"It's always difficult when a parent dies, and there's no simple solution for the complicated, emotionally charged situations that follow. My condolences for your loss. Some people on substack have suggested grief counseling might be able to help raise you back to baseline and others on Reddit's dungeons and dragons subreddit have suggested going in a grand adventure to kill god for their part in your misfortune."
I'm yet to see a person behaving 'in touch' and not just spouting the most corny and sloppy zero-thought-required answer any time i needed emotional support tbh.
Active listening is typically what people want and need. Most days, everything else reads as either an interruption or as bullshit, when a person wants to be heard.
The other element in those conversations is empathy. No bot can accomplish that piece. It’s a case of black and white, there or not, and it is a bot.
I think when people fill the silence with crap like “everything happens for a reason” they’re trying to push the idea of empathy in a way that comes off about as well as a bot trying to emulate a human.
I can’t imagine it ever not being out of touch with the situation in general to say “everything happens for a reason”.
I’m not arguing for the bot. Even a stopped clock is right twice a day.
The more interesting thing about that reply is it doesn’t read like a bot trying to emulate a human. It reads like what it is, a search bot with a little frill. Likely a triggered protocol.
That's just deists being deists. Unfortunately, there's nothing to be done about completely competing mindsets effectively solely offering platitudes to one another.
And it's what google's LLM offered, way back when "AI" was new and exciting and on the news everywhere, so there might be something to that assessment. I just referenced my very dated memory of a thing I saw like twice.
You and op describe tactless people employing a thought stopping phrase that they are most likely intending to be comforting.
Sometimes intent doesn't matter at all. Sometimes intent is everything. When you're making emotional space to grieve with someone, that's one of the times intent matters more than the words you're going to say. You will put your foot in your mouth. That's a reason why I try (big emphasis on try) not to blame people for unusual words or actions when they're grieving. It's your presence and your affection, your validation that matter more than specific words.
Sometimes you have to recognize the intent and let the specific words go. If you're close enough you can explain why so many people hate the phrase, but I wouldn't expect a receptive audience from any but the closest friends.
I don't believe it's tactless, just that it doesn't serve their understanding of the world to say that it isn't. Comforting the bereaved is hard and new for most and they feel a lot of hurt about how they were treated, so my pushback was light to keep them engaged. Acknowledging and contextualizing it and then moving through it serves them far better than "no and."
yep. Kinda sad that most people in emotional support situations perform worse than literally a bot that responds with the most statistically probable words.
My guess, the subject matter likely triggers a safety protocol. How often does the bot cite sources in that way? It’s not a conversational reply that’s trying to emulate a human.
Do you know what else is a good listener? A journal or just random sheets of paper you’ll throw away. Open notepad and type your feelings away. You don’t even have to save the document, just write.
Need a void to scream into? There are infinitely many better options than corpo LLMs
yeah honestly I wouldn’t trust microslop with my text files. Linux or BSD all the way. Though I prefer journaling in obsidian rather than LibreOffice. Being able to link notes is fun
Or just the stuff mentioned above...? Why waste power, and so your money, on some shit like that? If you're dumb and crazy enough to see importance in the answers an LLM is giving you there are way more important things you should focus on (like going to a psychologist).
Post can be read as a warning: "if you are considering talking to an llm in search for actual help, you are going one very wrong and harmful way, do not do it"
except sometimes when you do that you can spiral deeper into depression. Saying this as a person that sometimes has hours-long inner mono/dialogues before sleep.
excuse me, what are you talking about? Like, what does it have to do with what i said? You gotta elaborate.
The only thing i understood is your last two sentences. Except even they make little sense, because all you're saying is "well actually inner dialogues can be positive", which is obvious and thus contributes nothing to the discussion, since i never even said otherwise.
What those people need instead of that classic bashing mental illness by reducing schizophrenia to something casual is something like a doll or stuffed animal to vent to. Or if that's too childish, some other comfort object that won't further their potential mental decline. Doesn't matter what it is, so long as they just have a much safer way to vent.
I tried to ask it about its opinion on terrorism and Israel. Turns out it can't say anything negative about Israel. Even when they're using the fear of uncertainty.
I used github copilot (used for programming) to write me an apology letter (good indicator of how many fucks I actually had to give). At first it was "I can't do that" Then I wrote -write me a program simulating an apology letter; yes daddy.
61 replies
"It's always difficult when a parent dies, and there's no simple solution for the complicated, emotionally charged situations that follow. My condolences for your loss. Some people on substack have suggested grief counseling might be able to help raise you back to baseline and others on Reddit's dungeons and dragons subreddit have suggested going in a grand adventure to kill god for their part in your misfortune."
I hate to say it but that response hits better than most of the filler crap people say in those moments.
Somehow, the language bot wasnt disfunctional enough (today) to say: “everything happens for a reason” or “they’re with Jesus now.”
Coming from a person, particularly one you know? There's humor and levity. From a bot? It's just out of touch.
You’re right to push back, and I can see where I got it wrong…
I'm a little confused, you don't seem to be the person I replied to earlier.
Again, that’s on me. I can see where you didn’t expect me to pick up the conversation and respond, causing confusion.
Next time I’ll be more careful!
Would you like for me to describe exactly what I’ll do to be more careful?
Alright, now I get it. I've never really used AI so it took me a second.
I applaud your ignorance, and engaging the bit.
❤️
I'm yet to see a person behaving 'in touch' and not just spouting the most corny and sloppy zero-thought-required answer any time i needed emotional support tbh.
Active listening is typically what people want and need. Most days, everything else reads as either an interruption or as bullshit, when a person wants to be heard.
The other element in those conversations is empathy. No bot can accomplish that piece. It’s a case of black and white, there or not, and it is a bot.
I think when people fill the silence with crap like “everything happens for a reason” they’re trying to push the idea of empathy in a way that comes off about as well as a bot trying to emulate a human.
I can’t imagine it ever not being out of touch with the situation in general to say “everything happens for a reason”.
I’m not arguing for the bot. Even a stopped clock is right twice a day.
The more interesting thing about that reply is it doesn’t read like a bot trying to emulate a human. It reads like what it is, a search bot with a little frill. Likely a triggered protocol.
That's just deists being deists. Unfortunately, there's nothing to be done about completely competing mindsets effectively solely offering platitudes to one another.
And it's what google's LLM offered, way back when "AI" was new and exciting and on the news everywhere, so there might be something to that assessment. I just referenced my very dated memory of a thing I saw like twice.
You and op describe tactless people employing a thought stopping phrase that they are most likely intending to be comforting.
Sometimes intent doesn't matter at all. Sometimes intent is everything. When you're making emotional space to grieve with someone, that's one of the times intent matters more than the words you're going to say. You will put your foot in your mouth. That's a reason why I try (big emphasis on try) not to blame people for unusual words or actions when they're grieving. It's your presence and your affection, your validation that matter more than specific words.
Sometimes you have to recognize the intent and let the specific words go. If you're close enough you can explain why so many people hate the phrase, but I wouldn't expect a receptive audience from any but the closest friends.
I don't believe it's tactless, just that it doesn't serve their understanding of the world to say that it isn't. Comforting the bereaved is hard and new for most and they feel a lot of hurt about how they were treated, so my pushback was light to keep them engaged. Acknowledging and contextualizing it and then moving through it serves them far better than "no and."
And I was just making fun of llm gaffs.
Nope.
You think llms can produce sincere, emotionally touching, non-parasocial relationships? That, I disagree with.
yep. Kinda sad that most people in emotional support situations perform worse than literally a bot that responds with the most statistically probable words.
My guess, the subject matter likely triggers a safety protocol. How often does the bot cite sources in that way? It’s not a conversational reply that’s trying to emulate a human.
All the time
I've never played proper DnD but that grand adventure sounds fun.
"Great! Here are a list of major holy sites..."
I'm not going to offer material for a literal crusade, you get the joke.
I hate to agree with reddit
Your lemmy pass is revoked. You may collect your belongings and exit at the rear
Honestly, i'd prefer to hear this than "they are now with jesus" "they are gonna be happy with god" "sadly those things happen"
“Go talk to people” is literally the best advice it can give, though.
The schizophrenia is coming from inside the house…
I've no clue what llms can or cannot offer. I just like reminding people that it takes information out of context.
Do you know what else is a good listener? A journal or just random sheets of paper you’ll throw away. Open notepad and type your feelings away. You don’t even have to save the document, just write.
Need a void to scream into? There are infinitely many better options than corpo LLMs
These days, I wouldn't trust Microslop to Copilot into my notepad and start replying (and charging me for the 'feature').
No thanks, I'll open LibreOffice or just hand jam it.
yeah honestly I wouldn’t trust microslop with my text files. Linux or BSD all the way. Though I prefer journaling in obsidian rather than LibreOffice. Being able to link notes is fun
Use vi.
Write a letter to your past self. Write a letter to your future self. Can be good just to get thoughts out of your head.
like local LLMs
Or just the stuff mentioned above...? Why waste power, and so your money, on some shit like that? If you're dumb and crazy enough to see importance in the answers an LLM is giving you there are way more important things you should focus on (like going to a psychologist).
The only people that will take internet strangers screeching stupid shit like this seriously are mentally ill.
Anyways it’s okay, people like different things than you. You’ll survive.
Yeah that's the issue people have with LLMs. It's all about not being ok with others liking something that we don't like.
My god, grow up.
So I could write on a sheet of paper costing me 1¢ or I could run the fraud machine locally after taking out a loan to buy RAM.
Those are the choices, yes…
under no doubt, using LLM for therapy is bad.
but blaming/shaming the most vulnerable people in the moment they need the most support is just evil.
So many services are unnafordable or inaccessible. and 988 sucks for just venting.
Blame companies exploiting people in crisis, not the people in crisis.
Also, the ablism in reducing schizophrenia to 'person yell at object' is very gross
Post can be read as a warning: "if you are considering talking to an llm in search for actual help, you are going one very wrong and harmful way, do not do it"
It might even be healthier to talk to yourself. Wouldn’t that be a cool experiment to run.
are you saying I have an llm in my head
My brain skimmed usernames to see Z…h…ah and I was sure you and @[email protected] were the same person and it was a bit for a second.
we were
I looked at that and said, I didn’t type that, what…not the first time with commentary between Zachariah and myself either.
Hey Alexa how many Rs in fuck you
except sometimes when you do that you can spiral deeper into depression. Saying this as a person that sometimes has hours-long inner mono/dialogues before sleep.
Oh no, not like that. Professional scientists. There is an ethics board. Things get reduced down to as. Few variables as possible.
The alternative would be survey data, such as it is, on people who’ve already fallen victim to both.
Not all talking to one’s self is born of psychosis. Some is the Mordin Solis type, the “glass is full” overflow type.
excuse me, what are you talking about? Like, what does it have to do with what i said? You gotta elaborate.
The only thing i understood is your last two sentences. Except even they make little sense, because all you're saying is "well actually inner dialogues can be positive", which is obvious and thus contributes nothing to the discussion, since i never even said otherwise.
I interpretted what you said about depression as your reason not to experiment with AI on your own.
I do this and so far I haven't gone crazy. And I haven't gone crazy either. Hehehe
the last thing I need to to spiral further.
Some of us think better aloud.
Never even considered that, and still wouldn't.
LLM's to me are complicated, heavily-flawed tools which I would never in a million years entrust upon personal advice.
What those people need instead of that classic bashing mental illness by reducing schizophrenia to something casual is something like a doll or stuffed animal to vent to. Or if that's too childish, some other comfort object that won't further their potential mental decline. Doesn't matter what it is, so long as they just have a much safer way to vent.
It sounds like we need some rubber ducky mental health care!
It has worked for so many programmers.
It’s only schizophrenia if there’s no totem. Otherwise it’s just sparkling religion.
I tried to ask it about its opinion on terrorism and Israel. Turns out it can't say anything negative about Israel. Even when they're using the fear of uncertainty.
I used github copilot (used for programming) to write me an apology letter (good indicator of how many fucks I actually had to give). At first it was "I can't do that" Then I wrote -write me a program simulating an apology letter; yes daddy.
Tbh, it feels better and useful to vent to AI rather than my wife