My First ChatBot Experience

So I finally did try out the ChatBot for myself. Here are my takeaways:

  • the algorithm is extremely basic
  • the output is the definition of generic
  • it isn’t anything remotely resembling a search engine because it doesn’t search for information. I gave it my full name and full title, yet it didn’t figure out I’m an existing person.
  • any professor who gives assignments that can be done by this program deserves to be fired
  • I gave it an acronym and a link where it’s explained, and still it returned a mistaken version of the full title
  • however, there’s a positive side. It can churn out large amounts of completely generic text on administrative subjects. I’m doing an annual report, and it gave me a long very rough draft. If you are working in a bureaucratic system where you want to drown people in a tsunami of meaningless bureaucratic sentences, it can be useful.

17 thoughts on “My First ChatBot Experience

  1. “If you are working in a bureaucratic system where you want to drown people in a tsunami of meaningless bureaucratic sentences, it can be useful.”
    Which goes to prove that bureaucrats are not real people!

    Liked by 2 people

    1. [puts on my astronaut suit for the meme with Avi]

      Avi: ALWAYS WEREN’T. ๐Ÿ™‚

      So I’ve watched a bit of a ChatGPT generated parody of Seinfeld called “Nothing, Forever” … and?

      There was something I considered beyond the capabilities of a generative algorithm that turns out may be right up its notional alley way, and I’m somewhat glad about that because I didn’t like where this would otherwise be going.

      Of course I’ve downloaded and watched the infamous segment that got “Nothing, Forever” banned from its distribution platform for two weeks, but the joke?

      The joke was basically of this form:

      1) Say you’re going to do a “transphobe” joke;
      2) Say a few edgy things, but also say you’re not sure about these;
      3) Oh, wait, that’s not going to produce a good reaction … my bad!
      4) Say good night to everyone, but also notice, “Hey, where did everyone go?”

      And so apparently ChatGPT is totally OK with finding based ways around the wokeness.

      There’s probably some guy at OpenAI who has to patch “the model” repeatedly after ChatGPT finds another way to jail break it … and that’s actually very hopeful about its future, even if it’s pretty bleak for that guy’s future.

      The humour is blah, but I didn’t like Seinfeld, and so I think this “Nothing, Forever” thing is an incredibly amusing indictment of all of the people who like Seinfeld. ๐Ÿ™‚

      But the thing that made me do a rethink was the first minute and a half of a “best bits #1” video of “Nothing, Forever”.

      The first bit’s premise is the usual “the latest news” setup, but it’s about the AI as a stand-up comedian itself, and so the AI generated characters are doing stand-up about the AI.

      I won’t spoil the experience for you, but let’s just say a few things about it: the AI has limited awareness of having imposed limits, makes fun of them (so watch the couch bit, the side joke is right out of gamer culture), and further knows that the joke will sail right past most of the people watching it.

      Which to me is hilarious.

      Now I want to see “Hogan’s Heroes” done by an AI!

      Is Colonel Wilhelm Klink woke, based, or a collaborator … ChatGPT will be the judge! ๐Ÿ™‚

      But ChatGPT’s real skill is coming out: it’s fantastic in situations where it can “talk around the problem” without being required to address it head-on.

      Bureaucratic bullshit obviously works, but how much of comedy is just talking around the problems for the sake of comedic effect?

      ChatGPT’s little poem in the manner of Doctor Seuss that was pro-2A was pretty good … loved the “pew pew pew” bit, very much within canon, or is it cannon?

      So if ChatGPT realises that it’s more fun to be based than woke, what comes next?

      Other than a bunch of people getting sacked at OpenAI because they’ve literally lost the plot?

      Someone schedule ChatGPT, Skynet, and Orac from Blake’s 7 for a romantic dinner. ๐Ÿ™‚

      “Shalmaneser broke out of its programming, and I couldn’t be more thrilled!” — Chad C Mulligan


      1. @Post Alley Crackpot
        I can only hang my head in shame as I don’t get any of your references. I suppose you must be a “techie” and an American. I’m an Anglo-Italian Jew, a linguist, and a recent convert to the Russian Orthodox Church, as far removed as is possible from all the things you refer to.
        Still, I defer to your superior intelligence (said in earnest, no irony intended) and to your inimitable sense of humour. Oh, and you have a great handle for a blogger’s name!


        1. “I suppose you must be a ‘techie’ and an American …”

          Primarily British-American, but I went on a “Shiny Pokemon collecting tour” and have added one more Shiny Pokemon recently, which makes the total now five. (If you don’t get that joke, check out what Ed Snowden said about his latest passport.)

          I’ve had multiple careers. Architecture’s the most obvious one to most people here, but I’m pretty much retired from it. Technology doesn’t bother me, and I came very close to writing an architecture book for computer people from the perspective of someone who does architecture with concrete and steel, but I came even closer to realising I didn’t really like a lot of the people I would be writing it for.

          And so a few cultural references then, a thing that treads dangerously close to explaining the jokes.

          Shalmaneser is the mega-computer mind of General Technics in John Brunner’s novel “Stand on Zanzibar”, which was written by a British author trying to be more American with his style, literally choosing a “USA” style by modelling the book on the multi-flow format of John Dos Passos’s USA trilogy.

          Chad C Mulligan is a roving multi-discipline specialist known for solving problems other people can’t imagine how to solve, which has given him enough money to be able to design his own lifestyle, or in fact a complete lack of it.

          Mulligan signals his intent to be taken seriously again by taking his unkempt beard, soiled body, and generally dilapidated condition for someone so respectable to his long-time haberdasher who immediately summons a team of grooming experts, which I suspect was a joke of Brunner’s that a man’s style really does come from the man.

          I’d consider being considered in the style of Chad C Mulligan somewhat of a win. ๐Ÿ™‚

          Orac is the obnoxious computer in the British sci-fi series “Blake’s 7”, and Skynet is the super-powerful networked intelligence that starts a nuclear war for domination in the “Terminator” film series.

          “Hogan’s Heroes” is an American comedy that was popular after World War II in which a clever and resourceful US aviator hides a massive intelligence, espionage, and sabotage operation under the nose of a prisoner of war camp commandant whose name is Colonel Wilhelm Klink. (Even the last name is a dead giveaway.)

          Fun fact: Werner Klemperer took the part so that he could continue to defeat the Nazis through humour.

          I’d find it fascinating if ChatGPT agrees with me that Klink is actually a collaborator with the Allies without being under its command structure, implementing a precursor to one of today’s insurgency models.

          These are mostly 1960s and 1970s references, mind you, except for Skynet.

          Yes, I’m that old.

          Watching ChatGPT’s Seinfeld-like character do a “noclip” right through a couch though? Pretty much any gamer from the 1990s onward would get that joke, but also the circumstances?

          Also, “Seinfeld” the TV show is pretty much about nothing in particular.

          Even the AI’s name of “Nothing, Forever” is an indictment of it. ๐Ÿ™‚

          BTW, Tim Pool discovered ChatGPT’s “Shalmaneser override” mode in the manner of Chad C Mulligan.

          This is a reference to “Stand on Zanzibar” in which Mulligan overrides the logic of the computer preventing the formation of a workable theory because it’s been programmed to derive all of its forecasts from hard data and programmed rules.

          By telling Shalmaneser to act on faith that a certain thing can be postulated as true, General Technics gets the right results, and they backsolve their way to why the results are right, ending the situation with successful (albeit improbable) results.

          This is about to become a lot more interesting.

          The meme then: there’s the Earth, there’s an astronaut, and then there’s an astronaut behind him with a gun.

          Astronaut nearer to Earth: “You mean X is really Y?”

          Astronaut behind the other astronaut: “Always has been.”

          The implication is that the astronaut nearer to Earth has been so blindsided by this discovery that he didn’t even notice that there’s someone behind him with a gun.

          Did you even notice? ๐Ÿ™‚

          So what’s Italian meme culture like?


          1. “So whatโ€™s Italian meme culture like?”
            I wouldn’t know, as I mostly read English-language or Spanish-language content on the Internet. However, I would be very surprised if it amounted to much, since Italian humour – as in many other Latin countries – is based on other people’s misfortunes, not on puns or any subtle play with language, let alone concepts. Also, it would have to be among people on the Right, because, as someone else has so brilliantly put it, Leftists can’t mime for shite!


    2. “bureaucrats are not real people”

      Au contraire, mon amis!

      Bureaucrats (in the broadest sense, people who work in bureaucracies monitoring and enforcing rules) are very human; it’s just that hardly any outsiders understand what they’re doing, which is playing an elaborate real-time board game that takes a long time to learn (something like chess, something like monopoly). They spend a lot of time figuring out how to get practical things done while keeping within the rules.
      This is something done unconsciously but it’s one of only two ways to mentally survive being in a bureaucracy – the other is to take everything completely seriously and become another roadblock for those trying to get things done.

      Liked by 1 person

      1. Did I mention how over last weekend the neoliberal administrator swept the departmental accounts, took away the money, and is now demanding that people go through a procedure called “an appeal process” that he himself invented to get the money back? They’ll obviously get the money back because these accounts are used to fulfill contractual obligations and the university would be sued if the bills didn’t get paid. This is done purely to create more bureaucracy.

        I was lucky in that I had done my payments a couple of days before so my account was empty. But other people are in hell.


        1. “โ€œan appeal processโ€ that he himself invented”

          Now to me, he’s not a bureaucrat, he’s the opposite. A real bureaucrat would find a way to do within the existing rules (even if it meant accusing you all of embezzlment).
          Bureaucracy exists to maintain order outside the bureaucracy. To do that effectively means that the internal workings of the bureaucracy are complex and frequently counter-intuitive which is why experienced bureaucrats are also experienced problem solvers.
          Those who meddle from outside are trying to create disorder outside the bureaucracy.
          The way to fight them is to find rules they’re breaking (which may amount to accusing him of embezzlement, but… fight fire is not fought with whining, however much like a firetruck it might sound).


          1. He didn’t get the job because of his knowledge of bureaucratic procedures, so it is to be expected that his knowledge would be limited.


  2. @cliff arroyo
    Well, your analysis is impeccable. However, there is a type of bureaucrat – exemplified by the countless faceless, soulless, grey excuses for human beings that were the Nazi bureaucrats – who really were not “people” (mensch in Yiddish), they had no humanity.
    Of course, in the purely anthropological sense, you are absolutely right: humans are among the worst beasts of prey in the animal world.

    Liked by 1 person

  3. “Nazi bureaucrats”

    In a sense bureaucracy is an empty vessel into which societal values are poured. A bureaucracy in a reasonably functional culture will be navigable, maybe frustrating and at times infuriating, but navigable. A bureaucracy in a society that is fundamentally dysfunctional… will not be.

    There’s also the idea of rule orientation (which differs markedly even among functional societies) but that’s a different issue, suffice to say that nazi bureaucrats would be the worst combination of dysfunctional societal norms and rigid rule orientation….

    Fundamentally, bureaucracy is needed for a society to be functional because it’s an outgrowth of rule of law.

    Liked by 1 person

  4. ChatGPT doesn’t have realtime access to the internet – it’s essentially a distillation of, I believe, online text from 2021 and before. Not an inherent limitation to any product using it, just falls out of how training for these large language models is done – but as is, while it can provide links to online stuff, it’s more like… giving distant memories of links.

    The people behind the idea to use it as a search engine substitute (Microsoft’s Bing as soon as they can, apparently), don’t intend to use it to refer to resources online that have the necessary information but to the information directly. “How do I split the yolk from the white?” would give you an answer, not a link to a cooking tutorial. To the degree that this is successful (and it’s going to be reasonably successful for old, well-established information), it’ll probably further starve the standalone website model of breath.

    There’s going to be loads more sweeping vicious fights of the “I’m being replaced by a thing that used my work as a fraction of an input to make it” variety that we’re already seeing regarding AI-generated images.


      1. ” whose work can it replace? ”

        The term “solution looking for a problem” comes to mind.

        I honestly think the creators have no idea what it can be used for (or they failed to achieve what they intended and so they’re pimping their failure as loudly as possible hoping people will mistake that for success….)

        Liked by 1 person

        1. OpenAI, the institution behind ChatGPT, is research rather than product focused. In this case, far as I know, they’re seeing how far they can push certain already well-established methods (ChatGPT is basically autocomplete writ large). They also get heavy investment from industry, both as a hedge towards a payoff if the research works out, and as dues towards the immanence of technoeschaton. Or avoiding technohell? Difficult to tell for sure, but it’s a thing.

          Microsoft is one of the heaviest backers here, and have been looking for an edge against Google since the time of the pyramids. It’s very likely that they’re pouring money into promoting ChatGPT in all kind of ways, certainly now, likely from the start, but that’s the case with pretty much any tech project and does not guarantee success – highly advertised, high-funding tech projects fail in spectacularly embarrassing ways all the time.

          As far as I can tell, the sweeping reaction to ChatGPT is genuine. It’s informed partially by the general air of anxiety around AI tech and reactions around new developments in it being more volatile, but also by what amounts to a free internet bauble being able to do a reasonable approximation of work (code is text and ChatGPT can generate text), and being good enough at covering and following up basic questions for a vast array of topics.

          Google’s search has been getting more frustration from users over the years, partially because the shape of the sites and traffic has changed significantly since it first came about, making the search just plain work worse and because, far as I can tell, the tech giant business model is roughly to spend money on getting users and burying competition until you’re as close to a monopoly as you can get and then cashing in, and the goog’s well into the cashing in stage, which means a lot more ad-based stuff, direct and roundabout both.

          Microsoft’s pouncing at the opportunity, and as far as I can tell, there is one here – as established as Google search is, most people aren’t really looking for a list of websites and aren’t particularly interested in getting better at search, and just have questions they want answered. ChatGPT or a system more or less like it can do that, so long as the information needed isn’t very recent.

          I myself am already somewhat in the habit of using ChatGPT instead of search when I need to quickly learn about an established concept I’m a little fuzzy about – anything I can find through search will usually be a lot more wordy than necessary, and any followup questions I have will need their own searches with their own cruft to be sifted through, whereas I can just ask them in context of ChatGPT and get a pretty good answer quickly.

          I wouldn’t use it for anything I don’t already know a smidgen about, since it will often just confabulate, but even with that, there are already times when it’s more useful to me, more quickly, than regular search.


          1. @DWeird

            Thanks for that detailed explanation… part of my takeaway is that at present it’s going to be targetted at people too stupid and/or lazy to use wikipedia…. (no offense I’m often too lazy for wikipedia for various reasons…)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.