It gives r/TwoSentenceHorrorStories kinda vibes, I like it.
post
I hate how people are using AI for this now. It used to be that you had to actually do that work of kidnapping your therapist and getting them to roleplay as your ex-wife at knifepoint.
Uploading entire conversation to a data harvesting company is so wise...
Oh my God it's a joke.
Me in the 90s: computers are great, I wish everyone has access to them and know how tonise them.
Me today: hey genie, can I get one more wish to undo my last one.
Neither is posting the same meme that's been posted 5 times in a week, but here we are.
¯\_(ツ)_/¯ if the conversation was via a centralised platform like WhatsApp, SMS, or Discord, they probably already have it...
I've seen many screenshots of people genuinely doing that. It's disturbing.
"My BF and I are arguing, help me win."
"Tell me how to respond to this girl I'm interested in."
This is still better than what my mother would do when her and her second husband were arguing. She would call me or one of my brothers to try to get us to take her side in the argument.
We would have much preferred her taking it to AI than trying to drag us into her disagreements.
"Tell me how to respond to this girl I'm interested in."
This might actually be a win. It will accidentally train them to be polite and courteous.
It depends on the usage.
If the user is reading the response and then thinking about it and adapting the suggestions to their own style of speaking, maybe they'll learn something over time. Studies have shown that students who use LLMs as a tutor do improve their understanding of material.
If they're copy/pasting like most of the world, then there's evidence to suggest that they'll actually get worse at the thing they're offloading to the LLM.
I'm guessing this is mainly a joke, but it's really not training them to do anything except offload the "labor" of relationships to a third party and that women are video games where selecting the right dialogue options will get you your preferred ending.
Who doesn't like a happy ending?
Not everyone has good role models. Would you rather people get advice from Andrew Tates or Grok? Because you might want it to be someone other than those options but right now that is what a lot of people are picking between and we have no way of reaching and stopping them from picking poorly. As awful as Grok is, it manages to be somehow better than the other options.
You know tgere's people that upload complete, unedited customer data with a prompt "write personalized emails for these customer to sell x"
To be fair, those conversations probably happened on the data harvesting company's platform.
ChatGPT, where is the best place to buy methamphetamines and a hooker locally?
god i miss craigslist.
Isn't Craigslist still alive but barely kicking? Some wild and very overt things in the missed connections section.
Edit
Still alive and horny

If you need an AI to tell you that you don't deserve either. You got to get in your car and drive around it two o'clock in the morning or it isn't worth it.
Lmao, try accounting software with automatic uploads for the entire staff.
Supposed to do accountants math, cause that's something computers struggled with so much before /s
That's the thing you decided was the topic of importance to comment about...? Trying to not read too much into this but I can't deny that your comment is very weird, there's so much worse going on.
How is this slop? I don't think this is an AI generated image.
The image quite literally contains a text chat with a bot. It could not more clearly be ai generated. It's "haha look at my chatgpt log its so funny" content.
When even the LLM knows you're fucked up.

all 44 comments