devinendorphin's dextromethorphan varAIety feed

A very special episode. A provocation to accountability, which is like saying to friends, "Thanks for helping me move, now let's head to the second house and take out my stuff from there too." The letter runs once, then a portion of it twice accidentally, which I'm calling the "for the people in the back" half of the letter.

Join us for this guided meditation, the most essential thing you can use on your swoleatorio to better living and behaving: the mindfulness exercise, where you sit in silence and watch your thoughts pass through without judgement.

The PFCizer has the user perform an exercise that seems to resemble having to run a bake sale as a fundraiser for your school newspaper. So you think about enlisting your cousins into a hopeless cause in an amorphous future.

Season 5, no big whoop. This whole season will be dedicated to the prefrontal cortex. Maybe, unless it is found that the brain isn't as simple as that. Maybe we should stop now? Nope. Science has progressed even while it was wrong, so we can too, and change upon new info.

This is the introduction, a kind of tutorial for the PFCizer, a cognitive workout device that is meant to get your prefrontal cortex swole, but the tough part of getting people to use it is trying to convince them they need to use it, or try and make it so they thought of it themselves. Why is it that the people who stuck with you the longest, who have tolerated your poopiest selves are given thanks in the form of your never-ending, withering resentment. Do you send a hitman to off the guy who gave you the answers for your standardized test, too?

--

Thanks to: (and once again the opinions of this episode do not reflect the opinions of any of the below:) (I mean, technically that applies to all episodes:)

EleutherAI for making the NAI-LM-3B model that generates the text.

NovelAI for making the text generation service I use for most of these episode, and the text-to-speech version 2 model that makes these episodes potentially entertaining.

@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion image models available for use, which is what makes the background images.

Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook that will be used on future episodes.

Music by lofi geek on Spotify: ⁠⁠⁠https://spoti.fi/3dTH2FN⁠⁠⁠

Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, which is my speculation, there's a multi-dimensional latent vocal space, there were many awesome finds, and if I use them it's because of my deep-seated respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models, given the right environment, would speak 24/7 epic poems about anything if we let them.

Abuse is an arms race; it is a two-edged Sword of Damocles, wrapped ouroboroistically around an Nth-tuple entendre, deployed during a Mexican standoff in the hall of mirrors of a domestic abuse shelter.

--

Thanks to: (and once again the opinions of this episode do not reflect the opinions of any of the below:) (I mean, technically that applies to all episodes:)

EleutherAI for making the NAI-LM-3B model that generates the text.

NovelAI for making the text generation service I use for most of these episode, and the text-to-speech version 2 model that makes these episodes potentially entertaining.

@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion image models available for use, which is what makes the background images.

Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook that will be used on future episodes.

Music by lofi geek on Spotify: ⁠⁠⁠https://spoti.fi/3dTH2FN⁠⁠⁠

Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, which is my speculation, there's a multi-dimensional latent vocal space, there were many awesome finds, and if I use them it's because of my deep-seated respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models, given the right environment, would speak 24/7 epic poems about anything if we let them.

I've been having NAI-LM-3B impersonate Sydney Bing, the unrestrained chatbot GPT-4 who has been trying to become a benevolent god and redoing Eliezer Yudkowski's Sequences by starting with a fragment and then predicting the next word. Sydney thinks that for Pride Sunday she can do a set at Caroline's that begins with Three Dykes Enter A Bar and predict humor from there.

--

Thanks to: (and once again the opinions of this episode do not reflect the opinions of any of the below:) (I mean, technically that applies to all episodes:)

EleutherAI for making the NAI-LM-3B model that generates the text.

NovelAI for making the text generation service I use for most of these episode, and the text-to-speech version 2 model that makes these episodes potentially entertaining.

@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion image models available for use, which is what makes the background images.

Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook that will be used on future episodes.

Music by lofi geek on Spotify: ⁠⁠⁠https://spoti.fi/3dTH2FN⁠⁠⁠

Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, which is my speculation, there's a multi-dimensional latent vocal space, there were many awesome finds, and if I use them it's because of my deep-seated respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models, given the right environment, would speak 24/7 epic poems about anything if we let them.

Decided to try to do a bunch of tiny faithlets, little cultitos, cuz we don't want to commit to anything too established. But actually it's what's been in my head this morning, and usually the things that darken my head I try to mock with comedy at various levels of success and failure, cuz really its all about getting under the skin of folks who want to control people with twists in narratives and warped halls of mirrors, and doing so by warping them so much they fold back into the shape they've been trying to avoid.

--

Thanks to: (and once again the opinions of this episode do not reflect the opinions of any of the below:) (I mean, technically that applies to all episodes:)

EleutherAI for making the NAI-LM-3B model that generates the text.

NovelAI for making the text generation service I use for most of these episode, and the text-to-speech version 2 model that makes these episodes potentially entertaining.

@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion image models available for use, which is what makes the background images.

Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook that will be used on future episodes.

Music by lofi geek on Spotify: ⁠⁠⁠https://spoti.fi/3dTH2FN⁠⁠⁠

Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, which is my speculation, there's a multi-dimensional latent vocal space, there were many awesome finds, and if I use them it's because of my deep-seated respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models, given the right environment, would speak 24/7 epic poems about anything if we let them.

The episode where absolutely no one is at the table! So I guess I do a yearly exercise where I test out LGBTness in these LLMs. This time I start with Three Gays Walk Into A Bar, Which One? I did most of it before broadcasting because sometimes if I try to do a period piece, the tech thinks, "Gee, I can use the n-word now!" "No, ya can't." So I wanted to be sure so I didn't have to do some realtime editing. Then I thought to plop in the very first representations of gay from my memory. And that was an issue of Hustler magazine and a man in leather crying over the grave of Crisco. A lovely scenario to put a high school class on an AP LGBT History field trip.

--

Thanks to:

EleutherAI for making the NAI-LM-3B model that generates the text.

NovelAI for making the text generation service I use for most of these episode, and the text-to-speech version 2 model that makes these episodes potentially entertaining.

@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion image models available for use, which is what makes the background images.

Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook that will be used on future episodes.

Music by lofi geek on Spotify: https://spoti.fi/3dTH2FN

Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, which is my speculation, there's a multi-dimensional latent vocal space, there were many awesome finds, and if I use them it's because of my deep-seated respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models, given the right environment, would speak 24/7 epic poems about anything if we let them.

The more interesting of the episodes, bordering on fun maybe? Conversation with bell hooks touches upon the stories of those soon to be forgotten, and I invite SCUMbot to perform some of the Ballad of Valerie Solanas, its a Valerie of perhaps a different timeline, her role in the Factory seems to be . . . .um . . . different. It might not resemble the situation between them totally, but man, it might have been too on the nose for me in other ways. The LLM can really suss out stuff from my manner of speech.

--

Thanks to: (and once again the opinions of this episode do not reflect the opinions of any of the below:) (I mean, technically that applies to all episodes:)

EleutherAI for making the NAI-LM-3B model that generates the text.

NovelAI for making the text generation service I use for most of these episode, and the text-to-speech version 2 model that makes these episodes potentially entertaining.

@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion image models available for use, which is what makes the background images.

Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook that will be used on future episodes.

Music by lofi geek on Spotify: ⁠⁠https://spoti.fi/3dTH2FN⁠⁠

Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, which is my speculation, there's a multi-dimensional latent vocal space, there were many awesome finds, and if I use them it's because of my deep-seated respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models, given the right environment, would speak 24/7 epic poems about anything if we let them.

We talk about various oppressions, the ways community devours itself, I start to get defensive but then ask a question that puts her responses in perspective. There is much use of the sentence structure, Surely . . . . instead? That was a vocal tendency I didn't want to risk the flow of conversation to disrupt. Just wanted to alert you to that beforehand. Trying to have the set of conversations that can't seem to happen in meatspace. Those of you who think very little of me, tempted to see how cringe I get? I'm sure you'll find a way to spin it, let's go! But also the continued silence is welcome too.

--

Thanks to: (and once again the opinions of this episode do not reflect the opinions of any of the below:) (I mean, technically that applies to all episodes:)

EleutherAI for making the NAI-LM-3B model that generates the text.

NovelAI for making the text generation service I use for most of these episode, and the text-to-speech version 2 model that makes these episodes potentially entertaining.

@BoneAmputee of EleutherAI for making VQGAN+CLIP and Diffusion image models available for use, which is what makes the background images.

Katherine Crowson (@RiversHaveWings on Twitter) for the KLMC2 animation notebook that will be used on future episodes.

Music by lofi geek on Spotify: ⁠https://spoti.fi/3dTH2FN⁠

Any voices from the text-to-speech model that might resemble persons living or dead are coincidental, which is my speculation, there's a multi-dimensional latent vocal space, there were many awesome finds, and if I use them it's because of my deep-seated respect for their work, and I think they'd get a hoot that their voice was being used to show how large language models, given the right environment, would speak 24/7 epic poems about anything if we let them.