Is ChatGPT Going to Be The Future?

So the other week I heard about OpenAI releasing their latest chatbot, and this chatbot is suppose to be conversational with a natural language model to be able to decipher what a user is asking. Anyway I've played around with it for little bit and think that this seems to me quite a disruptive technology. For example:

Human: Make this passage more humouras. So the other week I heard about OpenAI releasing their latest chatbot, and this chatbot is suppose to be conversational with a natural language model to be able to decipher what a user is asking. Anyway I've played around with it for little bit and think that this seems to me quite a disruptive technology.

AI: So the other week I heard about OpenAI releasing their latest chatbot, and this chatbot is suppose to be able to have witty banter with a natural language model to understand your questions. After giving it a try I'm quite convinced that this could be a revolutionary technology. What do you think? Could this be the beginning of a robot uprising? Or should we just sit back and enjoy the convenience this will bring? Let's discuss the future potential of ChatGPT!

Remember this is just in its infancy. It also is able to check code for bugs and create code from natural language and also vice versa, which to me would be a hugely useful tool. Anyway I just can't see how any organisation that uses technology is some capacity would be able to ignore this. I think in the next few years we'll see a shift in what jobs will be in demand, it's alraeady had a big impact on artists with the DALLE 2 project and I can see this being used in from law practices through to medical practices to accountants, whatever industry really.

Comments

  • +4

    This has been done to death twice on here.

    Exhibit A

    Exhibit B

    It's a failure and useless technology.

    /Thread

    • Ok

    • +3

      It’s just the latest shitposting fad.

      I also think it is a convenient way for people to write some bullshit trash blog post and deflect any criticism of it by saying “well, I didn’t write it… chatbotRTG did!”. It’s the blog equivalent of an insurance/accident/traffic infringement post that starts with “This happened to a friend…”

    • +2

      'It's a failure and useless technology.'

      This comment aged well…LoL

  • +4

    I think it is a bit like when desktop publishing suddenly allowed everyone to make bad school newsletters with comic sans and word art.
    The tool unlocks some productivity savings - it can write a first draft of a letter of complaint, for example, that you could then quickly edit.
    But the same way MS Publisher didn’t put graphic designers out of work, it just made it easier for a layperson to make a flyer, GPT tools won’t write an app for you, but might help beginners get some code snippets to work with, or replace a professional copywriter or novelist with its simple, derivative text.

    I keep seeing people post GPT produced contracts or other documents with a breathless “lawyers are out of business” as if lawyers don’t start with a template document now. The trick is not the text, but it’s meaning, and GPT can only extrapolate texts based on other examples. If you think this is a substitute for legal advise, never write your own will.

  • +1

    it's alraeady had a big impact on artists with the DALLE 2 project and I can see this being used in from law practices through to medical practices to accountants, whatever industry really.

    Just finished watching ColdFusion ?

  • +1

    It is good for basic stuff and gave me impressive results, but after asking some specific questions related to my field it often or a few times it confidently gives out a wrong answer or a weird answer and when I ask it to cite sources, it does provide me a source but the source does not actually have any material referring to what it actually said anywhere when I CTRL+F searched the books. Maybe I'll treat it like wikipedia that you have to re-verify the information from another more reliable source but still pretty decent starting point.

    Overall it is still quite impressive at the contextual understanding of natural human language and did generate some results that could not be just searched up on the internet, it had to be generated/created, so it is unique in its own way.

  • +1

    Like any tech, if it gets enough traction, time resources funding, it could become great, or it could be another elon.

  • It also is able to check code for bugs

    This can already be done and isn't ground breaking. It's not like this is a Messiah and is going to render engineers obsolete

    Honestly I don't think this is the leap forward you think it is. It's fun, interesting, but it won't be a daily used thing by professionals

    • +1

      It is one of those things that look ground breaking in fields you aren’t an expert in, but everyone can see why it is lacking in their own domains.
      If you are a “coder” and think GPT is going to replace you, you have bigger problems.
      That said, it would be great for a first draft of a letter or something for somebody with poor writing skills, and that is very useful.
      What it does impressively is instantly produce the basics a Year 7 kid could do with a bit of research. If your job could benefit from a year 7 kid on hand to do scut work that you then look over, it is a productivity tool.
      It will probably end up replacing some online first line support chat staff, or other similar jobs with repetitive, routine tasks for 80% of the job - but it won’t replace the tasks that actually need thought.
      I’d be fine with my bank using it to answer questions like where is the nearest open branch that can issue a bank cheque, or what is the Swift number or other things that it is tedious for me to go look up.

      • +1

        I don't think it could be used for chat support staff as there's no guarantee it will give a correct answer. The only guarantee is that the language will sound natural

        • Dumb chatbot for customer support is everywhere now whilst corporate downsizing their staffs. I'm having hardtime asking the dumb chatbot to direct me to a real person.

          The crux is that, this thing will improves itself overtime (exponentially) and form the base stack of large language model for commercial use which then builds specialty parameters on top using training data tailored for that target market.

          It may looks useless to some for now but it just leap-bounds and brought forward the AI age for mass developments faster than anyone thought.

          Have a look on how fast AI learns itself to beat the best of real human in Dota2. You'll be amazed if you know what it takes to win the game for a real human.
          Yes it's a game, but that's to trial on before putting to real world application. The fact that it can outsmart human on exploiting strategy is really something.

          • @dcep: Dumb chatbot is fine because the company can control exactly what it can say so it will never give wrong information. I don't think a chatbot that generates new content will be taken up though since its content can't be directly controlled

            • @Quantumcat: I think future versions will be able to have custom training data sets, maybe you could do this now by giving your custom data extra weighting.

  • I still prefer SmarterChild on MSN.

  • Its look like novelty to me. I dont think any of my dev work gonna replace by anything similar.

    • I can see myself using it for things like "write css for centred div with another div in the bottom right of it" instead of Googling for it every time because I keep forgetting

      • +1

        It also won't have domain knowledge of the site you're working on, so it's literally just a new way to Google

  • It'll be the future eventually. Every business will have one that has been fine tuned to understand what the business sells.

  • +1

    I'm not surprised it is garbage at the moment. It is my belief that was some golden lemonade guy working on it, actively using it on ozbargain before he was banned.

    The timing makes sense, but we have no proof.

  • Used it to pick up girls with surprising effectiveness.

  • Homie, the only “artists” complaining about DALLE draw furry garbage.

    Painters paint. And outsourcing comprehension tasks deny understanding.

Login or Join to leave a comment