Like many cases of ChatGPT being blamed for (weird social issue), it seems difficult to get a complete sense of what these people would be doing in the absence of ChatGPT, but I thought these were...
Like many cases of ChatGPT being blamed for (weird social issue), it seems difficult to get a complete sense of what these people would be doing in the absence of ChatGPT, but I thought these were a bit novel in that they involve multiple people interacting in a sense with an AI rather than just one person talking away at an AI while spinning out into delusional psychosis. It would be very weird to have a third “person” involved in my marriage weighing in on matters.
I imagine that without ChatGPT, they would have to duke it out alone, with the confidence, compassion, stamina and patience of someone actually human. I read the article thinking it's being used...
I imagine that without ChatGPT, they would have to duke it out alone, with the confidence, compassion, stamina and patience of someone actually human.
I read the article thinking it's being used in place of something like a close loyal friend taking your side to console you. But even those friends will get tired of your shit. They will develop views of their own. They will tread carefully where they don't have a complete picture. They will know when you're being one-sided. They will be concerned with your relationship. They will be concerned with your mental health. They will have a mental health of their own to be concerned with.
The insidious thing is, if someone in a relationship did make a new friend and all that person did was validate and agree with no pushback, my first reaction would be "what are they manipulating...
The insidious thing is, if someone in a relationship did make a new friend and all that person did was validate and agree with no pushback, my first reaction would be "what are they manipulating this person in order to gain?". That's the behavior of a manipulative person trying to gain something from someone. Either money, access, or to "steal" that person for themselves. LLMs don't have motivations, but their system directive to be agreeable and validating still results in the same feedback loop that is used by manipulative people to destroy relationships.
I could understand it if the third party was a trained marriage counselor, but Chatgpt? I'm sure they fed some online copies of relationship advice into it but that's crazy trusting it to...
I could understand it if the third party was a trained marriage counselor, but Chatgpt? I'm sure they fed some online copies of relationship advice into it but that's crazy trusting it to negotiate marriage hardships.
One of my closest friends refuses to listen to anyone complain about people she doesn't personally know. If she's going to hear complaints, she wants her own experience of the person so as to have...
One of my closest friends refuses to listen to anyone complain about people she doesn't personally know. If she's going to hear complaints, she wants her own experience of the person so as to have a counterbalance to whatever the complaint of the moment is. It seems like ChatGPT would benefit from that sort of separate subjective experience, though I know it's not really possible with the models that currently exist.
Llms are replacing reasoning and critical thinking in the times you need them most because those are the hardest moments for us to get through. It’s a shortcut, but the problem is that in a...
Llms are replacing reasoning and critical thinking in the times you need them most because those are the hardest moments for us to get through. It’s a shortcut, but the problem is that in a relationship it’s the journey that matters (including inevitable arguments).
I'm really afraid of children also growing up with LLM-driven parental advice regardless of how sound the advice. People have been using LLMs for things that it's clearly not suited to do so and...
I'm really afraid of children also growing up with LLM-driven parental advice regardless of how sound the advice. People have been using LLMs for things that it's clearly not suited to do so and its basically impossible to tell what the long term consequences of it is going to be? Back when I nuked my Reddit account that was involved in mental health subreddits I did so in part because I figured it would result in unhealthy, toxic therapy-esque conversations. Despite having been correct about that, it seems like I may have actually underestimated just how bad it could get.
Recently I was at a conference. That particular conference was one I had been there before, one year ago. Since LLMs are a hot topic too(as it involved supportive software engineering), I was secretly hoping that my views on LLMs and other forms of what we call genAI(a term I dislike, but that's a story for another time) was more of an outlier than I thought. That I was surrounding myself in online bubbles and needed to address that instead.
Turns out, that was not the case - and some also brough up concerns that I hadn't heard of before but sounded valid to me. I really did not like that as many of the people there were better educated on the topic and also have been more directly involved in some of these negative consequences. It kind of indicates my opinion is more grounded than I like it to be.
This might be a bit of self-centered vent around the topic. Still, I really need to led this out. Given that I play Go I've followed machine learning on and off since 2016, mental health gets involved and my current field of software engineering is heavily affected by this. Despite not being at the center of it at all I guess it's difficult for me to not experience a lot of feelings about all this.
Like many cases of ChatGPT being blamed for (weird social issue), it seems difficult to get a complete sense of what these people would be doing in the absence of ChatGPT, but I thought these were a bit novel in that they involve multiple people interacting in a sense with an AI rather than just one person talking away at an AI while spinning out into delusional psychosis. It would be very weird to have a third “person” involved in my marriage weighing in on matters.
I imagine that without ChatGPT, they would have to duke it out alone, with the confidence, compassion, stamina and patience of someone actually human.
I read the article thinking it's being used in place of something like a close loyal friend taking your side to console you. But even those friends will get tired of your shit. They will develop views of their own. They will tread carefully where they don't have a complete picture. They will know when you're being one-sided. They will be concerned with your relationship. They will be concerned with your mental health. They will have a mental health of their own to be concerned with.
The insidious thing is, if someone in a relationship did make a new friend and all that person did was validate and agree with no pushback, my first reaction would be "what are they manipulating this person in order to gain?". That's the behavior of a manipulative person trying to gain something from someone. Either money, access, or to "steal" that person for themselves. LLMs don't have motivations, but their system directive to be agreeable and validating still results in the same feedback loop that is used by manipulative people to destroy relationships.
I could understand it if the third party was a trained marriage counselor, but Chatgpt? I'm sure they fed some online copies of relationship advice into it but that's crazy trusting it to negotiate marriage hardships.
One of my closest friends refuses to listen to anyone complain about people she doesn't personally know. If she's going to hear complaints, she wants her own experience of the person so as to have a counterbalance to whatever the complaint of the moment is. It seems like ChatGPT would benefit from that sort of separate subjective experience, though I know it's not really possible with the models that currently exist.
Llms are replacing reasoning and critical thinking in the times you need them most because those are the hardest moments for us to get through. It’s a shortcut, but the problem is that in a relationship it’s the journey that matters (including inevitable arguments).
I'm really afraid of children also growing up with LLM-driven parental advice regardless of how sound the advice. People have been using LLMs for things that it's clearly not suited to do so and its basically impossible to tell what the long term consequences of it is going to be? Back when I nuked my Reddit account that was involved in mental health subreddits I did so in part because I figured it would result in unhealthy, toxic therapy-esque conversations. Despite having been correct about that, it seems like I may have actually underestimated just how bad it could get.
Recently I was at a conference. That particular conference was one I had been there before, one year ago. Since LLMs are a hot topic too(as it involved supportive software engineering), I was secretly hoping that my views on LLMs and other forms of what we call genAI(a term I dislike, but that's a story for another time) was more of an outlier than I thought. That I was surrounding myself in online bubbles and needed to address that instead.
Turns out, that was not the case - and some also brough up concerns that I hadn't heard of before but sounded valid to me. I really did not like that as many of the people there were better educated on the topic and also have been more directly involved in some of these negative consequences. It kind of indicates my opinion is more grounded than I like it to be.
This might be a bit of self-centered vent around the topic. Still, I really need to led this out. Given that I play Go I've followed machine learning on and off since 2016, mental health gets involved and my current field of software engineering is heavily affected by this. Despite not being at the center of it at all I guess it's difficult for me to not experience a lot of feelings about all this.