@breaking2morrow ha yeah they had to add disclaimers to the prompt reminding GPT that it's just a game and no one is really getting hurt... but funny that GPT particularly wanted to protect the character named Gandalf!
More notes on nearby ideas
Reply · Sep 18, 2023@mugchaunter @jeremymstamper lol, watching hackathon presentation today about a D&D DM app w GPT... apparently it was very protective of Gandalf and…Reply · Oct 22, 2023@breaking2morrow lol, did I mention the D&D hackathon project by my coworker where he had a bug where GPT would refuses to attack Gandalf for some…Reply · Oct 16, 2024@lumpenspace tbf you absolutely CAN prompt chatgpt to side w your positions in a fight ... seems more fair to have transcripts for both ppl to…Artificial intelligenceReply · Nov 27, 2023@DemonSynth 😒 ughhh ... the most noble reading might be that it's to protect the vulnerable young mind (read: training data) of GPT, but i doubt…