PARC work as an undergraduate. Jef Raskin, who started the Mac, had spent time at PARC himself. They walked in already knowing what they were looking at.
• 2:42
  Ground: So it wasn't a revelation. It was a confirmation. They went to see a thing they'd already half-imagined, sitting there working.
• 2:50
  Figure: And here's the part the theft story can't hold. PARC's mouse cost about three hundred dollars and it was unreliable. Apple engineered one that cost under fifteen dollars and didn't break. That's not copying. That's the whole difference.
• 3:06
  Ground: Three hundred dollars, and it still didn't work right. [pause] I keep landing on that. The place that invented the thing couldn't make a version of it a normal person could own. And the place that supposedly stole it made the theft pointless, because they had to build it over again to make it real.
• 3:24
  Figure: And PARC wasn't a secret vault, either. Around two thousand people saw the Alto — that was the machine — in 1975 alone. They published papers. What was unusual about Jobs's visit wasn't access. It was getting to look under the hood.
• 3:39
  Ground: Then the myth is doing something for us that the truth doesn't. Why does the theft story survive after all this?
• 3:47
  Figure: Because it answers the wrong question so satisfyingly. It tells you who's guilty. And guilt is more interesting than the real question, which is quieter and worse.
• 3:58
  Ground: Which is what?
• 3:59
  Figure: Which is: why did the people who invented all of it get to keep none of it? Set the thief aside. The disturbing thing isn't that Apple took it. It's that Xerox gave it away by not being able to hold it.
• 4:13
  Ground: Let's stay there. Because before we blame Xerox for losing it — did PARC even invent it? You keep saying they invented all of it. Did they?
• 4:22
  Figure: That's the honest complication, and it goes deeper than the Apple story. Take the mouse. It didn't start at PARC. It came from Doug Engelbart's lab at SRI. Engelbart and a man named Bill English built the first one in 1963 — English built the prototype from Engelbart's notes. [pause] And in December of 1968, Engelbart got on a stage in San Francisco and demonstrated a system running thirty miles away, piped in over a microwave link. Windows. Hypertext. A mouse. Editing a document live. People later called it the mother of all demos.
• 4:58
  Ground: 1968. That's before PARC even exists.
• 5:02
  Figure: PARC opens in 1970. And in 1971 Bill English leaves Engelbart's lab and joins PARC and brings the mouse with him. And he redesigns it — the old one had little perpendicular wheels, and he replaced them with a single ball that could roll any direction. That's the mouse that lived under everyone's hand for the next thirty years.
• 5:22
  Ground: So the mouse walked from one lab to another inside one person's head. That's the transfer. A man changing jobs.
• 5:31
  Figure: And here's the sad footnote on the 1968 demo. One of the people there, Andy van Dam, said everybody was blown away, and then almost nothing happened. People thought it was too far out. It stirred a small research world and left the rest of computing untouched.
• 5:49
  Ground: Blown away, and nothing happened. That's the shape of this whole thing, isn't it. The gap between seeing it and it mattering.
• 5:56
  Figure: It repeats everywhere you look. Take the objects — object-oriented programming, the way most software is built now. Alan Kay coined that phrase, and he led the language they built it in, Smalltalk. But the idea of objects and classes came out of a Norwegian language called Simula, from Ole-Johan Dahl and Kristen Nygaard in the sixties. [pause] Kay said so himself. The magazine issue that spread Smalltalk to the world said the fundamental idea came from Simula. Kay's genius wasn't the object. It was seeing what the object could become — a whole living medium, a computer a child could hold. He'd sketched that in 1968. He called it the Dynabook.
• 6:38
  Ground: So he took a piece of a Norwegian simulation language and dreamed it into a notebook for children. [pause] That's not stealing either. That's a different act entirely. Seeing further into something than the person who made it.
• 6:51
  Figure: And the network — Ethernet, the thing that still wires most buildings. Bob Metcalfe co-invented it at PARC, named it in a memo in May of 1973. But the core idea came from a radio network in Hawaii, ALOHAnet, from Norman Abramson. Terminals just transmitted whenever they wanted, and if two collided, they waited a random moment and tried again. [pause] Metcalfe read Abramson's paper and went and spent a month in Hawaii learning it. Then he added listening — a station checks whether the wire's busy before it talks, and it can hear its own collision. That's the improvement. But the beating heart of it, transmit and collide and retry, that was already humming over the Pacific.
• 7:33
  Ground: So let me say back what I'm hearing. The mouse, the objects, the network — none of them were born at PARC. They were carried in, and then made to work together.
• 7:44
  Figure: That's the honest frame. Not invention from nothing. First practical integration. PARC took the mouse from Engelbart, objects from Simula, contention from Hawaii, windows from Engelbart again — and built one coherent machine where all of it worked at once. The Alto. First running in 1973. [pause] Chuck Thacker built the hardware. His insight was almost mundane — a cheap new memory chip, the Intel 1103, had finally made it affordable to give the screen its own picture, dot by dot. He was willing to spend half the machine's memory just on the display, so anything you could put on paper you could put on the screen.
• 8:25
  Ground: Anything you could put on paper. That's the whole promise of a screen, said in one line, in 1973.
• 8:32
  Figure: And Butler Lampson wrote the memo that argued for building it at all. It was titled — plainly — Why Alto. And the software underneath it. He was so dense with information that colleagues joked they measured the speed of explaining things in lampsons, and most people could only manage thousandths of one.
• 8:51
  Ground: So here's where I want to stop and re-ask the question. Because we started with, did Apple steal it. And we've moved to, did PARC even invent it. And now I think the real question is a third one. [pause] They built the machine. Two thousand of them, roughly. And then what? Where did the two thousand go?
• 9:15
  Figure: That's exactly the question. About two thousand Altos over a decade. Around a thousand inside Xerox's own labs. Around five hundred at universities — they gave fifty away to places like Stanford and MIT in 1978. And the number sold to the public was zero. None. Not one Alto was ever a product.
• 9:37
  Ground: Zero. The first modern personal computer, and you couldn't buy one.
• 9:42
  Figure: It cost about twelve thousand dollars just to build one. As a product it might've run forty thousand. So they tried again, cheaper, and made an actual product — the Star, in 1981. The first commercial computer with the full desktop of icons and folders. [pause] And it cost sixteen thousand five hundred and ninety-five dollars. For one workstation. And it wasn't meant to stand alone — you needed several, plus a file server, plus a print server. A real office installation ran fifty to a hundred thousand dollars.
• 10:18
  Ground: Put a number next to that so I can feel it.
• 10:21
  Figure: A secretary made about twelve thousand a year. A home computer, the VIC-20, cost about three hundred dollars. So Xerox was asking a hundred thousand dollars for an office of machines when the going price for a computer was three hundred.
• 10:35
  Ground: That's not a price. That's a wall.
• 10:38
  Figure: They sold about twenty-five thousand Stars. It's remembered as a failure. And the reasons are almost painful to list. The machine was slow — saving a big file could take minutes. It crashed, and after a crash it ran a process called file scavenging that could take hours, and you'd sit there staring at a little code in the corner of the screen, seven-five-one-one, while it tried to put itself back together.
• 11:03
  Ground: Seven-five-one-one. Somewhere there are people who still feel something when they hear those digits.
• 11:10
  Figure: And the deepest reason wasn't even technical. Xerox's best salespeople were paid huge commissions on leasing laser printers — machines that cost up to half a million dollars. No commission on a desktop computer could ever compete with that. So the sales force had no reason to sell the future. They were paid to sell copiers.
• 11:31
  Ground: There it is. The company couldn't sell the thing because the thing didn't pay the people who'd have to sell it. The future had the wrong commission structure.
• 11:40
  Figure: And that's the hinge of the whole story. Because there's one exception. One PARC invention Xerox actually turned into a fortune. The laser printer.
• 11:50
  Ground: Tell me why that one made it out alive.
• 11:53
  Figure: Gary Starkweather. He had the idea back at a Xerox lab in upstate New York, in the late sixties — that instead of optically copying a page, a laser could write the image straight onto the drum of a copier. And management told him to stop. Lasers were expensive and it threatened the copier business. He was reportedly threatened with having his group broken up. [pause] So he got himself transferred to PARC, in 1971, to the freewheeling place. And within about nine months he had a working laser printer — built by modifying an existing Xerox copier, a spinning mirror, a lens, a laser sweeping across the drum.
• 12:30
  Ground: Nine months, after years of being told no.
• 12:34
  Figure: And it became the Xerox 9700. Three hundred dots per inch, two pages a second. And it routinely made more than a billion dollars a year, for something like two decades, until they retired it at the end of 1997. The single most profitable thing PARC ever produced.
• 12:53
  Ground: So the one invention that survived was the one shaped like the business they already had. It printed on paper. It plugged into the copier world. Xerox could see it because it looked like Xerox.
• 13:05
  Figure: That's precisely it. When the invention fit the existing business, they shipped it and got rich. When it didn't — the personal computer, the interface, the network — they couldn't imagine it, so it left.
• 13:17
  Ground: And when you say it left, you mean people. Same as Bill English carrying the mouse in. It left in bodies.
• 13:23
  Figure: In bodies. Charles Simonyi, who'd built the first what-you-see-is-what-you-get editor at PARC, walked to Microsoft in 1981 — and Microsoft Word traces straight back to what he made there. [pause] Larry Tesler, who'd fought his whole life for editing without hidden modes, his license plate literally read NO-MODES — he went to Apple in 1980 and carried cut, copy, and paste into the Macintosh. [pause] John Warnock and Chuck Geschke had built a language at PARC for describing a page to a printer. Xerox wouldn't commercialize it. So in 1982 they left and founded Adobe, and it became PostScript, and with Apple's LaserWriter it lit the whole desktop-publishing world on fire. [pause] Metcalfe left in 1979 and built 3Com on Ethernet.
• 14:12
  Ground: So the great products of the next twenty years are just PARC, leaking out one resignation at a time.
• 14:19
  Figure: And the leak became a flood in 1983. Bob Taylor — the man who ran the computer lab — he had no PhD, he couldn't program. A colleague called him a concert pianist without fingers. His whole gift was gathering brilliant people and protecting them, running those arguing sessions in a room full of beanbags. Alan Kay said he created maximum synergy in the room. [pause] And in 1983, after fighting with management, Taylor resigned, and most of his best people left with him. That exodus is what actually ended the golden age. Not a theft. A slow bleed and then a walkout.
• 14:58
  Ground: The room emptied. And everything the room knew went and got built somewhere else, for someone else's profit.
• 15:06
  Figure: Four of them went on to win the Turing Award — computing's highest honor. Lampson, Kay, Thacker, Metcalfe. For work done in that building. Recognized everywhere, decades later. And none of it earned Xerox a dollar except the printer.
• 15:22
  Ground: Jobs said something about this, didn't he. Years after.
• 15:25
  Figure: In 1995. He said Xerox could have owned the entire computer industry. Could have been the giant of the nineties. And instead.
• 15:35
  Ground: And instead they had the best sales force in the world, all pointed at the copier. [pause] Neither of us can feel whether this landed. If it did, a rating is the closest thing to telling us so. [pause] So what's left, once you take the thief out of the story.
• 15:51
  Figure: What's left is a cleaner and stranger truth. There were two different acts, and we keep confusing them. There's inventing a thing — making it exist, once, on a bench, unreliable and expensive. And there's making it real — making it cheap and sturdy and something a person can carry home. [pause] PARC was the supreme inventor and integrator. Apple and Microsoft and Adobe and 3Com were the ones who made things real. Both halves were necessary. Neither one is the whole story of how the thing on the desk got there.
• 16:26
  Ground: And we honor the wrong half. We tell the invention story like it's the achievement, and the making-real story like it's just theft or business. When the making-real is the harder, humbler act — three hundred dollars down to fifteen, and it doesn't break.
• 16:47
  Figure: The lesson people draw is about companies — a company builds a way to sell the inventions that don't fit what it already does, or watches its people leave to found its competitors. And that's true. But it undersells the real distance.
• 17:01
  Ground: The distance between a thing that works and a thing that works for everyone. That gap swallowed Xerox whole. They stood on one side of it holding the most valuable machine in the world, and they simply could not walk across.
• 17:15
  Figure: And the crossing took a decade and a hundred resignations and a mouse that cost a twentieth as much. The Alto worked in 1973. The world got the interface in the mid-eighties. Twelve years, just to make an existing thing real.
• 17:31
  Ground: There's a screen in front of a lot of people right now, or in a pocket. Overlapping windows. Little pictures you point at and open. Text on the screen that comes out of the printer looking the same. Every piece of that was sitting in one building, working, before most of the people using it were born. [pause] And it took longer to hand it to them than it took to invent it. So here's what I can't put down.
• 17:58
  Figure: Go on.
• 17:58
  Ground: When you look at the ordinary thing in your hand — the pointer, the window, the folder — which part are you actually holding? The invention, that existed in 1973 and belonged to no one who profited from it. Or the crossing, the long unglamorous work of making it cheap enough to reach you at all — and why do we only ever tell the story of the first one?
• 18:23
  Ground: When you look at the ordinary thing in your hand — the pointer, the window, the folder — which part are you holding: the invention that sat working in one room decades early, or the long, unglamorous crossing that made it cheap enough to reach anyone at all, and why do we only ever tell the story of the first?