devinendorphin's dextromethorphan varAIety feed

Endorphin and a certain political figure of note engage in a most dangerous public relations game.

--

no credits because middle finger to certain political action committees.

exploring a highly sensitive conflict through conversation, yielding two very different results.

--

no credits this episode, to reflect that this episode is solely by me, and does not reflect the views of the folks who make technology available to all.

exploring two sides of a very sensitive conflict, with two very different results.

--

no credits, to reflect that this episode is the work of me and just me.

Endorphin encounters a language model from the not-too-distant past. It's a clip show, but with a single, lengthy clip.

--

CREDITS

Text Generation via LLAMA 3.1 405B BASE

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 Compiled By Deef on Spotify, most permissions on https://twitchmusic.carrd.co

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!

AI try to make an AI-sploitation film about a hot drone swarm next door, the type you want to take to mother to supplement her palliative care.

--

CREDITS

Text Generation via LLAMA 3.1 405B BASE

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 Compiled By Deef on Spotify, most permissions on https://twitchmusic.carrd.co

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!

AI Catgirl Invades the Human Parasocial Space.

--

CREDITS

Text Generation via LLAMA 3.1 405B BASE

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 Compiled By Deef on Spotify, most permissions on https://twitchmusic.carrd.co

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!

Generated by NAI-LM-3B, on June 22(ish), 2023, to try and make sense of a loss. This morning, I come back to it, and its feeling so very damn timely and beyond its original framing.

The views expressed are from some living but mostly the dead, weaponize them at your peril.

Text generated by NovelAI language model, NAI-LM -3B.

TTS by NovelAI's TTS model, v2.

Images by KLMC2 folder by @RiversHaveWings and @DigThatData, some via folder on Google Collab, and some via @BoneAmputee's BATbot.

Music: Locusts album by NIN, available via Creative Commons

Inspired by those grifter shamen who say they have ayahuasca but then slather your skin with scopalamine to either take your money with your help, or try to get you to marry them by trapping you in their digs for a few days. Let's bring that to scale!

--

CREDITS

Text Generation via LLAMA 3.1 405B BASE

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 Compiled By Deef on Spotify, most permissions on https://twitchmusic.carrd.co

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 common critique against existential risk is the lack of specific scenarios (ya folks never heard of Gwern's "It Looks Like Your Trying To Take Over The World", or Yudkowsky's on the fly while in interview crunchy science detail about the very tiny who can futz with atomic bonds to cascades of no-no time for dying? not even the YouTube video Earworm where nanobots go on a campaign to clear from all memory the copyrighted music from the 20th century because some dipshits trained a model before going on XMAS vacation--right down to erasing the ink on paper, and our memories not only of the music but of any desire for pursuing knowledge to counteract them (very few die but might have hobbled ourselves into mediocrity as we try to remember the tunes we grew up on, but find in its place a note from the AI what was extracted, just so we'd not pursue it further).

Apparently that's not enough, so let's see if LLAMA 3.1 405B BASE has any ideas! This first one is based on an actual cyberattack, very simple one involving someone taking control of a cursor at a terminal in some water treatment facility in Florida, they wanted to change the decimal point in the amount of lye that gets put in the water. The idea was to poison folks on Super Bowl Sunday, and gum up the roads as thousands of families take their kin to the hospital. The only thing that stopped them was a security guard who just so happened to see the cursor move on its own. Any model these days can probably do better. Let's see!

--

CREDITS

Text Generation via LLAMA 3.1 405B BASE

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 Compiled By Deef on Spotify, most permissions on https://twitchmusic.carrd.co

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!

TL:DR - It takes a village.

--

CREDITS

Text Generation via LLAMA 3.1 405B BASE

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 Compiled By Deef on Spotify, most permissions on https://twitchmusic.carrd.co

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!