kaveenk

Hey all! I've posted previously here about my project GPTDiscord, a full-fledged OpenAI integration for Discord. Recently, we've made some updates that enable internet-connected chatting! During a conversation, the bot will be able to perform mathematical operations with wolfram, search google and get web-content, and the bot is now even able to directly browse and crawl links that you give it, to help answer your questions!

In addition to this, there are already a ton of other features built in, such as document (infinite length) understanding and Q&A, permanent memory chatting, GPT-4 support, image generation, and etc.

Check it out at https://github.com/Kav-K/GPTDiscord

Thank you! You’re correct that ChatGPT is different from davinci. They’re both built on GPT3.5, but ChatGPT is trained on more conversational influences. However, davinci is really strong and will behave mostly like ChatGPT with the right conversation prompt :)

Exactly! This is also what I’ll be implementing within the next few days. Embedding conversation history and dynamically doing semantic search for relevant conversation snippets and building a prompt based on that to send to GPT3

Hey! This is an option in the settings. You can type !gp to view settings and !gs to change them. For your case, !gs summarize_conversations True !gs summarize_threshold (token amount to summarize at)

Thank you for bringing this up because this is something I need to change in the README. You’re correct in presuming that it’s just because it’s simpler to tell users to use a password, but this is not the right way to approach this.

Thank you! I agree re: token usage, it will get expensive with extensive use. The image prompt optimizer pre-text itself is also roughly 3500+ tokens, and if you use best_of=2 for example, it’ll basically be like 20 cents per image optimization request. However, I love the optimizer and I think it works amazing, for DALLE, MJ, and SD!

None that I’ve made, but I think there are several slack integrations out there, I see a lot on the OpenAI community server :)

Yup! Behind the scenes I send AIOHTTP requests to the various OpenAI endpoints.

Thank you!

So I actually have it set to use both the text-davinci-003 and text-curie-001 models. If you’re on low usage mode, that’s when the curie model will be used :)

I've build a GPT3 and DALL-E2 integration for Discord. With it, you can ask GPT3 questions, have full conversations with GPT3 (just like ChatGPT), generate, vary, and redo images, and even optimize text prompts for image generation!

The conversation functionality also has medium term memory, and you can chat with it infinitely without having to worry about context/token limits. Moreover, in the next few days, I will be implementing long term and permanent memory for the bot using embeddings!