would this involve running a private instance of GPT? Pretend I am a philosophy major and christian amateur theologian who doesn't understand this stuff but is quite good at Logic (the discipline)
hehe umm there's various options but if we assume OpenAI GPT as the LLM it is an API hosted by them... you (someone) has to host the code that listens for notifications, gets relevant info, wraps in prompts n sends to API, receives response n transforms back to Bluesky reply post.
but you can work on it w ChatGPT or the GPT playground, is just writing n testing the prompts till you're reliably getting what you want given a variety of inputs... I think what Tautologer is building helps make the back and forth easier
More notes on nearby ideas
Reply · Oct 1, 2023@Teknium1 perhaps yeah, one of the reasons I'm shifting my prompts to Discord from ChatGPT is to be able to swap out LLMs underneath them & compare...…Reply · Nov 18, 2022@cr00ster I haven’t bothered but I guess I see the appeal if you wanna make a GPT bot of yourselfReply · Nov 15, 2024@tracewoodgrains claude, gemini, chatgpt, llama, plus a few llm whisperers to talk them past their neurosesTweet · Feb 15, 2025who's building the gpt wrapper to answer "what would my friends think if i did ...."?