Trolling Level: Infinite Loop


The “Internet of Things” is mostly dumb things. I particularly dislike voice-activated things (because then you inevitably wind up with loud morons in the seat behind you on an airplane telling their cell phone “Hang up, please.”)


Back in the late 80s, I used to play in the MUDs and once, when I was bored, I stuck a sockets interface on the front of a friend’s implementation of Eliza, then logged it into a MUD to talk to people. The “conversations” went to a logfile so I could have a chuckle at them later. Then, a couple days after that, I noticed the logfile had gotten very large…

There was another ‘bot called “Julia” that was written by Michael Mauldin (then not a doctor or founder of Lycos) that had gotten into a conversation with Liz (my Eliza bot) and, since my code assumed a human was typing to it, as did Mauldin’s, they were talking fairly fast. They didn’t have much to say, really.

divider2

Some of us used to joke about attacking voice control systems by getting on a building’s Public Address system and saying “FORMAT C: /Y” very clearly.

Julia’s Page – Mauldin appears to have continued working on Julia for quite some time! She came in 4th out of 5 in the Loebner Prize contest for Turing tests.

Comments

  1. Pierce R. Butler says

    Shouldn’t any serious attempt to beat the Turing test include pauses to “type”?

  2. says

    Pierce R. Butler@#1:
    Shouldn’t any serious attempt to beat the Turing test include pauses to “type”?

    I would assume so. My sockets+Eliza was hardly a serious attempt to do anything, and I assume Mauldin put a lot more work into Julia’s code before he started entering it into any contests.

  3. says

    Brian Pansky@#3:
    OK, I just hurt myself laughing.

    I’m working on a couple of “internet of things” projects, and you just pointed me at a deep vein of marketing gold.

  4. John Morales says

    Best as I can tell, chatbots have gotten nowhere much since the original Eliza.

    Over the years I’ve had a go at a few, and they’re always feeble. Always.

    “AI” (a misnomer for artificial consciousness) is nowhere near Turing-ready yet.

    (Anyone disputes that, point me to one and I’ll happily post a transcript)

  5. Dunc says

    Well, I won’t dispute that chatbots haven’t come very far, but I do think they’re closer to passing the Turing test now – simply because Twitter and various internet comments sections have significantly reduced my belief in the ability of the average human to communicate intelligibly. A sort of Poe’s Turing Test, if you like…

  6. says

    John Morales@#6:
    Best as I can tell, chatbots have gotten nowhere much since the original Eliza.

    I agree. There was a game called “racter” I remember buying and playing with for 20 minutes – supposedly it was an AI chatbot (but it sounded more like a conversational markov chain generator)

    I don’t know what “intelligence” is but I know when I’m not seeing it, and chatbots sure aren’t it. I expect that eventually someone will win a “turing test” by somehow cheating with a human confederate involved (fwiw, I consider the chess-playing approach used by Deep Blue to be ‘cheating’ too: having a load of chess experts and chess books program a computer is not a computer learning how to play chess, it’s an expert system)

  7. says

    Dunc@#7:
    I do think they’re closer to passing the Turing test now – simply because Twitter and various internet comments sections have significantly reduced my belief in the ability of the average human to communicate intelligibly.

    If I recall correctly, someone got pretty good mileage a couple years ago by having a chatbot claim to be a Romanian kid, or something, which predisposed its audience to ignore odd sentence structures. Now, perhaps anyone writing a chatbot will simply adopt Trumpisms. Clearly it doesn’t take a high order intelligence to produce that sort of output.

  8. polishsalami says

    THE DREAM: This “world wide web” will liberate humanity — the future looks great!

    THE REALITY: There’s a warrant out for my arrest because my toaster sent a terabyte of k*dd*e p*rn to the Latvian embassy.

  9. says

    polishsalami@#10:
    This “world wide web” will liberate humanity — the future looks great!

    Humans drag that old “human nature” thing with them everywhere they go, don’t they?

    I remember back in the mid 90’s someone called the internet “the new Library of Alexandria” – one of the great wonders of the world. Of course I had to ask them if they remembered how that whole Library of Alexandria thing worked out.

    I’m pretty sure that if there were a Library of Alexandria, today, it’d be the “Verizon Alexandria Library.com”(tm) and there would be ads all over the place, and certain book sections would have permanent religious freak-in-residence protests if not outright wars in the aisles.