Before I even began this whole thing on Twitch, I spend Thanksgiving weekend 2020 on AI Dungeon trying to see how elaborate we can make a college full of jesters, and a setting all about teaching what makes comedy funny. After like 13 episodes in one weekend, including one where Scott Thompson guest starred and did a one person play about a small mining town on strike in Canada--very pointedly no gays to be found--I knew something was up, if it seemed the system was able to match the multi-layers of comedy-anti-comedy I was laying on their weird-ass casserole.
That was how davinci-3 did it, now trying LLAMA 3.1 403B, but this episode we are doing a simple scenario.
--
CREDITS
Text Generation via LLAMA 3.1 403B Base model.
Voices from the TTS v2 of the NovelAI text generation service, a product from a company now called Anlatan.
@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion (AND MANY MORE) models available for use via BATbot.ai, which is what makes some background images.
Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook.
David Marx (@DigThatData on Twitter) for souping above KLMC2 notebook to make use of initial images, which takes KLMC2 abilities to next level. Multi-prompt scheduling still baffles me, but I'll tackle them someday.
Music By Harris Heller via Streambeats Twitch-Friendly Lo-Fi Playlists on Spotify
Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, and if I use them it's because of my respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models would speak 24/7 epic poems about anything if we let them. In the meantime, because I use them, I will not be monetizing any of this. This is evidence of capabilities for the public good. And my resume/CV!
