The Technium

Turing’d


[Translations: Japanese]

Over the years I’ve had many opportunities to work with professionals from various fields. In every endeavor, computer technology is utterly transformative. But not every field gets this. Some scientists, licensed experts, and professionals are allergic to new technology.

I had an epiphany recently on why some varieties of professionals are more welcoming of disruptive technology than others.  I realized the types of pros who are most eager to employ the latest technology are those fields which have already been Turing’d.

We have this long list of tasks and occupations that we humans believe only humans can do. Used to be things like using tools, language, painting, playing chess. Now, one by one they get Turing’d. A computer beats them. Does it better.

Images

So far we’ve can check off arithmetic, spelling, flying planes, playing chess, wiring chips, scheduling tasks, welding, etc. All have been Turing’d.

Computer scientists are great to work with, because in general they are completely fearless. They were Turing’d long ago. They grok that many of the tasks they used to do can be done much better by computers. On the other hand, doctors as a rule are loathe to accept new technology because what they do is hard to delegate to computers. Ditto for a lot of biologists.

Within biology there are certain fields that have already been Turing’d. For instance, phylogeny, the study of taxonomic trees, how different species are related to each other. Figuring out phylogenetic trees turns out to be something computers can do better than even the smartest, most learned humans — even though nobody believed this only a little while ago. So phylogenists are Turing’d and very open to new ways of doing things.

Phylogeny

Taxonomists and field biologists, on the other hand, still believe that computers can’t recognize or classify organisms as well as human can. Ooops. They are about to be Turing’d. Doctors, too.

Once you are Turing’d it is much easier to believe other occupations which we humans used to do uniquely, can be done by computers. You tend to be open to disruptive technology in all parts of your life.

Have you been Turing’d?




Comments
  • http://www.whywenothithard.com Mike

    So far there aren’t any robot kickboxers :).

  • http://wanderingstan.com/ Stan James

    I’m turing’d in career, and also in leisure: Playing the original “Adventure” as a child one day, it occurred to me that one could write a program that would try every possible sequence of commands and inevitably win the game. Suddenly the game seemed a lot less interesting than writing that program!

    “Turing’d” is a good and important new adjective.

  • Shaye Horwitz

    What I wanna know is in what order we can expect other stuff to be Turing’d. How long till, say, novel-writing? Or take the chess example you gave – how long till go gets the same treatment? Voice acting? (I’m guessing go, then VA, then writing – but there are lots of others.)

    Some of this, of course, would be more a hardware issue than software, particularly anything that involves something other than “manipulation of information” (as broad as that is). But that’s robotics. And it has an equivalent, too, though software’s a part of that as well.

  • http://www.psonnets.org/ Michael

    Can a computer write a better poem?

  • http://www.SalesDuJour.com Gary S. Hart

    When computers create computers on their own, we’ll be in trouble.

    Colossus ‘The Forbin Project’.

  • Samy

    I like this article. I Stumbled here. +1 for the Heinlein drop.

  • http://simplyted.blogspot.com Ted Holmes
  • http://www.milsworks.net/blog Robbo

    Fascinating stuff. As a puppeteer for many years I felt semi-Turing’d with the advent of CG animation – but my most immediate thought was: dreaming.

    Do autonomous self-learning computing systems engage in anything which could be considered even remotely comparable to dreaming?

    While there’s a fair range of insights into what exactly dreaming is – I’m firmly in the camp of it being the two sides of the brain wresting with how to organize the flood of information absorbed by all of our senses throughout the day. Making sense of our senses. Inspiration lurks within there too.

    To have specific thinking tasks replicated by computers is one thing – to have these “machines” then go forward and engage in dreaming will be quite another. That is something I look forward to seeing.

    Cheers.

  • Tom Buckner

    I deliver mail. In principle we’ve been semi-Turing’d for years; we were an early adopter of OCR machines for mail-sorting. Even so, hard-to-read addresses and other such judgment calls, and physically delivering mail, require humans. However, that could all be changed if it were cost-effective. There are fully robotic warehouses, after all: we postal workers merely do much the same thing those robots do, but instead of warehouse with controlled conditions and standard shelf heights, we operate out in the elements under unpredictable conditions, delivering to millions of awkward locations.

    E-mail and other modern technologies have obviated much of what I do, though certainly not all.

    It would be hard to invent a job more inherently uninteresting.

  • Ron

    http://en.wikipedia.org/wiki/AI-complete

    We have yet to probe the limits of data mining and bottom-up AI, but like crowdsourcing, the lack of authority can be a mixed blessing. In AI as in life, there are a lot of evolutionary dead ends.

    A legal-argument search engine, to retrieve appropriate precedents for an argument in my brief, would need a lot of contextual knowledge to be helpful, not only on legal concepts, procedures, jurisdiction, and so forth, but also in the context of the case I am trying to make overall. It will take a long time for the lawyers to be Turing’d.

    The parallel human brain does not map very well to the Turing machine.

    • Kevin Kelly

      Ron, yes it will be a long time before lawyers are Turing’d — but on the way will be AI-assisted lawyering, which will cause all kinds of havoc, denial, and fright.

  • http://www.ridinghoodmusic.com Sebastiaan Elsenburg

    O, but that apostrophe is terrible: why not just say:

    ‘Turinged’.

  • http://mollykleinman.com Molly

    Librarians have absolutely been Turing’d, or at least some of us have, but many seem unwilling to admit it. Librarianship as a field is split at the moment, between the embracers of disruptive technology, and the loathers. Embracers will win, of course.

    And while we have been Turing’d, that doesn’t mean we’re out of a job. But that’s not what you seem to mean by Turing’d. You’re talking about fields and professions that still exist, despite the computers that do the work better. I’d love to hear you expand on that distinction at some point.

  • http://blog.amusecorp.com/index.html/188 Vasu Srinivasan

    Try asking google image search for a baby it lists a man’s picture. My 3-year old yelled, “Hey!! That’s not a baby. That’s a man.”
    Check this out:
    http://blog.amusecorp.com/index.html/188

    Certain things can be Turing’d only in theory,as reality is aeons away. Ask Marvin Minsky.

  • http://cube3.com LarryR

    ” Used to be things like using tools, language, painting, playing chess. Now, one by one they get Turing’d. A computer beats them. Does it better. ”

    Computers dont do anything beyond the rules within them. I take issue with Painting– unless you only mean the task ala “welding”, as i take issue with “language” if you refer to it’s USAGE as in a poem, story, or emotive communication.
    Occupations based on repeated tasks are all subject to the faster gunslinger, or faster mechanism. Old story played out for many eons.

    Im still puzzled by the religion of technology since its a fools game as any dogmatic system.

    Is technology ONLY the current Religion to be used to foster the US and Thems of control, power, and authority? The 1985 Apple Commercial showed us “the computer” as liberator from dogma and authority….but as the truth of the first decades of the mass computer age has began to show clearly technology is just that hammer. The human that throws it is still the only who decides where it lands or why it’s thrown.

    1+1=2 never really defined Humanity or it’s success or failures. It only assists in mediating them.

    1+1=3 and the response and valuation of that equation has always been the true HUMAN equation.

    Whats BETTER? why is faster better?
    Thats the only question worth trying to answer….and like Corben found out, our answer and the machines are not nessacarily the same.

  • http://maheshcr.com/blog Mahesh CR

    What about self-deception? Satire? Self-mockery? Humor? Will computers hit the bottle? Will they emerge from loss wiser?

    Would they have stopped evolving the Dinasour because reason ruled they were not quite right?

    Not all pathways of life are reasoned. The lines of our intent run deep, perhaps to the very start of creation. The machine shall supplant man from reason’s realm to varying degrees of effectiveness.

    Life is curious mixture of tangible biology and intangible consciousness. The latter shall elude the machines until the end, they have escaped the maker of the machine for so long!

  • http://cognections.typepad.com/lifeblog charlie

    Really, getting Turing’d just means that what you do can be described by a set of procedures and computed. A good place to see what professions will next be Turing’d is to look at what gets outsourced. Tom Peters always calls out that things that can be automated, will be.

    What do you think?

  • http://www.debategraph.org David

    Turing Turing’d!?

    Can Computers Think?

  • http://www.drbradley.com Craig, MD

    Physicians are amusingly/alarmingly resistant to letting any part of our world be Turing’d. But there are huge parts of medicine that could be better managed by, if not fully Turinging them, at least mechanizing. Clinical pathways and electronic decision support; automatic dosing; even something as seemingly personal as selecting referral centers and specialists could lead to enhanced patient care if it were more based on calculated outcomes and guided referral.

    But much of our profession fights it tooth and nail.

    As a corollary then, to the point of the post, most medical offices and even hospitals so reluctantly accept computerization of even the mundane tasks. You may have a bank that you can log on to automatically and transfer money, but if you have a physician from whom you can extract information or a refill electronically, you’re in a small (but slowly growing) minority.

    Nice post.

    @Mike: yeah, but an even modestly designed robot could be built that could not be beaten by a kickboxer using only kickboxing skills.

  • http://www.nealsheeran.com Neal Sheeran

    flying plans marked off the list? I don’t think so. While computers have things more efficient and accurate, there is still a pilot, who is human, “flying the plane.” With UAVs – there is still a pilot, he just happens to be sitting in a trailer on the other side of the planet.

  • http://junesix.org Patrick

    I believe that any occupation that consists of repetitive tasks, data aggregation, or step-wise algorithms is ripe to be Turing’d. The early computer-aided robotic assembly lines took care of the first part. The cheap availability of massively parallel processing (and cloud computing) is now taking care of the second and third. We’ve already seen it with online shopping, buying airline and hotel tickets, and even doing taxes. Medicine is definitely next on the block for examination. While we’ll always need doctors, it’s no surprise that the algorithms doctors learned from years of med school and training to diagnose and treat are already being dissected and Turing’d as we speak.

  • http://fury.com Kevin Fox

    I’ve always interpreted the Turing Test to relate to a computer’s ability to emulate a human to the point that its ability is indiscernible from a human’s.

    While it’s no mean feat to program a computer to weld or fly a plane better than a human, I think this is a different accomplishment distinct from being ‘Turing’d’.

    Consider a pre-Turing Turing test of the painting vs the photograph. In the 1800s, painting was a purely human endeavor. The ability to build a contrivance, a mechanical thing that would capture a visual representation of a scene hitherto impossible without an artist wielding a stylus or brush, was a major accomplishment and on the purely objective metric of detail the camera won hands down. In later years it would also surpass the best painters in color fidelity as well.

    But that doesn’t make a camera a better painter than a human, nor is a camera a device that creates works that are indistinguishable from a human artist’s work. It merely provides a different kind of perfection.

    A computer program may be able to play chess ‘better’ than a human, but unless it plays in the same manner as a human then Turing is the wrong stick to measure by. Turing’s test was never about machines that could surpass humans, but ones that had advanced enough AI (whether hard or soft) that they were indistinguishable from humans.

  • http://www.tonebank.dk William

    Computers can’t get ideas. But it would be sweet if they could, they would probably be very imaginative.

  • Jon Miles

    Here is a field that is starting to be Turinged: the history of language evolution. A 2005 paper in Science, “Structural Phylogenetics and the Reconstruction of Ancient Language History” applied the same techniques of computerised phylogenetics described above to grammatical features of languages. The paper showed that it was possible to unravel relationships that were obscure before, and to break the 8,000 year barrier after which divergence of vocabulary is too great to show relationships.

    I was amused to see the reactions to this paper on language discussion boards. They started with “This is nonsense”, then moved to “Nothing new here – people have looked at grammatical features before”. They don’t know yet that they are being Turinged.

    The link to the abstract of the paper is http://www.sciencemag.org/cgi/content/abstract/309/5743/2072?maxtoshow=&HITS=10&hits=10&RESULTFORMAT=&fulltext=language+phylogenetics&searchid=1&FIRSTINDEX=0&resourcetype=HWCIT

    • Kevin Kelly

      Jon, thanks for the example and link.

  • http://simplyted.blogspot.com Ted Holmes

    Kevin, this is a great point.

    Perhaps an intelligent system passing the formal Turing Test could fall in love, write a heart wrenching poem and add something eloquent to this discussion. But we need a way to recognize the more granular, narrow ways Artificial Intelligence is presently evolving. I think you’ve nailed it nicely.

    Just as an unassembled robot COULD be said to fail the test of an actual robot, it is a small leap to assemble the parts so that they can communicate as a working unit.

    We can already see chunks of A.I. emerging from the Network. And many of them are being networked to work together.

    So here is a small list of ‘turing’d’ systems which only recently were considered solely the domain of humans. (Hope the links come out alright)
    Automated:

    decision making

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22use+automated+decision+%22&btnG=Search

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22Software+Takes+Airplane+Controls+Before+Crash%22&btnG=Search

    analysis

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22use+automated+analysis+%22&btnG=Search

    testing

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22automated+testing+is+used%22&btnG=Search

    learning

    http://www.google.com/search?num=100&hl=en&safe=off&q=+%22machine+learning+is+used%22&btnG=Search

    creativity, invention, design and autonomous self-improvement

    http://www.google.com/search?num=100&hl=en&safe=off&q=creativity+machine+%22john+koza%22&btnG=Search

    simulations

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22machine+simulation+is%22&btnG=Search

    even perception such as:

    vision

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22machine+vision+is%22&btnG=Search

    hearing

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22machine+hearing+is%22+&btnG=Search

    smell

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22E-Noses%22+%7C+enoses&btnG=Search

    touch

    http://www.google.com/search?num=100&hl=en&safe=off&q=machine+%22+sense+of+touch%22&btnG=Search

    and taste

    http://www.google.com/search?num=100&hl=en&safe=off&q=%22robot+*+taste%22&btnG=Search

    Glad you started the discussion. Gotta run. My job hasn’t been taken yet :)

    Ted

  • http://www.updike.org/ Jared

    I disagree that spelling has been Turing’d. Computers can’t overcome their user’s ignorance to get them to spell homophones correctly. (to vs. too, your vs. you’re, there vs. their vs. they’re, its vs. it’s and so on)