Six Pixels of Separation - The Blog
February 15, 2011 8:44 PM

A Deeper Shade Of Blue

The lower east side of Manhattan was buzzing and unlike any other usual New York City summer night, you walk fast.

You zig and you zag as if where you're heading is your ultimate and final destination in life. The city is tough. It feels very self-involved. It's a Billy Joel song. It can feel like a clenched fist. On this particular Wednesday night in July 2009, it was even harder to keep my wits about me. I had just finished the sold-out public launch of my business book, Six Pixels of Separation, and my literary agent was whisking a small group of us to a private dinner to celebrate. As one of the few non-New Yorkers, I struggled to keep walking pace with people who regularly chew up 5th Avenue the same way most of us exercise in the morning. The pace was more intense because I was deep in discussion with Stephen Baker. My head was spinning... I'm a big Stephen Baker fan.

You don't know Stephen?

At the time, Baker was a Senior Writer for BusinessWeek, had launched the seminal BusinessWeek Blog, Blogspotting (along with Associate Editor, Heather Green), and had just published the amazingly interesting book, The Numerati - which looked at how some of the smartest people in the world were leveraging mathematics, data and analytics to alter human behavior. To say that I was "boxing out of my weight class" would have been an understatement. I must have held my own because Stephen and I not only enjoyed some great food and drinks together with our group that night, but we stayed connected. In the spirit of Valentine's Day this week, it seems like the perfect opportunity to see what Baker's latest passion project is all about.

Computer vs. Human Being.

"IBM's Deep Blue computer beat Gary Kasparov in 1997 in chess," said Baker via Skype chat late last week. "If you think about chess as a game, it's a limited domain... which is saying something, because there are still billions of possibilities within the limitation of a 64-square board, whereas Jeopardy involves knowledge, which is limitless and language, which is limitlessly complex. IBM had to teach this machine they called, 'Watson', to understand very strange and puzzling Jeopardy clues and then to go and hunt - within its own database - to not only find the right answer, but to bet on it with a strong game strategy in a very limited amount of time. Watson does all of that. It's entertaining to watch."

Jeopardy - The IBM Challenge.

You may have seen the television commercials for Jeopardy - The IBM Challenge (the three-part television event started airing yesterday, February 14th, 2011). You may have seen the PBS Nova documentary, Smartest Machine On Earth, on it. You may have read something about this new supercomputer in these pages over the past little while. Well, Baker, literally, wrote the book on it. Final Jeopardy - Man vs. Machine and the Quest to Know Everything, tells the behind-the-scenes story of how IBM's new supercomputer, Watson, would challenge two of Jeopardy's greatest all-time champions (Ken Jennings and Brad Rutter) to see if a computer can actually beat a human (or two).

And no, this isn't just a parlor trick with a faster computer that's better at using Google than you are...

"That wouldn't be fair," laughs Baker who released a digital version of the book, Final Jeopardy, a few weeks ago that was missing the last chapter (which will only be sent once the final episode of Jeopardy has aired). "Just like other player's on Jeopardy, Watson has to have the information in its 'head,' so no... it's not a machine that can manipulate Google faster or better than someone else. Google actually works with our brains - not as a brain. People input on average three words into Google that creates a parameter around a bunch of web pages, and the user has to use it's own brain power to not only choose those words, but then to figure out what the results are that were spit back to it. You use your brain before and after with Google. Watson does this all itself. It makes sense of the words, searches out the answer, calculates the exact answer and then goes into the gaming mechanics."

This all begs the question: what is it about human beings and our desire to create something that is bigger, better, smarter and strong than us?

Baker thinks that this may have something to do with the fact that we - as a species - feel somewhat alone in this universe. "We put ourselves to work to make these friends," he suggests. "If you look at history, we've been making and building machines to augment and replace our limbs - tools of steel and steam power - which have lead to wheels and cars and wings and airplanes. We have done all sorts of things to augment our body, strength and movement, and now with the information age, we're moving to the head and brain. We started with calculators, but now we're in the realm of words and knowledge, because we are starting to have machines that are powerful enough to do this sort of thing."  

While there is no current business application for Watson, it's clear that this type of computational thinking has tremendous business potential (at some point in the future).

This could well-be the future of search, it could be what a true business assistant might look like or just the very beginning of true artificial intelligence. Currently, Baker says Watson is the equivalent of 2800 computers working together and that it fills an entire refrigerated room with way too many algorithms that run at multiple times. According to Baker, the producers of Jeopardy actually built the set around Watson's infrastructure and brought the set, crew and talent to IBM to film the three-episode escapade. Ultimately, it is instances like this that force us to think about a moment in time where Watson (or something like it) is able to replace human beings to become the ultimate employee?

Does Watson actually think like a human being?

"When discussing a machine that behaves in certain ways like a human, it's very hard to divorce it from human processes," Baker said in a recent Blog post. "This was a struggle for me in the book... it could be argued that Watson does not think, know, or remember. I tried to avoid attributing those words to it in the book. It does, by contrast, process, calculate, recall, and estimate. You can analyze Watson's information processing and debate whether it actually thinks. I've gotten into those discussions, but I don't dwell on them in the book. It's 'thinking,' if that's what you want to call it, is very different than ours."

Rise Machines. Rise.

I believe they said the same thing about the HAL 9000 in 2001 - A Space Odyssey or any computer that became self-aware and then reaped devastation on humanity from movies like The Matrix and The Terminator to scores of Science Fiction novels. Baker assures me that unlike those technologies, Watson is still fairly neutered. "We won't see that in our lifetime," he laughed.

The above posting is my twice-monthly column for the Montreal Gazette and Vancouver Sun newspapers called, New Business - Six Pixels of Separation. I cross-post it here with all the links and tags for your reading pleasure, but you can check out the original versions online here:

By Mitch Joel


Comments Comments Feed
  • Posted by Meghan Ward
    Mitch Joel

    Mitch, I'm also a writer and curious to know how newspapers/publications feel about you reproducing your articles on your blog. My assumption is that publications want exclusivity, for the most part. How do you negotiate this? Do you ask the papers for permission? What is their response? Have you ever reproduced magazine articles here on your site?

    I'm currently reading "Six Pixels" and am really enjoying it. Superb.
    Thanks!

    Reply
  • The limitless potential regarding the future of machines is incredible. All we have to do is look at what has transpired in our lifetime. You touched upon HAL 9000, Terminator, etc., and for good reason. Who is to say the future will be secure from machinery armed with artificial, complex and innovative brains?

    Perhaps not in our lifetime? Okay... But at the current rate of technological advancements, who is to say what will or will not be.

    Reply
  • Posted by Bryan Person
    Mitch Joel

    Mitch: I'm fascinated by this Watson story, and have been following it closely -- including reading all of Baker's book, save for that forthcoming final chapter -- over the past few weeks.

    Watching the matches themselves has been entertaining -- Watson absolutely wiped out his human counterparts in the second half of the first match yesterday -- but I'm also interested in the long-term implications that Baker lays out in his book. A Watson-type Q&A device could help doctors with a medical diagnosis that they might miss on their own, for example.

    I've seen concerns in some corners that Watson's success is some indication of computers becoming "too powerful." I don't see it that way at all. Instead, there's a real opportunity to harness a Watson-like computer for what it's good at: quickly searching and making sense of LOTS and lots of information that our human brains wouldn't be able to store, anyway. In the end, it still would come down to the human to weigh the evidence presented by the computer and make the appropriate decision.

    Reply
  • Posted by KevinMGreen
    Mitch Joel

    Some coverage has certainly suggested that "Watson" is the beginning of the "Rise of the Machines." For the most part, it's been easy to say that Watson is controlled and that we won't see rogue computers taking over society any time soon.

    While that is certainly true, I wonder about the increasing laziness of humans. Could a "Watson" based machine replace cognitive reasoning and human emotion for some people? Will some individuals second guess their own knowledge for that of what a super computer tells them?

    The entire project is fascinating and the ensuing conversation and dialogue has also been (for the most part) rich in valuable debate.

    Reply
  • Posted by Amod Munga
    Mitch Joel

    Great piece of writing as always, Mitch.

    To your question about why we as human insist on building machines that are better,smarter or stronger than us, I venture this opinion: it's not that the machines are better, smarter or stronger than us. It's more how do we stack up as the "smartest" creatures on the planet.

    We're testing ourselves. Are we so strong, so smart that there's nothing on this planet, not even created by our own hands that can better us?

    We've run out of mountains and the titans are a myth. So we're creating dragons to slay. Will the dragons one day decide they've had enough of being toyed with? At the risk of sounding like a conspiracy theorist, I think Phillip K Dick may have given us the answer.

    On a different tip, it could also be that when these machines are beaten, the person or people who vanquish it can claim they are better than the people who built it. So maybe it's just a primitive dick-swinging contest and we haven't really evolved at all.

    Either way, it's entertaining...until one of them comes looking for Sarah Connor...

    Reply
Add a Comment

Please complete all the fields below, including the spam filter (to prove you're not a robot).

  1. Fill in your email address to have your Gravatar photo included with your comment.
  2. Please type the word pixels here:
TrackBacks

TrackBack URL: http://www.twistimage.com/movabletype/mt-tb.cgi/2333