Engelbart As UberTool?

Yesterday I described UberTool, an imaginary company planning to push a set of tools through a mutual-improvement process; their team would improve those tools, and then use those improved versions to improve them further, and so on through a rapid burst until they were in a position to basically "take over the world."  I asked what it would take to convince you their plan was reasonable, and got lots of thoughtful answers.

Douglas Engelbart is the person I know who came closest to enacting such a UberTool plan.  His seminal 1962 paper, "Augmenting Human Intellect: A Conceptual Framework" proposed using computers to create such a rapidly improving tool set.  He understood not just that computer tools were especially open to mutual improvement, but also a lot about what those tools would look like.  Wikipedia:

[Engelbart] is best known for inventing the computer mouse; as a pioneer of human-computer interaction whose team developed hypertext, networked computers, and precursors to graphical user interfaces.

Doug led a team who developed a rich set of tools including a working hypertext publishing system.  His 1968 "Mother of all Demos" to 1000 computer professionals in San Francisco 

featured the first computer mouse the public had ever seen, as well as introducing interactive text, video conferencing, teleconferencing, email and hypertext [= the web].

Now to his credit, Doug never suggested that his team, even if better funded, might advance so far so fast as to "take over the world."  But he did think it could go far (his Bootstrap Institute still pursues his vision), and it is worth pondering just how far it was reasonable to expect Doug’s group could go.   

To review, soon after the most powerful invention of his century appeared, Doug Engelbart understood what few others did — not just that computers could enable fantastic especially-mutually-improving tools, but lots of detail about what those tools would look like.  Doug correctly saw that computer tools have many synergies, offering tighter than usual loops of self-improvement.  He envisioned a rapidly self-improving team focused on developing tools to help them develop better tools, and then actually oversaw a skilled team pursuing his vision for many years.  This team created working systems embodying dramatically-prescient features, and wowed the computer world with a dramatic demo. 

Wasn’t this a perfect storm for a tool takeoff scenario?  What odds would have been reasonable to assign to Doug’s team "taking over the world"?

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • Grant

    Robin,

    I certainly can’t speak with an certainty on Engelbart’s situation, but I believe the barriers to innovation are often much more mundane than great ideas like mice and email. Demoing such advancements and actually using them as cost-effective tools are two entirely different things. IMO, the innovations needed to make such grand ideas successful are often invisible to industry outsiders. How useful would his advancements have really been without faster, smaller and cheaper transistors, or reliable packet-switched networking?

    I actually think “big” ideas are much more often the easy part than most people think. The devil’s in the details, and large markets are great for figure out how to make a better widget that powers doohickeys and thingamajigs. I’m skeptical any small team could get close to matching the performance of a market as a whole and really achieve the positive feedback loop that we see in computer science today.

    I’d bet a large number of people had ideas to build things like mice and GUIs, but didn’t bother because they had no idea if it would ever be cost effective.

  • Emile

    Grant: The trick isn’t only in having a “big” ideas, but in successfully convincing concerned people (investors, engineers) that the idea is indeed a good one. Only then can the resources become available to solve all the technical hurdles.

    From the investor’s point of view, plenty of people are pitching “great” ideas. Figuring out which ones are really “great” isn’t a trivial problem.

  • http://facelessbureaucrat.blogspot.com Bill Harshaw

    Seems to me you’re ignoring the maintenance rule. Everything you build requires maintenance/overhead. Two guys in a garage have very little overhead, but as the enterprise grows, the percentage of effort devoted just to housing, equipping, and paying the worker, and to supporting the results grows and grows. For there to be a takeoff, you can’t think just one big enterprise. Need to have new enterprises spawning off the mother. So you need to ignore the organizational lines, because people move among organizations. Think “Silicon Valley”, not Bootstrap Institute.

  • homunq

    Of course, we now know that mice and hypertext are powerful stuff. But they are not powerful stuff because they make an individual computer user massively more productive. They are powerful stuff because they enable more people to use computers (mice) and to leverage each other’s productivity (hypertext, when networked). These are the wrong technologies for an UberTool scenario.

    In fact, I think that there are no right technologies for UberTool. Human brains, agriculture, and industry also prospered due to network effects. You can’t build them in a box. And when someone tells you that AI will be any different – that a “hard take-off” is possible – you can effectively zero out your probability that that person will be remembered as a successful AI pioneer when AI eventually succeeds.

  • homunq

    OK. “Possible” doesn’t earn them a near-zero probability. But “probable” does.

  • Ian C.

    Poor Engelbart. He gives the world the mouse and what’s his reward? Being called an Ubertool.

    I think that for a tech to take over the world it has to be generally applicable across many fields. And unfortunately the more general something is, the more pie in the sky it will sound to an investor.

    AI is obviously the height of generality, so it does have the potential for world dominance. But Engelbart was on to something too.

    He realized that even though companies perform different specialities, and they improve their techniques differently, they could all improve their technique-improvement process the same way. He realized that at the meta-meta level everyone unifies. In that sense he too was being general enough to take over the world (or at least all companies). He didn’t because his hypertext/wiki ideas probably weren’t enough, but the identification of that unification alone deserves props.

  • John Maxwell

    I agree with Grant. I think that innovation is a slippery pig, and that it is rarely the visionary who profits from his vision because of the mundane details that need to be worked out to make something a success in the marketplace. Xerox PARC tried to build on Engelbart’s vision, but the Xerox Star was a failure in the marketplace. Apple was the one to figure out how to make it profitable, but only after the Lisa was a failure. Hypertext didn’t really take off until Tim Berners-Lee invented the World Wide Web. Furthermore, neither Engelbart nor Xerox PARC thought of the spreadsheet, which was the killer-app for early PC use.

    There is also the productivity paradox that economists observed in the 1980s. In spite of huge investments in IT, economists weren’t seeing productivity improvements among white-collar workers. It took a while for white-collar workers to figure out how the new tools could make them more productive.

  • http://jed.jive.com/ Jed Harris

    I worked for Englebart in 1969-70, I worked for a little while in Alan Kay’s group at Xerox PARC, I knew the folks in the Xanadu project moderately well and followed it closely for several years, and I worked at Apple and then in venture capital for years.

    Arguably I never made any revolutionary contributions of my own but I was an active observer and have some warranted observations about these questions.

    First, I did think at the time that Englebart’s, Alan Kay’s, and Xanadu’s visions (and to a lesser degree the technical contributions by Englebart and Kay, but not Xanadu) had revolutionary potential. That was why I spent time with them. My expectations more or less fit the later reality. So at least some revolutionary ideas are recognizable at the time.

    Second I didn’t see the implications of other even more revolutionary things I was exposed to. I was there when the second IMP was installed in 1969 but didn’t intuit the implications of the internet. I did figure out implications of cheap local networks in the mid-1970s but didn’t see the social implications of the internet until we got early Mosaic demos at Apple. Etc. Furthermore except possibly for Paul Baran in one RAND report, nobody really laid out any vision of a networked society before the early 1980s. Even Baran (and Vinge, Gibson, etc.) missed most of the impact.

    So some really huge implications are hard for anyone to see much in advance.

    Most important for Robin’s question, in the projects I’ve seen, almost none of the actual value was “internalized”.

    • SRI and Englebart didn’t create the products that generated value from Englebart’s ideas.
    • Xerox and Kay didn’t create the products that generated value from the Xerox PARC ideas (though Xerox tried, and arguably thereby retarded the wider impact of those ideas with IP encrustations).
    • The Xanadu project failed partly because it was so successful at keeping all its IP secret, so it got no external adoption, help, or constructive criticism (all of which it desperately needed).

    My conclusion from all this is that no single organization or project could succeed the way Robin describes. I agree with other comments that the larger network of innovation is required — complementary innovations, variants that work better, translation into important niches or applications, etc.

    In fact all my observations indicate that if innovations are kept secret or their IP is fully protected, it usually kills them. The only exceptions I’m aware of is when they are very self-contained “point technologies” like xerography or the vulcanization of rubber.

    The opposite examples are the internet and the web, both of which succeeded partly because their IP was all completely open, and their development was broadly collaborative. If they had been owned and protected, say by telecom companies, we would not have seen anything like the current social value they generate.

  • haig

    “Taking over the world” is a very vague metric. Does it mean becoming the most financially successful economic entity? Does it mean influencing the ideas of generations of people? Does it mean physically conquering the world through might? If speaking strictly in the economic sense, then Englebart and his team didn’t take over the world because they didn’t have a business model that allowed them to exploit their innovations in the marketplace and create value directly as a business (or if they did, then it was either poor or poorly executed). In hindsight, we can see how he could have formed some ubercompany that combined apple,microsoft, cisco, google, etc. in one superentity but he didn’t. If I remember correctly, I believe he formed a consulting company instead of selling product/services; he saw (rightfully though much too idealistically) that the most important ideas of his were not the tools, but the new ways of communication and cooperative work that they made possible.

    But if you are talking about influence as passed on by the memes he created, then It can be argued that Engelbart’s vision did take over the world, he just didn’t position himself financially to capture the wealth created, other organizations did that.

  • Brian Macker

    Robin,

    (Ubertool == Capitalism), except without the inefficiencies of communism implicit in having all of capitalism run under a single company with no price structure. Get it?

    Wasn’t this a perfect storm for a tool takeoff scenario?
    By a single collective entity? No, that would be the U.S.S.R during the communist era. They were by far the single collective entity with the most tools. They were making tools to make other tools. Machinery doesn’t get made with sticks and stones. So they were following the model.

    What odds would have been reasonable to assign to Doug’s team “taking over the world”?
    Pretty close to zero, if pitted against a bunch of other companies working on a free market model.

  • http://blogjack.net Glen Raphael

    Engelbart’s team had a Not Invented Here problem. His UberTool tried to combine a bunch of component software – a text editor, an email program, graphics software… – many of which you could find better implemented as individual components somewhere else at the time. So the UberTool would be a step backwards for many users. Though we remember the ideas from his group that turned out to be quite good, we forget that at the time he was selling other ideas bundled with those that turned out to be goofy or impractical or unmarketable. For example, his vision included that people who were using a mouse or light pen in one hand would to use a Chorded Keyboard in the other so the user could efficiently type and point at the same time.

  • Tim Tyler

    Many people today do at least use left-handed ctrl-c, ctrl-v and ctrl-x “chords” in conjunction with right-handed mouse operation quite a bit.

  • http://blogjack.net Glen Raphael

    Sure, but what they don’t do is learn how to type on a one-handed chorded keyboard and use that entirely instead of a standard keyboard as Engelbart and his team expected they would.

    Apple has often managed to force new hardware standards on the world by deciding at just the right time this is a better way to do things; if you try it you’ll be more productive. (Machines with no floppy drive, pen-based PDAs, multitouch input phones and many other moves all had their skeptics). On the other side, Microsoft has often managed to sell monolithic this does everything application suites that aren’t quite as good for some purposes as specific specialized apps from different vendors. So it’s not inconceivable that someone could have sold the world a monolithic application requiring multiple new interface methods. But the limiting factor on doing this isn’t foresight. There’s also marketing, luck, timing, persistence, charisma, design talent, development talent, managerial ability and more. It would be quite a coincidence if one small group at SRI had excelled in all of those at once and my sense of it is that general business savvy was a weak point.

  • Doug S.

    “One hand on keyboard, one hand on mouse” is standard for computer game playing, but yeah, I see what you mean. We don’t type documents on video game controllers, after all.

    Other candidates for “could have taken over the world?”

    1) IBM, back in the days when it was THE computer manufacturer. Similarly, Apple Computer could have ended up with a Microsoft-like domination of the home computer market, but they lost out to the IBM compatible machine that ran Microsoft’s version of DOS. And speaking of Microsoft, haven’t they pretty taken over the world already? Imagine what a James Bond villain do with a few back doors into the Windows operating system…

    2) AT&T, back when it had full control over its phone lines, to the extent that it was illegal in the U.S. to connect a non-AT&T phone to their telephone network.

    3) 19th century energy companies. Today, you need electricity to do pretty much anything – could people back in the era of Thomas Edison and Nikola Tesla leveraged that into a Take Over The World scenario? I’ve made half-joking remarks about the Singularity having occurred in 1876, when Thomas Edison invented the industrial research laboratory, but there really is a grain of truth to it.

  • Pingback: AI-Foom Debate: Post 1 – 6 | wallowinmaya

  • Pingback: Overcoming Bias : Friendly Teams

  • Pingback: Overcoming Bias : That Old SF Prejudice

  • Pingback: Links for December | More Right