Via regular reader feeblemind comes this clip of the IBM-designed Watson on Jeopardy.

It’s a very impressive display of computer programming from the short video available. Call it artificial artificial intelligence at this point. But the future is looming in this display of computer programming prowess.

So, I wonder what track a true AI will take. Will it be HAL 9000 or HOLMES IV? Or will it be Neuromancer? Or SKYNET?

(Actually, dystopian views of AI are actually pretty dominant in pop culture. Mike from The Moon is a Harsh Mistress is kind of exceptional these days. But then, Heinlein thought technology was good. Far too many Luddites these days think it’s bad – except for their iPhones, of course.)

Anyway, watch Watson kicking some serious butt in the video.

This entry was posted in Geek Stuff. Bookmark the permalink.

4 Responses to HAL 9000 Or HOLMES IV?

  1. Foxfier says:

    I’m not sure we’re able to make true AI, as in actually able to make moral choices. (Counter to Star Fleet, Data was a person.)

    I do think we’ll end up with machines that do the “trouble shooting” step program, even for life and death situations, which will pretty uniformly be bad…. (utilitarianism without even a face, goodie)

    OTOH, if we can get the UI for a psudo-AI good enough, they can make computer interactions a lot more comfortable. (I’m picturing a computer nurse that’s the umpty-great grandaughter of voice recognition phone systems.)

  2. bob sykes says:

    Hubert Dreyfus’ critique of AI (“What Computers Still Can’t Do,” 1994, MIT Press) is still relevant. There has been no progress in AI since the 50s. Even the theoretical stuff is a shambles. Deep Blue and Watson owe their limited successes to brute force elimination of possible choices. How humans do this stuff is still a mystery.

    The AI situation is curiously similar to tokamak fusion: no advances, none, since the 1970s.

    The use of computers in K-12 education has also been critiqued (Clifford Stoll, 1999, “High-Tech Heretic,” Anchor Books).

    There is one thing computers are very good at: well-defined repetitive jobs in manufacturing. Over the last 20-30 years, manufacturing output in the US has doubled (and remained a constant fraction of GDP), whereas labor productivity has tripled. That’s where all the jobs have gone, not to Mexico or China, but to the machines.

  3. Foxfier says:

    That’s where all the jobs have gone, not to Mexico or China, but to the machines.

    That would be true if we were building a lot of automated stuff over here– but we’re not.
    We’re building it elsewhere. (Not always because the operation is cheaper, since getting PERMISSION to build stuff is a cold iron dog.)

    The old “machines are taking our jobs” thing is just the old, re-warmed BS that’s been around all century.

  4. Bleepless says:

    In fact, technophobia has been a staple of sf since the beginning. “There are things mankind is not meant to know.” Higher tech rarely thwarts a techno-threat.

Comments are closed.