Post Av241bhUMEdREIRRuy by nazokiyoubinbou@urusai.social
 (DIR) More posts by nazokiyoubinbou@urusai.social
 (DIR) Post #Av241aa0WZ3ZknvxcO by sb@metroholografix.ca
       2025-06-11T16:54:49Z
       
       0 likes, 0 repeats
       
       My partner has two (!!) best friends who frequently use #chatGPT. Not only do they use it, they use it as an oracle, a diviner, a settler of arguments between friends, a therapist for their children, a spiritual adviser. Scary, scary stuff. They both feel personally attacked when she shares facts about the ownership, ecosystem, environmental and mental impacts, etc surrounding #llms. Please, if you have any resources that might be helpful to them, or to my partner in attempting to #help them, please share.
       
 (DIR) Post #Av241bhUMEdREIRRuy by nazokiyoubinbou@urusai.social
       2025-06-11T18:55:46Z
       
       0 likes, 1 repeats
       
       @sb I did see one thing someone said that might help hammer into people how bad LLMs actually are at information.  Get them to ask it about stuff they actually know a lot about.  I can definitely confirm that it doesn't take long for the LLM to start getting very wrong and then doubling down on the wrongness as you delve more into it.  (The key is they can't just ask one question and then walk away.)Might be worth a try if you can get them to do it properly.