Having a chat

By April 10, 2023Media, Society, Technology

ChatGPT is a natural language processing tool driven by Artificial Intelligence that allows a human to have human-like conversations and much more with the “chatbot”. The language model can answer questions and assist you with tasks like composing emails, essays, and code. Usage is currently open to the public free of charge because ChatGPT is in its research and feedback-collection phase. Since February 1st, 2023, there is also a paid subscription version called ChatGPT Plus1. At this stage, I opted for the freebie.

Over the last few days I happened to speak to a couple of people who do a fair bit of coding (in Python) in their paid employment and they both told me that they use ChatGPT to do their coding, with the humans at their end only checking the code. One of them told me that he obtained code from chatGPT in about thirty minutes which normally would have taken him a day to write. He went on to say that this will have a huge effect on the number of coders required. He suggested that only the best, most innovative coders, will be able to find jobs. He told me some of the things he thought ChatGPT could do, and I asked him if he could send me an e-mail listing some of the things it could do. He said, with a wry smile on his face that he would get ChatGPT to write me an e-mail to tell me what it could do. We both laughed, my cackle tinged with some embarrassment at my ignorance.

Most of the news media seem to be mostly concerned with the possibility that ChatGPT could be used by students to cheat in exams or essays2 or, horror of horrors, that it could be used to provide copy for news media3.

I decided to have a crack at this myself, so I joined OpenAI and asked some questions of ChatGPT4.

This is how it went:

Blot: I am just trying to obtain an understanding of the capabilities of ChatGPT and to work out if it will be of use to me.

ChatGPT: Hello! I’m ChatGPT, a large language model trained by OpenAI. I’m designed to understand and generate human-like responses to a wide variety of questions and topics. I have been trained on a large corpus of text, including books, articles, and web pages, and I have the ability to understand and generate text in many different languages. My capabilities include answering general knowledge questions, providing advice, generating creative writing prompts, assisting with language translation, and much more. However, please note that while I am capable of providing helpful information, I am not a substitute for professional advice, and I cannot guarantee the accuracy of all the information I provide. If you have any specific questions or tasks that you would like help with, feel free to ask and I’ll do my best to assist you.

Blot: I am a blogger and often write essays on topics with which I am largely unfamiliar, I thought that using ChatGPT might be able to give me a kick-start to such topics. Would that be something you would see as possible?

ChatGPT: Yes, absolutely! As a language model, one of my core capabilities is to generate human-like text on a wide range of topics. If you’re a blogger looking for ideas or inspiration on a topic you’re unfamiliar with, I can definitely help by providing some initial insights and ideas. You can ask me questions about the topic you’re interested in, and I can provide you with relevant information, statistics, and examples. I can also provide you with writing prompts and suggest possible angles for your essay. However, it’s important to keep in mind that while I can assist with generating ideas, I cannot write the essay for you. The final product will still be your own work, and it’s important to do your own research and fact-checking to ensure the accuracy of your writing. If you have any specific topics in mind, feel free to ask and I’ll do my best to assist you!

Blot: OK. I don’t intend to have you write anything for me, just to point me in the right direction and perhaps highlight aspects I may not have considered.

ChatGPT: I’d be happy to help! If you have a specific topic in mind that you’re interested in, feel free to let me know and I can provide you with some initial insights and ideas. Alternatively, if you’re not sure what to write about, I can suggest some popular or trending topics within your niche or industry to help you get started. As a language model, I have access to a vast amount of information and can provide you with statistics, examples, and perspectives on a wide range of topics. I can also help you narrow down your topic or suggest different angles to approach your writing. Let me know what you have in mind, and I’ll do my best to assist you!

Blot: OK. I’ll have a think about it and ask something when I think of something suitable. Thanks.

ChatGPT: You’re welcome! Take your time and feel free to reach out whenever you have a topic in mind or need some assistance. I’m here to help!

I still find it difficult to come to grips with the fact that the above conversation was between me, an ageing human, longer in the tooth than I care to mention, and a ‘computer’. When I started working, some decades ago, we had one minicomputer in the building and only the geophysicists used it to crunch numbers. We all had to do lessons on it but then we all promptly forgot those lessons. This command-line driven beast seemed to be useless for the sorts of things we did. Now here I am having a conversation, in English, with a machine. 

Sources

  1. https://www.zdnet.com/article/what-is-chatgpt-and-why-does-it-matter-heres-everything-you-need-to-know/
  2. https://www.abc.net.au/news/2023-01-26/chatgpt-sparks-cheating-ethical-concerns-in-schools-universities/101888440
  3. https://www.theguardian.com/commentisfree/2023/jan/24/chatgpt-artificial-intelligence-jobs-economy
  4. https://www.zdnet.com/article/how-to-use-chatgpt/

29 Comments

  • Arthur Baker says:

    Interesting, but not unprecedented in this forum. There is at least one ChatGPT response in an earlier discussion. It’s still there. Been there for quite a while, and if anyone realised it was bot-generated, they kept mum about it. If you can find it I’ll buy you a Mars Bar. If you can tell us the question that ChatGPT was responding to, I’ll buy you a packet of TimTams as well. (The double-coated ones in the blue packet are best, although you only get 9 biscuits instead of the usual 11. I think the extra chocolate more than makes up for getting fewer biscuits, but that’s just me.)

    • admin says:

      Arthur,
      As I am not fond of Mars Bars and Tim Tams are a bit too sweet for me, you’ll have to tell me.

      • Arthur Baker says:

        “You’ll have to tell me”. No I won’t “have to” do anything of the sort. You and Jon will just have to keep wondering. Blogs, eh? Marvellous, aren’t they?

        • admin says:

          Arthur,
          I don’t “have to” keep wondering. It may rankle, but there you go.

          • Arthur Baker says:

            “It may rankle, but there you go.”

            The day I get “rankled” by your responses is the day I … … well, never mind, wouldn’t want to provoke you into chucking another wobbly, would I?

          • admin says:

            Arthur,
            Being rankled seems to be a hobby of yours. Remember your months being rankled about the length of time the Liberals were in power?

          • Arthur Baker says:

            “Remember your months being rankled about the length of time the Liberals were in power?”

            Yay, that was great fun, wasn’t it? Incidentally, do you still believe that 8.7 equals 10? How does that affect your personal financial calcs and family budgeting? And how does the Australian Tax Office react when you unilaterally downgrade your annual income by 13%? Tax-time must be fun time at the Admin household. eh?

          • admin says:

            Arthur,
            It was simply bizarre on your part when you knew it was nothing to do with the point of the article. At the time, I gave you a list of real journalists with links to their articles saying precisely the same thing, and you replied ‘but they would ignore me’. It seemed to me that you just needed someone to bully.

          • Arthur Baker says:

            Right, so commenting on “the point of the article” is the only permitted response, and commenting on (or correcting) some detail of an article is forbidden in this forum and sometimes regarded as abuse. Thanks for clarifying that. Maybe you can also let us know what is your policy on commenters who differ from your point of view. That would be helpful.

          • admin says:

            Arthur,
            Cannot help yourself, can you? You seem to have purposely missed the point again. You stewed over this triviality for months but would not attack journalists who wrote precisely the same thing. I don’t have a problem with people disagreeing with me; in my game it happens all the time. It is when that disagreement gets down to personal abuse (which rarely happens in my game), that I won’t tolerate it. Life is too short.

        • Jon says:

          Your confidence suggests the one you are sure is AI is actually one of your own Arthur. 😉

          The one I’m thinking of was from a one time commenter. Don’t know whether BA can interrogate the blog for unique commenters in the last say 12 months??? but that would help me find the post I thought was bot generated.

          • admin says:

            Jon,
            I am not able to search for single commenters.

          • Arthur Baker says:

            G’day Jon. “Your confidence suggests the one you are sure is AI is actually one of your own Arthur.”

            Then find it, dear Henry, dear Henry, dear Henry.

            What happened to your previous thesis, the “middle of the road comment” from “a strange name” – “ie innocuous”? That wasn’t “one of my own”, was it?

            There’s a hole in your bucket, dear Henry, a hole.

          • Jon says:

            Nothing happened to my middle of the road interloper theory Arthur. I’ve had a quick browse but not uncovered anything like my vague memory suggested. Far too many posts and comments to wade through and no easy way to search so I doubt I’ll find it even if it exists. I did stumble upon a(nother) likely candidate though. It said something like – “this blog is very interesting and extremely well referenced.” Fits the bill closely I’d have thought.

            Can’t see any reason why two different posters couldn’t have used AI, but it seems a reasonable assumption that YOU could only be certain if you posted one yourself. Is my logic wrong?

          • Arthur Baker says:

            I did post it, Jon. If you can’t find it or distinguish it from a genuine real live subscriber’s post, that in itself sends all of us a message. What is that message? That’s for all of us to decide, in our own individual ways. Good luck with that.

          • Jon says:

            What’s the prize Arthur? “I don’t get out of bed for anything less than 10k.”

            I haven’t even glanced at the commentary around AI issues but it’s easy to see negatives such as how liars/scammers/cheating students/fraudsters etc, and propagandists/dishonest political leaders and enablers might use it, especially in collaboration with deep fake images and videos already doing the rounds. Given govt’s comatose responses to many digital issues, it’s probable that there won’t be any limits or legislation around its (mis)use for decades.

  • Jon says:

    Ask it to compile a list and brief summary of corruption under the Morrison government. Or, the hypocrisy of conservatives supporting dishonest and immoral/unethical politicians.

    Too many topics to go back to Arthur (I might eventually get round to looking) but I recall one in the last 6 months or so with a strange name attached which is a very likely candidate imo. Middle of the road comment – ie innocuous and without expressing an obvious opinion/leaning on the topic – and from memory he/she/it asked a question as though there was real interest yet didn’t bother to follow up.

    If AI bots were to be used openly and honestly then I’m sure there would be a way to embed an identifier to expose cheats (or store all requests and IPs), although crims and smarter computer operators would easily bypass that.

    Is there a code of use (like acknowledging sources, copyright, payment to original sources etc) surrounding chatbots? I suspect not – they will take whatever is accessible.

    Interestingly I saw a comment on a deals website where a lawyer was boasting about being able to charge $500/hr for basic legal work which drew a response from a savvy follower of AI advances to the effect of “enjoy it while you can because it will only be a few years before your skills will be largely redundant”.

    I’m not interested enough at this stage to do any research but prima facie it appears that AI has the capacity to concentrate wealth further AND cause real havoc to employment in many areas – as your programmer observed.

    • admin says:

      Jon,
      It will be interesting to see where it all ends up. After conversations with a few colleagues, I have a list of questions I will ask ChatGPT, and I will include yours. I have tried a few questions and the answers have been fairly straightforward and a bit bland. I might try a few more precise questions to see what comes up.

  • Mark Dougall says:

    Well that’s a shock. I assumed we were all AI bots.

    • Jon says:

      Might a name change – the Bot Report – be in the wind? I like that it’s the bizarro world version of the Bolt Report (presumably?).

      • admin says:

        Jon,
        That was the logic behind the name. To have a go at the Bolt malevolence, and to report on ‘blots’ on the political landscape.

  • Mark Dougall says:

    This blog is very interesting and extremely well referenced.

  • Mark Dougall says:

    If you believe in one of the various religions where we have been produced by some supernatural being then we are all artificial intelligence. I would suggest though that this is an example of RS. The opposite of AI. Real Stupidity.

  • Jon says:

    AI and lazy (USA) policing. Whooda thort that mix might be problematic?
    https://www.abc.net.au/news/2023-04-26/hidden-dangers-of-artificial-intelligence/102264038

    • admin says:

      Jon,
      Based on my experience with asking it questions on geology, I wouldn’t trust it to get anything correct.

  • Jon says:

    Daughter’s partner has been experimenting with ChatGPT. Got him to ask the age of Earth question. It gave the most common/accepted(?) accretion theory.

    • admin says:

      Jon,
      Yeah, that is what it did with the opal myth I asked it about. It said it all started from the Scott novel; which is the common belief but is probably incorrect.

Leave a Reply to Jon Cancel Reply

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

Bitnami