OpenAI spends about $700,000 a day, just to keep ChatGPT going. The cost does not include other AI products like GPT-4 and DALL-E2. Right now, it is pulling through only because of Microsoft's $10 billion funding
I actually started my journey into Lewd AI stuff with NovelAI. I stopped using it after awhile because I like chatbot rp specifically, not just something that will finish a story for me. Using Silly Tavern to try and emulate the NovelAI models into acting like chat bots just shows how not good they are at that.
Same, at least for anything but the tiny ones that will fit in my limited vram. Hoping a gpu that’s actually good for LLM will come out in the next few years that’s not 15k and made for servers.
I’ve heard it’s not as good, but I think NovelAI works pretty well, and explicitly allows that.
I actually started my journey into Lewd AI stuff with NovelAI. I stopped using it after awhile because I like chatbot rp specifically, not just something that will finish a story for me. Using Silly Tavern to try and emulate the NovelAI models into acting like chat bots just shows how not good they are at that.
Makes sense. In that case I guess your next best option is probably to buy or rent hardware to run the local models that are suitable for chat rp.
I have definitely been considering it. My current hardware gives me about an 80 second delay when I run an llm locally.
Same, at least for anything but the tiny ones that will fit in my limited vram. Hoping a gpu that’s actually good for LLM will come out in the next few years that’s not 15k and made for servers.