5 ms·

> Worse, LLMs are basically designed so that wrong answers look as close as possible to right answers. This needs to be shouted from the rooftops. LLMs aren't

> Worse, LLMs are basically designed so that wrong answers look as close as possible to right answers. This needs to be shouted from the rooftops. LLMs aren't machines that can spew bullshit, they are bullshit machines. That's what makes them so dangerous.

rickydroll 2y ago
1) How is this any different from social media influencers especially those in health and wellness? 2) I would argue that LLM's are designed to give an answer that is the closest possible answer to the right answer without being a plagiarist. Sometimes this means they will give you a wrong answer. The same is true for humans. Ask any teacher correcting essays from students.
- hn_throwaway_99 2y ago
  > How is this any different from social media influencers especially those in health and wellness? If you are trying to make an argument that LLMs are exceptionally shitty at giving well-grounded advice, and instead excel at spouting asinine bullshit, congratulations, you succeeded.
  - rickydroll 2y ago
    Let me pose a hypothetical. What would your reaction be if the OP got the bad recommendation for garlic-infused oil from TikTok? That is a believable scenario. You would not be far wrong if it drove you to shout from the rooftops that TikTok (and social media in general) is a BS machine. Why is it worse for the bullshit to come from a program (LLM) than it is from a human on social media? if not, why single out LLMs for their bullshit?
    - Gazoche 2y ago
      I mean, both are bad. Just because one exists doesn't make the other better by comparison. However LLMs are just so much worse because 1. they can be asked anything, on any subject, at any time, at the click of a button 2. They have even less accountability than a human when stating falsehoods or dangerous advice 3. They're really good at generating statements that sound believable or authoritative, even when devoid of truth
      - rickydroll 2y ago
        All of your statements are true just like these are 1) humans can be asked anything on any subject at any time at the click of a button. The response time differs but we can also add a delay of hours for LLM's to respond and make it seem the same. 2) humans have even less accountability than an LLM when stating falsehoods or dangerous advice. At least in the US, humans can spout whatever bull shit they want under the guise of the First Amendment. LLM's can be censored legally because they have no such legal privilege. 3) humans are really good at generating statements that sound believable or authoritative even when devoid of the truth. See US conservatives, corporate disinformation, advertising, and fundamentalist religions One could make the case that humans created LLM's in their own image. We identify the erroneous outputs as bullishit or hallucination because that's what we do. If we did not have the capability of lying, hallucinating, or the term I prefer, confabulating, it would be difficult to recognize in LLM output. It's tough for us to imagine a world where other beings didn't lie or confabulate all the women saying that, I'm beginning to wonder if in LLM is such a creature. It gives us confabulations because it doesn't know they exist and can't modify its behavior.
        nineteen999 2y ago
        > humans have even less accountability than an LLM when stating falsehoods or dangerous advice Where on earth are you getting this? As a human in most developed countries you can be very much held accountable for lying under oath. Other times you can be held accountable for lying: 1) when involved in an police investigation 2) when you are under arrest 3) when you are being lawfully detained by the police 4) if you are being asked about a crime you witnessed 5) when filing a police report 6) many kinds of financial disclosure documents (eg. taxes) 7) when you're advertising something 8) if you're committing fraud

rickydroll 2y ago

1) How is this any different from social media influencers especially those in health and wellness? 2) I would argue that LLM's are designed to give an answer that is the closest possible answer to the right answer without being a plagiarist. Sometimes this means they will give you a wrong answer. The same is true for humans. Ask any teacher correcting essays from students.

hn_throwaway_99 2y ago
> How is this any different from social media influencers especially those in health and wellness? If you are trying to make an argument that LLMs are exceptionally shitty at giving well-grounded advice, and instead excel at spouting asinine bullshit, congratulations, you succeeded.
- rickydroll 2y ago
  Let me pose a hypothetical. What would your reaction be if the OP got the bad recommendation for garlic-infused oil from TikTok? That is a believable scenario. You would not be far wrong if it drove you to shout from the rooftops that TikTok (and social media in general) is a BS machine. Why is it worse for the bullshit to come from a program (LLM) than it is from a human on social media? if not, why single out LLMs for their bullshit?
  - Gazoche 2y ago
    I mean, both are bad. Just because one exists doesn't make the other better by comparison. However LLMs are just so much worse because 1. they can be asked anything, on any subject, at any time, at the click of a button 2. They have even less accountability than a human when stating falsehoods or dangerous advice 3. They're really good at generating statements that sound believable or authoritative, even when devoid of truth
    - rickydroll 2y ago
      All of your statements are true just like these are 1) humans can be asked anything on any subject at any time at the click of a button. The response time differs but we can also add a delay of hours for LLM's to respond and make it seem the same. 2) humans have even less accountability than an LLM when stating falsehoods or dangerous advice. At least in the US, humans can spout whatever bull shit they want under the guise of the First Amendment. LLM's can be censored legally because they have no such legal privilege. 3) humans are really good at generating statements that sound believable or authoritative even when devoid of the truth. See US conservatives, corporate disinformation, advertising, and fundamentalist religions One could make the case that humans created LLM's in their own image. We identify the erroneous outputs as bullishit or hallucination because that's what we do. If we did not have the capability of lying, hallucinating, or the term I prefer, confabulating, it would be difficult to recognize in LLM output. It's tough for us to imagine a world where other beings didn't lie or confabulate all the women saying that, I'm beginning to wonder if in LLM is such a creature. It gives us confabulations because it doesn't know they exist and can't modify its behavior.
      - nineteen999 2y ago
        > humans have even less accountability than an LLM when stating falsehoods or dangerous advice Where on earth are you getting this? As a human in most developed countries you can be very much held accountable for lying under oath. Other times you can be held accountable for lying: 1) when involved in an police investigation 2) when you are under arrest 3) when you are being lawfully detained by the police 4) if you are being asked about a crime you witnessed 5) when filing a police report 6) many kinds of financial disclosure documents (eg. taxes) 7) when you're advertising something 8) if you're committing fraud