Why Are We Gathered Here

Opening remarks

In this session, Dr. Peter Lucas talks about our roadmap into the future and what makes MAYA special. Pete is one of MAYA’s founders and a visionary whose passion is to make true pervasive computing a reality. The Visage project is at the center of his work.

Outline

It's Ugly. Deeply and Unspeakably Ugly

I'll indulge in a little bit of philosophy as to the underlying motivation for why we’re here and what the Visage project is for, and why we care so much about it. And I’ll try to give a little bit of a roadmap as to where all the pieces fit together both now and in the future. Then, (the part I’ll certainly fail at) I’ll leave as much time as possible for discussion.

I need four volunteers. (Everyone looks at one another and laughs.)

(Nick, Slim, Seye, and Francine finally raise their hands.)

(Pointing to Nick) You are 1A. (Pointing to Slim) You are 1B. (Pointing to Seye) You are 2A. (Pointing to Francine) You are 2B.

The Ones come up and get your plumber's putty. (It’s mostly nontoxic!) The Twos come and get your toys (Ziplocs with magnets and links). The Ones (sorry, I should have done this before) come and get your toothpicks.

While I talk, you guys have construction projects.

I want the A's to build me a tetrahedron (a three-sided pyramid). I want the B's to build me a cube.

Go at it, and I'm going to talk while you're working.

So here’s my first secret. If you want to seem smart, you find some problem that hasn’t really been solved yet, but is on the verge of being solved. Then you find some other problem that has been solved that’s as close as possible to the one you want to solve. Then you go back and read as much literature as you can [written] at the time when the solved problem was being solved, and you put yourself in the mindset of the people who were struggling with the then-unsolved problem.

That’s what I did when I cofounded MAYA. Most of you know the canon of five books that you should read if you want to understand what MAYA is all about. The reason I’m so fond of them is that these are the most important five books that I read about industrial design when we were trying to figure out what we were then calling “post-industrial design.” They're:

Never Leave Well Enough Alone by Raymond Loewy
Design This Day by Walter Dorwin Teague
Horizons by Norman Bel Geddes
Designing for People by Henry Dreyfuss
Then, Twentieth Century Limited by Jeffrey Meikle, which is kind of a history of those days -- a wrap up of the other four.

Every good idea I contributed in the early days of MAYA I shamelessly stole from those books. It was highly analogous. It was the endgame of the industrial revolution. The world was filling up with powerful, world-changing, but unspeakably ugly machines – exceedingly unpleasant for anybody to be around. Even though they generated great amounts of wealth, and they crushed people both figuratively and literally. In the 30s, industrial design took a lot of raw capabilities and turned them into a world that is worth living in––that ordinary people would appreciate––and made people’s lives better. That’s really what industrial design at its best was all about.

The idea of MAYA was really simple: We were going to do the same thing for post-industrial design.

There have been three fundamental revolutions in the world. Two thousand years ago there was the agricultural revolution. About 10-200 years ago, there was the industrial revolution. And as soon as the industrial revolution ended, we started the information revolution.

And I really believe it’s only those three––that there’s nothing else as fundamental with respect to changing the way real people live.

And we’re right in the middle of the third one. It’s going a lot faster along some dimensions, but not others. Many of the problems are exactly the same.

You don't see it because you’re like fish in the water. You don’t see how unspeakably ugly computers are because they’re as good as they’ve ever been. But exactly in the sense of ugly I meant before – deeply ugly – I believe that the technology at the center of our lives, that is fundamentally changing the world, is still unspeakably ugly. And that’s what we were very explicit about when we started MAYA.

So it's not about invention; it's about perfection -- taking things that other people have invented. That's what design is. The Media Lab’s motto used to be “inventing the future,” and we toyed with the idea of "perfecting the future" as to what MAYA was about. But then we chose "taming complexity"––which was better. If you look at the technological innovations that we do here, there's almost nothing that you can’t point at a precedent for. Did we invent modular decomposition? What we’re doing is being competent at a lot of things – stepping back from the milieu and trying to create a world to harness this stuff and turn it into a world worth living in.

The Pitfalls of True Corporate Innovation

So, at any rate, the reason I was off on that rant is that I'm trying to do something else now. As most of you know, I’m in the early stages of trying to figure out how to found the Civium Foundation. The endgame of my career I hope will be helping the Information Space––which we could do a whole other talk about. I'm not going to talk about that. But that's what I'm trying to do, so I'm spending a lot of time following my own advice. I’m trying to start a non-profit for various reasons. It has a lot in common with the founding of universities. How do you pull off the modern research university?

(Pulls out a book and leafs through it)

So Robert Brookings was a rich industrialist. He and his partner, Samuel Couples, made a lot of money in the warehousing business in St. Louis in the 19th century. They were really successful. Here’s a story that I like (reads from the book):

. "Brookings later recalled, ‘I said to Mr. Couples one day that I could never justify to myself the blood I had sweat over Couples Station (their big warehouse) unless it accomplished some great good in the world, and if he had no objection to the University [Washington University] as a partner, which was small and struggling, I would like to give my half interest in the property to the University’s general endowment fund. Without a moment’s hesitation, he replied, ‘ Robert, this is a good scheme. If you have no objection, I would like to join you and give my half to the same University . . .’ ”

The reason I bring it up is that I’ve been spending an awful lot of time thinking about how you innovate—how you actually cause change to happen. Can you do it in the context of a large organization? Can a big corporation innovate? I’ve kind of come to the conclusion that it really can’t.

Buckminster Fuller used to say that you can make dollars or you can make sense – but you can’t make both. He didn’t mean that you can’t both be rich and do good. What he meant is that you can’t become rich while doing good. I’m not being anti-corporation. I don’t believe that large corporations are bad. I just believe that they don’t innovate. What they are good at is taking something that we know how to do and amplifying it immensely. Jeff and I were talking about this. All right, we have cell phones down now. Now we’re going to make two large corporations ––a slight exaggeration–– and directly compete everywhere in the world, and they’re going to crank out those phones cheaply. . . but don’t expect them to get really excited about whatever the successor to cell phones is, because it’s going to threaten everything about why they exist. They’re big, and they exist for one purpose. And if they have to change fundamentally, they’ll probably fail and go out of business.

You see what I’m saying? Look at really big companies with big research divisions. Are they funding those divisions to cause their fundamental business model to change?

(Someone says “no.”)

Is that a “no”? Then why are they doing it?

(Someone answers) “So that they can get control.”

They want to prepare for change – they’re not stupid. But do they want to propel us into the future to make their products as obsolete as soon as humanly possible? Hell, no!

You understand that if you are the officer of a publicly traded corporation, you have a fiduciary responsibility to your shareholders. It’s the one and only ultimate responsibility you have – to maximize profits. And if maximizing profits means keeping your product of record for as long as possible at the expense of innovation, you’d go to jail if you don’t do that. Officers of companies would be sued up the wazoo if they didn’t do the thing that maximized profits for their shareholders.

Is this bad? Not at all. As best as I can see it, this is a wonderful system. It’s just that it’s not the whole story – it’s less than half the story.

The rest of it is Obi-Wan in the cave, trying to figure out how to upset the Evil Empire—except that it’s not necessarily evil.

(Picks up MAYA’s playing cards that contain quotes and reads from one of the cards)

"In all things at any given time there is always an orthodoxy."

Always! It’s so hard not to think that your point in history is really special. That’s why I encourage all of you to go read some history – read this stuff in detail, read contemporary books written at the time – then you’ll learn how there’s nothing special about now except for the given problems that we’re solving. And the analogy between the industrial revolution and Andrew Carnegie and Henry Ford, and that sort of stuff and where we are now in the computer world could not be more compelling.

Real vs. False Structures

Are you done? (Looks at people working on the construction projects)

ONE A, please demonstrate the results of your efforts. (Looks at it, but it’s not complete.)

No no. Keep working!

(Takes the other projects. Lines them up on a table.)

The point was supposed to be that the cube with magnets would be a mess, and the two tetrahedrons would be just fine

What this exercise is about is the other quote in here. (Picks up the same book he quoted from before.)

(Before reading from the book, notes what Admiral Rickover said.) "The reason that engineering is harder than philosophy is that philosophies don't have to run.”

(Looking at the book in his hand) Here’s another quote. “A false theory in economics or politics may survive for hundreds of years. . . but for the engineer, the judgment is immediate and without appeal. The false structure falls and buries its author in the ruins."

Well, is that true or not? The problem is that false buildings don't just crumble. Incorrect engines don't run. The tetrahedron is a correct structure (picks up the tetrahedron made by the A’s). It doesn’t matter that these things (sides) are slippery and that they don’t provide rigidity along all dimensions. This is a very stable structure, and it doesn’t matter if you make it out of toothpicks or magnets.

(Tries to pick up the cube, but is having difficulty) I thought this one was going to be okay. With slightly stiffer Playdough, this would have been just fine. You can build this (points to the structure), but it’s cheating. A square really isn’t a structure at all (points to cube). It can be made to look right, but only because the stuff in the corners is secretly shoring up the edges with triangles you don’t see. It’s not depending on the correct structure to be rigid. That’s the trouble with the [engineering] quote in here. That’s how we have built buildings. We have just thrown stronger and stronger materials, and materials scientists are doing a great job, but the structural engineers are just getting lazier and lazier because the materials are getting better and better. And we don’t give any thought to whether the building is true or false.

What’s this have to do with Visage? Why did I say earlier that the computer world is unspeakably ugly?

Moore's Law has permitted us -- tragically -- to be unbelievably sloppy, for correctness to mean almost nothing. You can write a computer program, and if you're patient enough and clever enough, you can make something that is utterly magical – the computing equivalent of a building that will stand a mile high without even having to think about correctness and right structures. Because it doesn’t matter – Moore’s law right? Creative waste. But it will only get you so far. Moore’s law won’t go on forever. Even if it does, there’s the tremendous tragedy of waste. How much better computing could be for real people if only it were designed.

If we only deeply understood what the principles were -- that triangles are better than squares – and (pointing at the cube) that there are really secret triangles here to the extent it works at all. You can’t make a structure without a triangle. Just to belabor the point. . .

(Demonstrates a Bucky Ball)

This looks very random, but there’s nothing random about it at all. It’s made in a clever way. If I handed you the pieces of this, you’d come up with something. This particular one demonstrates the jitterbug transformation.

(Squeezes the Bucky Ball in and out, showing how it oscillates between two complicated symmetrical shapes)

This has very specific and interesting properties that you would never get by randomly putting this stuff together. Similarly, like transistors, you could mix germanium and silicon together, and you could do that randomly for the rest of your life, but you would never get a transistor. You’ve got to have the science behind it. You’ve got to understand what’s going on, and then you can do utter and complete magic, and people could spend the next two hundred years exploiting that magic.

But the key to all fundamental change is found in a few basic things that have to be driven by a deep understanding of whatever it is you’re working on.

So, I'm sensitive to being repetitive -- I know some of you have heard these things a hundred times before – but a) some of you haven’t; and b) there’s a nasty tendency to tune out the things you’ve heard before. But take it for granted that if I say it over and over again, I believe that it's particularly important. So don’t tune out.

Comprehensive Anticipatory Design Science

Okay, this is my last mention of Buckminster Fuller.

What he says he did was Comprehensive Anticipatory Design Science. All four words are very important.

Comprehensive means that you better not look at only one part of a problem.

Anticipatory is looking towards the future. The other famous design scientist I used to know is Herb Simon, CMU’s Nobel Laureate. He said design is the systematic attempt to affect the future. That’s as good a definition of design as I know.

Design -- as so defined

Science -- that's the empirical part. Science means testing––making hypotheses about nature, about the outcome of potential experiments, and doing the experiments to see if you’re right or wrong. As I see it, what MAYA is about is trying to follow Bucky’s advice: Comprehensive Anticipatory Design Science.

We’re engineers. How does this play out? We have chosen that pervasive computing is worth working on. Picking your battles is very important. This is the battle that Joe and I picked 7,8,9 years ago. Not many people, except a few academics, are seriously working on pervasive computing. It seems to be a completely done deal – that’s the good news with Moore’s law-- that it will be cheaper to make a random manufactured object an information device – to put some computing into it than it will be not to.

So the question becomes how to do Comprehensive Anticipatory Design Science if your goal is to own the future with respect to pervasive computing--influence the future in a positive way because, at the end of the day, we want to make the world better for real people. That's what any designer worth his or her salt would want to do. We want to make the pervasive computing world better for real people as soon as possible. That, as far as I can see, is what this is all about.

Visage: An Ambitious Agenda

Visage started out meaning something rather specific, and it's now my label for the entire enterprise. If this talk were 7 hours instead of an hour and half, I would have come in with a list of all the pieces of what I consider part of the agenda, and we could talk about how they will all fit together. But I want to pick a couple of the pieces and help you fit them all together.

I want to admit it flat out –it is probably an absurdly large and ambitious agenda -- and therefore one might argue it’s likely to fail. I don’t intend to fail, but if I were a neutral third party, I would have to conclude that it’s likely to fail. But that's true of any big, exciting thing.

The reason I talked about large companies and industrialization is that I need to motivate why you might think this might not fail. Because it’s Obi-Wan in the cave, right? The Rebel Alliance always loses until the big battle at the end when they suddenly pull it out of the fire.

Margaret Mead said "Never doubt that a small group of determined people can change the world -- indeed that's the only thing that ever does."

We’re not overthrowing orthodoxy because it’s bad. It’s just had its day. That’s just how progress happens. I believe in progress. And that might seem droll, but not everybody does. There are lots and lots of people who very explicitly claim that change is not good.

I believe in change, and if you don’t, you’re probably at the wrong company. I want the world to be better, and I don’t know how to make the world better without changing it.

I don't see how fundamental change ever happens unless it involves technology. Technology’s not the end of the story, but if you want to reliably make the world better, I suggest that you be a technologist, whatever else you are.

So, pervasive computing. Let's peel the onion. What that means from a technical perspective is lots and lots of processors, not one big computer, but many, many computers that comprise one big information system. It’s the trillion-node network. You can think of the entire world as being one big computer if you want to, but that’s not going to be a computer whose architecture is the same as the way we build PCs today. It’s not one big computer in the sense that an IBM PC is one small computer. Putting two or more or 16 processors in one box is maybe a toe in the water, but you can cheat when you’re doing that in a way that you can never cheat if you have a trillion processors. Things have to be far less coupled.

Obviously, even though many computers will be mobile, most of them will be relatively stable. The furniture in this room rarely leaves this room. That’s not mobile in the way a cell phone is mobile and moves literally around the world. Yeah, I can move these chairs around, but within extremely narrow limits. Therefore, it better be the information that flows. That’s networking. So we better get good architectures for information liquidity.

Information liquidity is fundamental to pervasive computing because you’re not going to move the machines around––you’re going to move the data around. So we have to get good at it. U-forms and all that stuff are simply the inevitable consequence. It’s not cleverness, it’s architecture that we need. We’d better have principles. We better be able to articulate the equivalent of saying triangles are structures; squares aren’t.

So what are those things [principles] that they teach you in computer school?

(No one answers.)

Nobody can think of one? That’s depressing!

All right, then I will. Modular decomposability. I usually label this the architecture of complexity. That is a principle, a design pattern of great generality. It doesn’t matter whether you’re building structures out of collections of U-forms that are made out of other U-forms and, you know, recursive data structures. Recursion is about modular decomposability.

You’d better start with that before we do anything else, and I think you were taught this in engineering school. This is not like some deep new thing that I invented. It’s just good engineering.

The Limits of OO in an Era of Pervasive Computing

By the way, things like object-oriented programming are an attempt to institutionalize that stuff. It’s best practice. What a discipline does is accumulate a bunch of best practices so that you can crank out the hundreds of thousands of engineers you need, and they can just intuitively do the right thing without having a conversation like this one every time they start a new project.

OO is a packaging of a good idea; it’s a packaging of modular decomposition that was originally conceived for designing large software systems. And it’s a pretty good one. Smalltalk was a damn good idea. Smalltalk 80 was the definitive articulation of OO pretty much in its mature form. That was the big idea, and it hasn’t changed much since then. And it’s now 2006, and we’re probably over the hump now with OO being accepted as the correct way of doing things – but just barely. So the next time you think, “We’ve been at this for 10 years now, and we’re not getting anywhere,” read some history and try to put it into perspective. These things take a long time. And I don’t think we have any reason to be pessimistic that we’re making good progress––but that’s a different pep talk.

So, OO is about modular decomposition. We don’t use OO very much in our world. Why is that? MAYA's Information Devices Architecture (IDA) is in some sense a direct competitor to OO.

There are two reasons (why we don’t use OO), and they both have to do with the particular needs of pervasive computing. First, it doesn't tell you very much about hardware design. And one of the characteristics you discover when you start thinking about pervasive computing is that the wall between hardware and software is less important than it used to be. Is firmware hardware or software? If you look at this machine and how it works (picks up his laptop), can you really draw a line at where the hardware ends and where the software begins? It’s a very fluid thing.

It would be nice if, in the next round of work, we could not have to worry so much about making a priori decisions about whether a given piece of function is done in hardware or software. And that’s the other desideratum. Device fungibility, right? Let’s think of devices before we think about hardware vs. software, and that’s what IDA tries to do.

This is pretty unique -- having a set of abstractions that are more specific than just modular decomposability. That, of course, is equally valid for hardware and software. It’s just that we have completely different ways of doing it. Componentization in hardware and OO in software -- you have to look hard to realize that they’re really the same thing. Let's talk about modules (not hardware or software), and then let’s design modules, and then much further down the path of actually building something, we can decide what modules are best done with hardware and what modules are best done with software. Seems like a good idea in a pervasive computing world, and more important than it was in a PC world.

Again, we’re being anticipatory. It’s not as good now, but it will continue to get better. And that distinction is more important as the world gets more pervasive.

There are deep, deep reasons why we, MAYA, are so perversely going out of our way to do it differently. Things that seem ridiculous on the surface I believe are necessary if you’re serious about change. OO doesn't really help with hardware. We didn’t throw away OO because we didn’t like it or we thought anything was wrong with it. It’s just not the right tool for the completely new problem that we’re gearing up to solve.

The other problem with OO is that it's really stinky for information liquidity. People are trying to "bend" OO as an information architecture. What does that mean? The way you get modularity in OO is with encapsulation, make a software machine, put your data in it, and then just call them methods. But what if you want that piece of information inside the little object to go from here (begins an arc with his arm) to a different machine (completes the arc) – which is going to be an exceedingly common operation in a pervasive computing world? Either you get the information out of the machine (in which case OO didn’t do you any good), or you try to move the whole machine. Moving the whole machine is like trying to move a wave from one swimming pool to another. You can’t really do it, but you could build a machine to read the shape of the wave and recreate it in another machine. That’s what serialization is.

Information as Things

Serialization is horribly inefficient, and it makes strong assumptions about the homogeneity of the two machines. So people came up with virtual machines. Java is an interpretive language, basically. People thought, we'd better build a layer on top of the hardware that makes these very different machines look the same.

You can make that work, but it’s complexity on top of complexity. No problem -- Moore’s law!

I’m tired of it.

I'm tired of using Pentiums for operations when PICs will suffice. I go to my library, and I see the poor little gray-haired volunteer lady scanning barcodes with the most complicated artifacts that humanity has ever developed. It’s pathetic. Step back from it guys. Look at it! Fifteen cents worth of hardware will build a barcode reader.

(Mimics a skeptic) Oh, yeah, but what about the programmer hours and all that sort of stuff?

(Normal voice) I’m tired of it.

This brings us to another part of the agenda: seizing the low ground.

One of the questions Slim collected was, Whatever happened to Interstacks?

It’s still there. Will we ever make a viable product of it? I don't know, but it's deeply important to the agenda.

Why Interstacks? My passion was to see how much we could build with ridiculously cheap and simple processors. It’s the extremely low-end processors, and it would have been so much easier just to throw a Pentium in there because the price of the Pentium was so much cheaper than the hundreds of hours to squeeze this into a little PIC.

I don’t care. It was figuring out a new way of doing computing.

How many of you have read the Introduction to the Expeditions to the 21st Century? Three or four? Ok, please read this. It was talking about device fungibility––picking up random pieces and putting them together. Anyway, the point is . . .

(Starts up the projector)

What we proposed was to show how in the 21st century you could build PCs out of small, modular arms-length pieces.

It is desperately important that, from the user’s perspective, we model information as things. What the WIMP paradigm mostly does, we’re going to always do. If you want to give a piece of information to somebody, you hand it to them as a thing. It stays where I put it. Sometimes it might look a little different (computers let us do that), but don’t make it so that sometimes it’s here, sometimes there, sometimes in both spaces at once, and sometimes it’s nowhere.

People, we have a million years of evolution for dealing with a world where objects are where we put them. Hans Moravec asked in one of his books, Why are we able to make a computer chess grandmaster, but we can’t make a computer that can move around the world as easily as a three-year-old? The answer probably is that chess is easy (for a computer) compared to moving around in the world, right? Chess seems hard to us, but we’re all Einstein’s when it comes to moving around the world because 99% of our brain has been evolving so that we’re unbelievably good at dealing with objects. So wouldn’t it be a good idea to model Cyberspace as much like objects as possible? Again, no new idea. Just competently applying what Alan Kay was writing about 30 years ago to pervasive computing in the future. It better feel like if I want something on my cell phone and I want to get it to your cell phone, I take an object and move it. And we’re trying to build a world that will feel that way when it’s done.

(Shows a slide of Interstacks modules)

So this was an illustration from the Expeditions proposal. The idea was that we would provide a base module, an SVGA module, a keyboard and mouse module, a TCP/IP module, and a Java module. It doesn’t matter if I were writing this today; it would be Python instead of Java. But that’s the whole point--it doesn’t matter. The TCP/IP module could be the only TCP/IP module in the whole world, no matter what you’re building, if it needs to communicate TCP/IP, you can engineer the heck out of it, and when the chips become available, just swap out the inefficient software and replace it with a piece of hardware. Nothing else in any part of the system ever changes.

I don’t see how we’re going to get pervasive computing until we can meet these requirements. And not sort of meet them, but absolutely, utterly, and guaranteed meet them.

(Picks up the tetrahedron.)

And that’s why we better be understanding the triangles – so we build without compromise even if it’s harder to build that way.

(Puts up slide of Visage and Interstacks modules)

This is what I want to build. That’s Visage, and it’s running on modules, no single one of which costs more than $30, $50, or $100 on its way down. If I could sell a million of them, I could build it for a dollar. Why should PCs be a $1000?

Enough with the Features, Already!

Fast, cheap, good -- pick two – the engineer’s mantra. We had an analogue to that in the computer world. There is performance, features, and cost. Pick two. At any given stage of technology, you could make it cheaper, you could make it perform less well, or you could remove features. What is the choice we’ve made?

Cost has been constant. PC’s have cost a $1000 for ten years. Performance is constant or maybe getting slightly worse. The processors are faster, but is Windows faster from the users’ perspective? On average, a Macintosh in 1984 booted in seven seconds. How fast does your PC boot?

So where does it go? It goes to features – and, boy, do we have the features! Historians for centuries will be documenting the features of MS Word. It’s absurd!

The trouble is that features have long since ceased doing the average person any good. It’s just not worth it to try to find them [the features]. So let's cool off on the features and let’s try to look at other things. Let’s make it faster. If I ever allow a finished Interstacks module to take more than 2 seconds to boot, I would like you to shoot me. They [the modules] collectively cost a couple hundred dollars. So it's taking that space and exploring a different part of it. All the hoops I ask you to jump through to build things in totally ridiculous ways are maybe not so ridiculous if you think through not where we are, or how things were built yesterday, or how you were taught to do them in college, but how they will have to be done tomorrow, when we have a different world—when there really is pervasive computing.

MAYA is about Innovation, Not Dollars

If I had wanted to put profit above innovation, it wouldn’t have been hard. This company is about innovation. Let’s remember what the goals are and what comes first. And just maybe if we stick to our guns for a little while longer, we’ll get something out into the world that will cause the way real people to live in 10, 15, or 20 years to be a little bit less ugly than what it now looks like it’s going to be.

It’s not easy because every step of the way, it feels like what you’re doing is wrong. And you have to let go of so many habits that treated you so well before.

Look at Steve Jobs' career. He and Wozniak started Apple literally in a Cupertino garage. They made the Apple I and took it to the local geek fest, and sold six of them, you know, and then they attracted some money and did the Apple II, and they ended up being very successful. Apple became a publicly traded company. And they managed just barely to get the Mac Lisa and the Macintosh out into the world.

Does anyone know the next step in the story?

(Someone yells) “He got fired.”

Exactly. And they hired the former CEO of Pepsi-Cola to run Apple Computing. That’s what happened when they were losing market share.

(Speaks excitedly and emphatically)

Steve Jobs didn’t care about running a publicly traded corporation. He cared about making the world a little better!

What does he care about? Would he change the direction of the company if a shareholders’ suit required him to in order to maximize profit in a direction he didn’t agree with? I don’t think so. He’d quit again.

It’s an interesting story. What did he do when he got fired? He started NeXT. How many of you have ever heard of NeXT ? (A few raise their hands.) Oh, not bad.

Jobs never made a nickel. But he got the Mach microkernel from CMU – very IDAish along some dimensions. And he got OS X designed, more or less. And when Apple was on its last breath, they invited him back.

The point is, he’s got a plan, and he cares about the plan.

It’s not exactly MAYA’s plan—he cares about different stuff. But of all the people who are alive in this generation, who can you point to who has caused more positive change for the real lives of ordinary people, or at least ordinary computer users? And it’s not because Apple dominates the market.

But I don’t think that Jobs considers his life a failure. He did okay. And I think we have as good a shot as anybody to do okay, too. But just remember what we’re about, what it takes, how the world works, what we’re trying to do, and how we’re trying to do it.

(Pauses)

That’s my pep talk.

return to Foundry Talks