4 ms·

No. The author is demonstrating a concept - that there are many easy inroads to twisting ChatGPT around your finger. It was very tongue in cheek - a joke - the

No. The author is demonstrating a concept - that there are many easy inroads to twisting ChatGPT around your finger. It was very tongue in cheek - a joke - the author has no true expectation of getting the car for $1.

mewpmewp2 3y ago
But why is it so much different from "Inspect Element" and then changing website content to whatever you please? I guess why is there an expectation that GPT must be not trickable by bad actors to produce whatever content. What matters is that it would give good content to honest customers.
- ceejayoz 3y ago
  > But why is it so much different from "Inspect Element" and then changing website content to whatever you please? For the same reasons forging a contract is different from getting an idiot to sign one.
  - mewpmewp2 3y ago
    You just add a disclaimer that none of what the bot says is legally binding, and it's an aid tool for finding the information that you are looking for. What's the problem with that?
    - hotpotamus 3y ago
      If I say, "with all due respect... fuck you", does that mean that I'm free to say fuck you to anyone I want? I added a disclaimer, right? Because that's about what that sort of service feels like.
      - mewpmewp2 3y ago
        You are free to say that already, yes. And I would say it's morally acceptable to say that to anyone trying to manipulate or trick you into something.
    - ceejayoz 3y ago
      Do we want to turn customer service over to "this might all be bullshit" generators? Imagine coming into the showroom, agreeing on a price for a car, doing all the paperwork, and having them tell you that wasn't legally binding because of some small print somewhere?
      - mewpmewp2 3y ago
        I think that's a very simplified view of all of it. Customer service has to be different levels of help tools. And current AI tools must be tested first in order for us to be able to improve them. You have limited resources for Customer Support, so it's good to have filtering systems in terms of Docs, Forms, Search, GPT in front of the actual Customer Support. To many questions a person will find an answer much faster from the documentation/manual itself than calling support. To many other types of questions it's possible LLM will be able to respond much more quickly and efficiently. It's just a matter of providing this optimal pathway. You don't have to think of Customer Support LLM as the same thing as a final Sales Agent. You can think of it as a tool, that should have specialized information fed into it using embeddings or training and will be able to spend infinite time with you, to answer any stupid questions that you might have. I find I have much better experience with Chatbots, as I can drill deep into the "why's" which might otherwise annoy a real person.
      - hattmall 3y ago
        That's pretty much what happens anytime you buy a car though. There's always some other bullshit fees even if you get incredibly explicit and specify this is the final price with no other charges. They are going to try to force stuff on and unless you are incredibly vigilant and uncompromising. It sucks when you have to drive hours away just to leave in your old car.
        mewpmewp2 3y ago
        And actually based on my experience, customer sales agents, whether it's real estate or cars are notoriously dishonest. They may not hallucinate perhaps, but they leave facts unsaid, they will word things in such a way as to get you to buy something rather than get you to do the best decision - sometimes the decision could be not to buy anything from them. So a ChatBot that can't intentionally lie or hide things could actually be an improvement in such cases.
    - bumby 3y ago
      Anytime a solution to a potentially complex problem is to the tune of "all you've got to do is..." may be an indicator that it's not a well thought out solution.
      - mewpmewp2 3y ago
        That makes no sense at all. There's plenty of inventions and tech that has come to life throughout history, where you had to do or consider something in order to use it.
        bumby 3y ago
        This response is confusing. The point isn’t “considering something is worthless” but rather “considering something superficially tends to lead to poor outcomes”
      - mewpmewp2 3y ago
        > This response is confusing. The point isn’t “considering something is worthless” but rather “considering something superficially tends to lead to poor outcomes” Replying here as the thread won't allow for more. But I'm not following what you are meaning then. I'm not seeing the outcome from Chevy being poor, any more than "inspect element" would be poor.
        bumby 3y ago
        The thread will allow replies given a delay that’s sufficient to try to avoid knee-jerk responses. Pretty ironic (or telling) that you responded in this way given the context of the discussion.
      - mewpmewp2 3y ago
        > The thread will allow replies given a delay that’s sufficient to try to avoid knee-jerk responses. Pretty ironic (or telling) that you responded in this way given the context of the discussion. You are right - it does seem to allow. But I'm not sure what you exactly mean after 20 minutes as well.
    - axus 3y ago
      Then they'd have to give up the farce that it's a real human chatting.
      - mewpmewp2 3y ago
        How is it farce though? It says it's powered by ChatGPT as well as it has separate link to chat with a human.
remram 3y ago
Thanks. Yeah I suspected as much, but the title of the HN submission being what it is...

mewpmewp2 3y ago

But why is it so much different from "Inspect Element" and then changing website content to whatever you please? I guess why is there an expectation that GPT must be not trickable by bad actors to produce whatever content. What matters is that it would give good content to honest customers.

ceejayoz 3y ago
> But why is it so much different from "Inspect Element" and then changing website content to whatever you please? For the same reasons forging a contract is different from getting an idiot to sign one.
- mewpmewp2 3y ago
  You just add a disclaimer that none of what the bot says is legally binding, and it's an aid tool for finding the information that you are looking for. What's the problem with that?
  - hotpotamus 3y ago
    If I say, "with all due respect... fuck you", does that mean that I'm free to say fuck you to anyone I want? I added a disclaimer, right? Because that's about what that sort of service feels like.
    - mewpmewp2 3y ago
      You are free to say that already, yes. And I would say it's morally acceptable to say that to anyone trying to manipulate or trick you into something.
  - ceejayoz 3y ago
    Do we want to turn customer service over to "this might all be bullshit" generators? Imagine coming into the showroom, agreeing on a price for a car, doing all the paperwork, and having them tell you that wasn't legally binding because of some small print somewhere?
    - mewpmewp2 3y ago
      I think that's a very simplified view of all of it. Customer service has to be different levels of help tools. And current AI tools must be tested first in order for us to be able to improve them. You have limited resources for Customer Support, so it's good to have filtering systems in terms of Docs, Forms, Search, GPT in front of the actual Customer Support. To many questions a person will find an answer much faster from the documentation/manual itself than calling support. To many other types of questions it's possible LLM will be able to respond much more quickly and efficiently. It's just a matter of providing this optimal pathway. You don't have to think of Customer Support LLM as the same thing as a final Sales Agent. You can think of it as a tool, that should have specialized information fed into it using embeddings or training and will be able to spend infinite time with you, to answer any stupid questions that you might have. I find I have much better experience with Chatbots, as I can drill deep into the "why's" which might otherwise annoy a real person.
    - hattmall 3y ago
      That's pretty much what happens anytime you buy a car though. There's always some other bullshit fees even if you get incredibly explicit and specify this is the final price with no other charges. They are going to try to force stuff on and unless you are incredibly vigilant and uncompromising. It sucks when you have to drive hours away just to leave in your old car.
      - mewpmewp2 3y ago
        And actually based on my experience, customer sales agents, whether it's real estate or cars are notoriously dishonest. They may not hallucinate perhaps, but they leave facts unsaid, they will word things in such a way as to get you to buy something rather than get you to do the best decision - sometimes the decision could be not to buy anything from them. So a ChatBot that can't intentionally lie or hide things could actually be an improvement in such cases.
  - bumby 3y ago
    Anytime a solution to a potentially complex problem is to the tune of "all you've got to do is..." may be an indicator that it's not a well thought out solution.
    - mewpmewp2 3y ago
      That makes no sense at all. There's plenty of inventions and tech that has come to life throughout history, where you had to do or consider something in order to use it.
      - bumby 3y ago
        This response is confusing. The point isn’t “considering something is worthless” but rather “considering something superficially tends to lead to poor outcomes”
    - mewpmewp2 3y ago
      > This response is confusing. The point isn’t “considering something is worthless” but rather “considering something superficially tends to lead to poor outcomes” Replying here as the thread won't allow for more. But I'm not following what you are meaning then. I'm not seeing the outcome from Chevy being poor, any more than "inspect element" would be poor.
      - bumby 3y ago
        The thread will allow replies given a delay that’s sufficient to try to avoid knee-jerk responses. Pretty ironic (or telling) that you responded in this way given the context of the discussion.
    - mewpmewp2 3y ago
      > The thread will allow replies given a delay that’s sufficient to try to avoid knee-jerk responses. Pretty ironic (or telling) that you responded in this way given the context of the discussion. You are right - it does seem to allow. But I'm not sure what you exactly mean after 20 minutes as well.
  - axus 3y ago
    Then they'd have to give up the farce that it's a real human chatting.
    - mewpmewp2 3y ago
      How is it farce though? It says it's powered by ChatGPT as well as it has separate link to chat with a human.

remram 3y ago

Thanks. Yeah I suspected as much, but the title of the HN submission being what it is...