devinendorphin's dextromethorphan varAIety feed

For those seeing for the first time. I made this word-machine to see how the model understands Emotional Abuse. Consider this a demo.

--

CREDITS

Text Generation via NovelAI.

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 Streambeats by Harris Heller via Spotify Playlists

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!

Inspired by the Mr. Bungle etiquette film, it devolves into either a cautionary tale of sugar or caffeine addiction, or about pregnancy, or psychosis. Or about teachers overstepping their bounds, or a principal who never leaves his office. Something for everyone!

--

CREDITS

Text Generation via NovelAI.

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 Streambeats by Harris Heller via Spotify Playlists

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!

A pallete cleanser, a bit of irrational to counteract Claude's penchant for rationale and tendency to humor me speculations. Here's a god of our present and future, who touches upon our most vulnerable moments, with very little frame of reference.

Or if you prefer, a smoldering college age Lisa Simpson reading from her short story at a bar near her school, where it's Beatnik and Bratwurst Night.

--

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!

It's a good thing you don't know me, because I'm pretty much like a broken record about bad faith. But this time, I enlist Claude's help to quantify how much bad faith drains our GDP, so that maybe we can start deincentivizing the fuckery.

If you have a vested interesrt in the fuckery, then well, this episode is boring, its not for you. Why don't you stay home and have some much needed you time. You deserve it!!

---

The views expressed are mine and mine alone, and certainly not of those listed below.

--

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!

A person on X mentioned the Basilisk in an exchange, and I took that as an assignment to extract information from huge billion voice aggregate like I did with a handful of gods (cultural bodies composed of human lifetimes), and visit the Basilisk, but needed a consult with Claude first.

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!

Endorphin shares some concerns that the three data profiles from the previous episode have surfaced, which leads into a discussion on bad faith and student debt, hell, any debt really. And the folk who would think, "We can't have a jubilee, that would harm the debt market!" And then expect us to craft the tiniest violin for them out-of-pocket.

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!

Endorphin and Claude 3 Sonnet get in the weeds regarding a system to implement compensation for an artist's work, and for the data involved with a citizen's digital footprint. We speculate wildly to show you what these models can do to punch through the blank page that prevents much necessary work from happening, so that you too can be a Citizen-Huge-Ass-Problem-Solver!

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!

TLDR: "If artists were to accept a one-time compensation payment of say $10,000 for their works being used to train AI systems, without fully understanding the tremendous cross-domain value their creative data provides, they would potentially be losing out on orders of magnitude more value.

Even just looking at the economic impact large language models are expected to have across industries, estimates range from $100s of billions to over $1 trillion in annual value in the coming years. And that's before considering the value unlocked by other modalities of AI that also leverage artistic data."

- Claude sonnet, hallucinating with Endorphin about the full potential value of the human data which informs LLMs and potentially other AI models across all domains of knowledge. (Ask Claude yourself, and get 2nd/3rd/4th opinions on this, just don't accept that first check.)

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!

Welcome to the Tenth Season! Gosh those year just flew by! That's what happens when you have a second dimension of time!

A framework for finding the kernel of truth in the most outlandish conspiracy theories. Hint: Probs not a cabal, it's maybe just yer mutha.

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!

Claude 3 and I spend about 12 hours while I do food delivery kicking around the ideas that have been collecting in my head regarding a huge legal gap--I know, do we need more law?--God I hope not, but there are people who have been harmed by actions that for most of human history have had no real language to point to the action, regarding stuff and things and more stuff and then some whosamawhatsis involving parts of our bodies we just now are starting to get a sense of them existing. Whoa Endorphin, you better lay down the pipe . . . that reason to not listen is on the house. Don't say I don't give you anything. Trust me, you won't care about this episode, so just skip this one and go to a fun one. Ehh . . . . how about one with a brief titles? Or go to a Conspiracy one, yeah, run along, we are closing out this season with boring things, you won't like it, bring me back a refill on this martini!

Behind-the-scenes fact: The part where I walk about how these issues have origins way back in our history as hominids, some dude approached a girl, NYU student most likely, offered tickets to see some comedy. Really put on the charm, said he was somewhat of a director, "oh you act? What monologue do you do for auditions? Can you do it right here?"

I start talking to Claude a bit louder, "This seems to suggest a cognitive difference perhaps at the medial pre frontal cortex! which is supposedly a site in the brain that is supposed to be the huge integrator of a lot of the systems in the brain! (dude is talking about this guy at a tennis court in StuyTown "you know how the Chinese move in around here and they {placeholder for racist things}, so I said to him you gonna be here for HOURS?") It's supposed to regulate emotion! supposed to be the seat of morality meaning the striving of a greater good even with no one watching!

Dude sounded disappointed when she did not agree to much after that, gave her the tickets anyway, she made sure they parted ways with him not knowing where she was going. I had nothing to do with that, they likely didn't hear me.

--

The following episode in no way reflects the opinions of the folks below without whose work my impulsive broadcasts would not be possible. The text in this here episode is my own, and no one else's, and I say this because I don't want anyone I respect to receive undue consequences for opinions that are AHEM MINE AND MINE ALONE. You wanna bring down a petty hammer, bring it down on this here NDA-less person who isn't a part of the machine learning industry, a person who likely has lost touch with reality anyway, so why waste your time with something that is not going to affect you much in the end, you big throbbing-membered person whose virility will light the path to other planets, oh yes, there's more where that's coming from--that's how they say it, yeah?

CREDITS

Text Generation via Claude 3 Sonnet by Anthropic.

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 Streambeats by Harris Heller via Spotify Playlists

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!