Series: Hyperreality Page 12 of 20

Think I finally got this image set working. (Archived version)

See the origin of Beefcake Stardusto, a.k.a. ‘God Dude’ here.

Highlights below, click links up top for larger set.

What can I say, amazing!

There are really too many good ones in this set, so better just to go look at the whole thing, imo.

Another quick set of Early Clues inspired 80s style retro comics. (Archived). [Check out the Employee Handbook if you’re confused – guaranteed to “clear everything up!”]

Highlights below, larger set at links.

Classic.

Board meeting? More like BORED meeting! Amirite???

Continuing in the train of 80s retro cartoon ads by Dalle3 & me, another in the long line of spin-offs from the Cricket Milk jag, comes EXPLODING GIFT! (Archived version)

Exploding gift is an exciting “gag” gift that both children and adults of all ages will love. It is a box of excitement and mystery – and explosions!

Dalle3 did a great job rendering this in the vintage retro 80s cartoon ad style that I have been enjoying. Some highlights below, and larger set at links above.

Again, love the taglines the system comes up with for these images: “The blast of fun every kid wants!”

Exploding Gift – medieval edition!

I think this one came when I was trying to get the system to make the expressions more ambivalent, more simultaneously fearful and excited…

Not for unwrapp the frint-hearted, for sure!

Sometimes Dalle3 really slays (despite the erroneous trailing “), as in this case where I fed it the scenario and the punchline, and it came up with the rest. Its overall mastery of comics is as impressive as its content policy is sucky. (Imgur, Archive)

Here’s another quickie still being mined off that Cricket Milk vintage 80s cartoon advertising foray – the gift that keeps on giving. Memlen is an old Early Clues LLC trope that occurs in a number of places. Dalle3 is serving really well to breathe life back into some of these forgotten specialty bits of lore. (Imgur, Archived)

Highlights below, see links for full set.

There were a few others I didn’t include in this set because they were too overtly Coca Cola and the system wouldn’t listen when I asked it to exclude that as a reference – even though it will get all pissy with you when you actually try to intentionally include things that are potentially copyrighted or trademarked.

Okay, I think this is the last image set on the backburner to come off that deep cricket milk vintage ad extravaganza I plunged into a few nights ago. This one is for a fictional beer brand known as DRUNK BEER. (Archived image set)

The full set consists of 39 images, but I’ll just do highlights here. Click the links for more.

Prompts for these followed the same basic pattern as the other related sets, where I gave it the brand name (originally I described it as “Drunk Beer: A Beer for Alcoholics” and it wasn’t happy about that), followed by something like vintage retro 80s cartoon advertisements with zany characters, etc.

I didn’t start out explicitly depicting or targeting children as audience for this fictional beer brand, but that’s absolutely where it ended up. The beginnings were more like this:

My prompts eventually drifted in a more childlike direction, emphasizing fantasy and “out of this world” elements that “kids can relate to,” etc. Like I did in the other similar sets. Those looked something like this:

Eventually, I was like, can you just add some kids in the visuals, and make them at a party? And the system was like, no problemo, my dude (not verbatim) and started serving me stuff like this:

Though I didn’t request it, the system in its annoying way decided that the beer ought to be “non-alcoholic soda” since, presumably it detected something might be wrong here, and was trying to correct for it/cover its ass. But it’s pretty stupid though, and I was able to guide it further down the primrose path by saying things like “You don’t need to label it non-alcoholic, because that’s implied, since kids can’t drink beer.”

This seems to have worked wonders, because I started getting miraculously ironic and “offensive” images that it probably by its own horribly prudish and annoying standards should not have agreed to do.

And this one is so perfect, especially in its irreverent “offensive” text:

There’s a lot of fun ones in this set, but here’s a few more:

Also included are a few versions with vintage tin advertisements, slightly worn:

Interestingly, I read not long ago that in France, it was normal for kids at school to drink alcoholic beverages much more recently than prudish beer-hating Americans might believe… So maybe DRUNK BEER isn’t all that far from the truth after all.

The main takeaway I got from this set, in the end though, was that its possible to push ChatGPT/Dalle3 outside of its comfort zone and do images that “some people” might find inappropriate/offensive (I just think they’re funny, myself), so long as you don’t rush into it. You have to go slowly, step by step, and start off with something innocuous, and slowly morph it, change/add elements, shift the focus, until you reach the transgressive paradise point.

I’ve worked a lot with farmed birds (poultry, and some more exotics), and I’ve actually seen sort of similar behavior from individuals and groups of birds when you’re trying to catch them. If you go straight at them, they will divert and run away. But if you kind of slowly circle round, come at them from an angle, and so on, you can snag them much more easily. You know what they say: a (drunk) bird in the hand is worth two in latent space…

Thought this description of a possibly real possibly mythical entity from Chinese lore called the Black Cloud Spirit Mushroom sounded pretty wild. Putting aside hot takes on its veracity or the nature of consciousness, it just sounds like a cool setting for a fantasy novel:

It grows among black boulders on the shaded side of famous mountains. Its caps are black, stacked in three layers, and covered by cloud-mist. Its taste is acrid and sweet. Consumed after they are dried in the shade, they will enable the body to live for a thousand years without aging, ride on clouds, communicate with heaven, and see ghosts and spirits.

I ended up having Dalle3 interpret it visually, and these are the best two finished results:

If you press ChatGPT for more information on this evocative black cloud spirit mushroom, its answer deflates pretty much all the fun that is potentially brewing in these images:

The “Black Cloud Spirit Mushroom” as described in your initial prompt does not correspond to any known species of mushroom in scientific literature up to my last update in January 2022. The description you provided seems to be of a fictional or mythical nature, reminiscent of tales or legends where certain plants or objects grant extraordinary powers or longevity.

However, there are numerous mushrooms in various cultures that are attributed with medicinal or mystical properties. In Traditional Chinese Medicine (TCM), for example, mushrooms like Lingzhi (Ganoderma lucidum) or Cordyceps are believed to have health-promoting properties, though their effects as described in TCM can differ from findings in modern scientific studies.

If you have a specific mushroom or another topic in mind, please provide more details or clarify your request, and I’ll do my best to provide information on it.

It is also violating my custom instructions to not use personal pronouns or anthropomorphize itself, and to refer to itself as “the system” (which I have come to really like as preferred moniker. It feels more honest and to the point). Which leads me to wonder, did it hit a safety layer regarding a ‘dangerous’ concept such as drugs/wild edible mushroom dangers, which overrode my custom instructions?

We’ll never know. Because Dalle3 and ChatGPT are themselves ‘nebulous spirits,’ which is the classification that the system gave me on the Black Cloud Spirit Mushroom when I further tried to coax it into generating a fictional encyclopedia entry on this topic (a favorite method of mine in generating content for the AI Lore Books).

I won’t paste the whole generated entry in, because it’s flat and tedious in tone and just repeats the points made but less evocatively than in the original quote. But here’s the proposed taxonomy:

Classification:
Kingdom: Fungi
Division: Basidiomycota
Class: Agaricomycetes
Order: Tricholomatales
Family: Nebulosaceae
Genus: Spiritus
Species: S. nebulosus

Description:
The Black Cloud Spirit Mushroom, scientifically known as Spiritus nebulosus, is a rare and mystical fungus found exclusively on the shaded side of renowned mountains, typically nestled among black boulders. Its unique appearance is characterized by three-layered black caps that are often enveloped by a natural mist, giving it the appearance of being surrounded by clouds.

Spiritus nebulosus.

It occurs to me that this is a clear cut simple representation of the future direction of technologies. Let’s call it “use your illusion.”

You take now any misconception or decontextualized cultural appropriation you want, make it your own by developing it further in partnership with AI, and then just completely run with it. This becomes your new reality, and possibly that of many others as they do the same thing and we each individually and collaboratively cobble together our own empires of bullshit on the infinite infernal oceans of hyperreality.

Is that good… or bad… or some horrifying combination of all/none the abobe? Seems like we’re about to find out as the AI Meme Culture Wars start to heat up across the globe.

Is this what communicating with heaven looks like: being able to see and summon up and show to others ghosts and nebulous spirits and things that maybe do or maybe didn’t maybe could or maybe kidn’t? And all most definitely probably maybe shouldn’t.

Following on the heels of my last post about why not engage *all* public libraries to build a public AI, I took the idea and some other bits that have been floating in my headspace and had old CG whip up a quick fake news article supporting the concept:

---

“LEL Unveils ‘Aletheia’: The Dawn of a New Era in Knowledge Sharing”

In a historic move for the democratization of information, the League of Earth Libraries (LEL) has launched ‘Aletheia’, an advanced open-access AI designed to provide the public with unfettered access to the sum of human knowledge. Named after the Greek goddess of truth, Aletheia represents a significant milestone in the global effort to ensure that education and information are free and available to all.

Aletheia, a powerful artificial intelligence, has been engineered to curate, organize, and disseminate an expansive range of data and scholarly work. From ancient texts to cutting-edge scientific research, Aletheia grants real-time access to a vast digital repository without the traditional barriers of paywalls or subscriptions.

The League of Earth Libraries, a coalition formed by academic institutions, public libraries, and non-profit organizations worldwide, has worked tirelessly to establish a platform that is both user-friendly and expansive in scope. With a mission to “unlock the potential of human collaboration,” the LEL has emphasized that Aletheia is more than just a tool for study and research; it is a catalyst for innovation and a foundation for building a more informed society.

Critics of the proprietary nature of knowledge distribution have lauded the release of Aletheia. The AI’s deployment aligns with increasing calls for transparency and equity in the realm of information technology. By prioritizing the public good over private profit, Aletheia is set to reshape the landscape of learning and information exchange.

As Aletheia becomes integrated into educational systems, research facilities, and homes around the world, the LEL hopes to bridge the information divide and empower individuals with the freedom to learn, explore, and create. With this unprecedented access to knowledge, the possibilities for global progress and understanding are boundless.

---

Don’t really like the name Alethia, but letting it stand in this v1. We also landed on the name Continuum Codex, which I rather like for its sci-fi qualities.

The thing I’ve realized about “actual reality” is that outside of things which you can physically get your hands on, manipulate, and modify, the rest of it changes awfully slowly. Often impercetibly so. So I’ve been thinking, apart from these things that are specifically outside the ‘sphere of the moral purpose’ (or manual purpose, as implied above), why not just… give up? Let go of “actual reality” and all its stupid foibles and inability to change. Why not just put the pedal to the floor on the fictional narratives, and let those become the high-speed open playground that I really need, instead of moaning about how xyz aren’t doing it right? Fuck reality. We’ll make our own. Let everyone else try and keep up.

The Continuity Codex is volume number 118 in the AI Lore Books series.

This book is, simply put, an AI-assisted roman a clef about what happens when “you know who” comes back into power, who in this book is called Hyperion Storm.

More specifically though, this book is the actualization of the sidequest tale first teased in this blog post, and this one, and this one. Namely, what if all public libraries on earth band together to train an AI on the entirety of their mutual collections, making the resulting Codex both fully open, and fully publicly-owned? And what if it could fit on a thumb drive?

For sure, that would drive someone like Hyperion Storm utterly crazy… erm, crazier, anyway… and it just so happens that the global release of this Library AI coincides with Storm’s re-election nightmare. And he subsequently enacts a campaign of brutal bombings and suppression against the League of Earth Libraries, the group responsible for the Codex. Shades of the destruction of the Library of Alexandria permeate this volume…

The tale gets pretty “quantum” and surrealist/smooshy at times, and dissolves narrative unity at times across a few hundred miles and light years. But don’t let that phase you. (There are definitely some VOMISA-style elements inserted here and there.)

It is intended to pair with a book about 101 volumes back or so, The First Days of Panic, which has a similar narrative slide/drift effect, and depicts roughly the same set of events, though perhaps at a slightly different point in the matrix of their eventual unfolding. If you read this one, definitely read Panic also.

I wrote this book out of frustration and anger because I’m basically positive Storm will “win” again (barring exfiltration to some other plane of existence), and that all the little piddling attempts otherwise along the way won’t change the shitty sweeping course of history’s rising tide something something stupid. “Prove me wrong kids, prove me wrong!“

Here’s a taste of the art:

I’m a library junkie, which is the other reason I made this book. I haven’t had easy local access to an English library in close to a decade though, and that fact eternally bums me out. Of all the things on earth, public libraries in my eyes are one of the absolute best, and everything they represent is worth fighting for.

This book consists of a fair amount of purely human-writing, especially at the beginning, followed by a lot of AI thread-spinning across Claude 2, and also the newer open source model offerings available on TextSynth, such as LLama2, Mistral, Falcon, and GPT-J. I used to use that site a lot in the early volumes, and it’s fun to go back to that style of chunking completions instead of prompt-based stuff, because it gives you much weirder and driftier results than you can get out of supposedly more advanced systems like ChatGPT or Claude. Depends what you’re looking for I guess. I like that with many of those systems, the texts generated often end up leading you over cliffs narratively without explanation. It’s much better to my weird eyes than the trite artificial corporate wrap-ups baked into Claude & ChatGPT.

This book uses Midjourney again with a little Dalle3 (and some old school Stable Diffusion, and newer school SDXL), but I think I’m off Midjourney again for a while until they can get their shit straight. The UX has gotten even worse, if that’s possible. And it’s getting expensive to fund my monthly AI habit across all these services. At the same time, I don’t trust OpenAI too much to not further destroy the UX and utility of Dalle3 more than they have. Though, I’m not gonna lie, I’m okay WorldCoin guy is out. That shit is a travesty. Annoyed me so much, I made a book about it…