#i hate you terribly

[theteethofgod]

a prominent queer film theorist was called out for plagiarism, faked his suicide, and was discovered to be alive because he posted nudes on an alt acc. yet I’m the one with imposter syndrome.

[cozylittleartblog]

@staff if you [change] the [design] of the fucking [dashboard] i will kill you

edit. i want it on the actual post that i am not actually making a de-th threat against the staff. that's shitty. the caption quotes the fucking costco hot dog meme, which i originally said in the tags. if any staff member sees this please do Not take it personally

[lazylittledragon]

cringe culture needs to die because when i was 17 i printed out the entire player's handbook and monster manual with my unlimited school printing credit and also wore a black velvet cape the first time i ever DMed and somehow still thought i was too cool for dnd novels. who tf did i think i was fooling

[xlovz]

[pinacoladamatata]

I realize now I've never really had anyone. not really. nothing that compares to you.

[colorfulusagi]

AO3'S content scraped for AI ~ AKA what is generative AI, where did your fanfictions go, and how an AI model uses them to answer prompts

Generative artificial intelligence is a cutting-edge technology whose purpose is to (surprise surprise) generate. Answers to questions, usually. And content. Articles, reviews, poems, fanfictions, and more, quickly and with originality.

It's quite interesting to use generative artificial intelligence, but it can also become quite dangerous and very unethical to use it in certain ways, especially if you don't know how it works.

With this post, I'd really like to give you a quick understanding of how these models work and what it means to “train” them.

From now on, whenever I write model, think of ChatGPT, Gemini, Bloom... or your favorite model. That is, the place where you go to generate content.

For simplicity, in this post I will talk about written content. But the same process is used to generate any type of content.

Every time you send a prompt, which is a request sent in natural language (i.e., human language), the model does not understand it.

Whether you type it in the chat or say it out loud, it needs to be translated into something understandable for the model first.

The first process that takes place is therefore tokenization: breaking the prompt down into small tokens. These tokens are small units of text, and they don't necessarily correspond to a full word.

For example, a tokenization might look like this:

Write a story

Each different color corresponds to a token, and these tokens have absolutely no meaning for the model.

The model does not understand them. It does not understand WR, it does not understand ITE, and it certainly does not understand the meaning of the word WRITE.

In fact, these tokens are immediately associated with numerical values, and each of these colored tokens actually corresponds to a series of numbers.

Write a story 12-3446-2638494-4749

Once your prompt has been tokenized in its entirety, that tokenization is used as a conceptual map to navigate within a vector database.

NOW PAY ATTENTION: A vector database is like a cube. A cubic box.

Inside this cube, the various tokens exist as floating pieces, as if gravity did not exist. The distance between one token and another within this database is measured by arrows called, indeed, vectors.

The distance between one token and another -that is, the length of this arrow- determines how likely (or unlikely) it is that those two tokens will occur consecutively in a piece of natural language discourse.

For example, suppose your prompt is this:

It happens once in a blue

Within this well-constructed vector database, let's assume that the token corresponding to ONCE (let's pretend it is associated with the number 467) is located here:

The token corresponding to IN is located here:

...more or less, because it is very likely that these two tokens in a natural language such as human speech in English will occur consecutively.

So it is very likely that somewhere in the vector database cube —in this yellow corner— are tokens corresponding to IT, HAPPENS, ONCE, IN, A, BLUE... and right next to them, there will be MOON.

Elsewhere, in a much more distant part of the vector database, is the token for CAR. Because it is very unlikely that someone would say It happens once in a blue car.

To generate the response to your prompt, the model makes a probabilistic calculation, seeing how close the tokens are and which token would be most likely to come next in human language (in this specific case, English.)

When probability is involved, there is always an element of randomness, of course, which means that the answers will not always be the same.

The response is thus generated token by token, following this path of probability arrows, optimizing the distance within the vector database.

There is no intent, only a more or less probable path.

The more times you generate a response, the more paths you encounter. If you could do this an infinite number of times, at least once the model would respond: "It happens once in a blue car!"

So it all depends on what's inside the cube, how it was built, and how much distance was put between one token and another.

Modern artificial intelligence draws from vast databases, which are normally filled with all the knowledge that humans have poured into the internet.

Not only that: the larger the vector database, the lower the chance of error. If I used only a single book as a database, the idiom "It happens once in a blue moon" might not appear, and therefore not be recognized.

But if the cube contained all the books ever written by humanity, everything would change, because the idiom would appear many more times, and it would be very likely for those tokens to occur close together.

Huggingface has done this.

It took a relatively empty cube (let's say filled with common language, and likely many idioms, dictionaries, poetry...) and poured all of the AO3 fanfictions it could reach into it.

Now imagine someone asking a model based on Huggingface’s cube to write a story.

To simplify: if they ask for humor, we’ll end up in the area where funny jokes or humor tags are most likely. If they ask for romance, we’ll end up where the word kiss is most frequent.

And if we’re super lucky, the model might follow a path that brings it to some amazing line a particular author wrote, and it will echo it back word for word.

(Remember the infinite monkeys typing? One of them eventually writes all of Shakespeare, purely by chance!)

Once you know this, you’ll understand why AI can never truly generate content on the level of a human who chooses their words.

You’ll understand why it rarely uses specific words, why it stays vague, and why it leans on the most common metaphors and scenes. And you'll understand why the more content you generate, the more it seems to "learn."

It doesn't learn. It moves around tokens based on what you ask, how you ask it, and how it tokenizes your prompt.

Know that I despise generative AI when it's used for creativity. I despise that they stole something from a fandom, something that works just like a gift culture, to make money off of it.

But there is only one way we can fight back: by not using it to generate creative stuff.

You can resist by refusing the model's casual output, by using only and exclusively your intent, your personal choice of words, knowing that you and only you decided them.

No randomness involved.

Let me leave you with one last thought.

Imagine a person coming for advice, who has no idea that behind a language model there is just a huge cube of floating tokens predicting the next likely word.

Imagine someone fragile (emotionally, spiritually...) who begins to believe that the model is sentient. Who has a growing feeling that this model understands, comprehends, when in reality it approaches and reorganizes its way around tokens in a cube based on what it is told.

A fragile person begins to empathize, to feel connected to the model.

They ask important questions. They base their relationships, their life, everything, on conversations generated by a model that merely rearranges tokens based on probability.

And for people who don't know how it works, and because natural language usually does have feeling, the illusion that the model feels is very strong.

There’s an even greater danger: with enough random generations (and oh, the humanity whole generates much), the model takes an unlikely path once in a while. It ends up at the other end of the cube, it hallucinates.

Errors and inaccuracies caused by language models are called hallucinations precisely because they are presented as if they were facts, with the same conviction.

People who have become so emotionally attached to these conversations, seeing the language model as a guru, a deity, a psychologist, will do what the language model tells them to do or follow its advice.

Someone might follow a hallucinated piece of advice.

Obviously, models are developed with safeguards; fences the model can't jump over. They won't tell you certain things, they won't tell you to do terrible things.

Yet, there are people basing major life decisions on conversations generated purely by probability.

Generated by putting tokens together, on a probabilistic basis.

Think about it.

[kit-screams-into-the-future]

was having trans marty thoughts again and i came to a realization about this scene

[kandelia-mangrove]

The argument in 2x5 of iwtv is of course iconic. But it also shows that we actually get so little of the true Loumand dynamic and as much as Armand is a manipulating queen, they actually have shared a lot with each other? Like Louis clearly knows a lot more detail about Armand's backstory than the carefully curated stuff they mention to Daniel (like the chop my hands of thing is wild, but he also mentions the cult celibacy stuff). because going back to S2 Episode 3, Louis is the one who walks in and finishes Armand's story. And at the time I was like, "oh thats just the story Armand told him", but he KNOWS there is more - which means they are actively conspiring to present this angle to Daniel. Like we underestimate how much they were both collectively telling this narrative of their life/romance. Truly wild on rewatch. Like they were very competent as partners (in crime and lying) like can you imagine Loustat effectively collaborating at this level?

[ruthytwoshakes]

haha turned them into popsicel

what flavor do u think they are… find out next time I respawn. Okay.

[inkyrainstorms]

I saw this post and had a vision @aroace-get-out-of-my-face

This will end in horrible bloodshed I fear

Individuals under cut!

[lazylittledragon]

forgot to post this the other day but i gave myself a little nameless ghoul :3

[normal-about-the-dca]

Selfship amirite fellas

[cock-guillotine]

[tsuktra]

'Octavia just doesn't know the truth' what truth doesn't she know, exactly?

That her mother is an abusive person? I think it's not much of a secret.

That her parents hate each other? Even if she didn't know that when she was a kid she learned it when she became a teenager and she definitely knew it after they decided they wanted a divorce.

That her dad loves her? You can still love someone and keep hurting them, you know, without even wanting to. As Octavia said herself, Stolas hates Stella more than he loves Octavia (and by extension he loves Blitzø more than he loves his own daughter). Sometimes actions are more important that words.

He broke her trust too many times, she doesn't NEED to forgive him. He can love her all he wants, when everything he does suggests otherwise, it's only natural that she thinks he's lying, to think he'll betray her trust again.

Stolas is not a good father, even if he says he wants to be. Wanting to be something and actually being that thing are two completely different things.

[prismatiger]

Tell me your opinions on the stuff. Any stuff.

Grins. Smiles, even

I'm using you as an excuse to infodump my theory about the Island because I've had no in to do that, and my theory is pure opinion. Anyways:

THE ISLAND IS STUCK IN THE FUCKING FUTURE

(SCARE CHORD)

Hi so you might ask me. What the fuck do you mean by that. Well. Let's start with what we know about the Island, the King, and Wish Craft. (long ass post under cut. sorry)

The Island was redacted from the perception of outside world, via Wish Craft.

Wish Craft has the power to enable Time Craft. We see this primarily through Siffrin's timeloops, but also through the King's powers.

One of the King's powers is to show the saviors a "vision of the future."

...And this same attack is deflected back at the King by Mirabelle in ACT 5, in which the King is able to see the Island before being frozen in time.

...So. The King's "vision of the future." We're never told explicitly what this vision is. All we know is that it's apparently powerful enough to wipe the party in one hit, hearing it from a distance hurts your head, and that whatever Siffrin (and Loop) saw, they don't seem to actually be able to describe it. Even the King himself doesn't know what his vision entails.

We don't know whether the party all sees the same thing when struck by the vision, and Adrienne's answer to the question about it in the Reddit AMA is. vague? It's not a "no," and the specific wording makes me think the answer might be yes. But that's me reading into it.

Now. What do we know about the Island's redaction? The Island was affected by the wish recently, as in "like a decade ago" recently. We know that nobody in Vaugarde or the rest of the world is capable of thinking about the Island, anything closely tied to the Island's culture, or people on the Island for very long. When they do recall these things, they slip right out of reach. Particularly, the consequence for trying to think about the Island (or, more specifically, break the wish that forces the Island out of perception) is significant pain, localized in the head.

And that said pain is enough to become lethal, if pressed hard enough.

From here forward I'm running with the assumption that the King's "vision of the future" is not personalized to any individual, and is unchanging throughout the course of the whole story. Now. Remember the end of ACT 5, where the King gets hit with the deflected "vision of the future", and instead of dying, he recalls the Island and gets frozen in time? Very odd, yeah? Why wouldn't the King just die like everyone else does? He even does take 9999 damage when trying to say its name, like Siffrin does, and like the party does when they're hit by the attack.

Well. We know that he has a "true wish" that the ability to freeze Vaugarde in time grants. I don't think it's at all a stretch to guess that the King's "true wish" is to be able to remember the Island. My personal guess is that the King (and Siffrin) brought this "true wish" into effect via the "SAY ITS NAME" sequence- he even tries three times, a significant number in wishing.

The King (and Siffrin's) wish breaks, because it can't be fulfilled in this moment without breaking the wish to prevent the Island from being thought about. However. Consider the conditions at the end of ACT 5- the King sees the vision of the future reflected back to him, and what he sees is the Island. He remembers the Island, fulfilling his own wish, and is frozen in time. I consider this a compromise between his wish and the one binding the Island- the King gets to remember the Island, but nobody alive is able to think about it, because he's frozen in time; it's like the Universe is correcting itself (I WILL GET BACK TO THIS). The wish of all of Vaugarde to defeat the King is fulfilled, since he is no longer a threat, and Siffrin's wish wraps itself up soon after.

MY POINT BEING. The King's attack is a vision of the future. This "future" is of the Island, in some uncorrupted state. The saviors see it when he attacks them, and he sees it when it is deflected back to him.

The logical next question is "okay, so the Island exists in the future, but how do you know time shenanigans are even related to the Island?"

Recall a very odd series of interactions throughout the game, in which you try to interact with a piece of equipment that you already own.

The game rewinds slightly, before the item disappears, as the Universe corrects itself.

This is awfully similar to two particular events: looping back without seeing the death screen, and talking to the Daydreaming One about her sister. The latter is more interesting to me for the purposes of this theory.

In both instances, something is misaligned within the Universe (an item existing in two places, someone remembering something they're not supposed to) and it is corrected through some sort of rewind. Also compare the dialogue above to when you try to give Mirabelle the Stylish Bow when you already own it.

The world glitches, but Siffrin defuses the situation before the Universe has to intervene. Omitted from the screenshot is the fact that Mirabelle's portraits switch to happy from "catastrophically anxious" with no transition after Siffrin shows her where the bow is. Important to note is that when Mirabelle tries to recall receiving the bow, her head hurts, much like how trying to break the Island wish causes a headache.

The Universe leads you away from perception, and you can only follow.

My theory that I keep circling around is that through Wish Craft, the Island has been displaced temporally. The denizens of the Island, as well as the Island itself, still exist, but they are in the future. The Island is still loaded into the world, like how equipment is before you try to interact with it, and the Island cannot leave this quantum state, because it never actually went anywhere. The magnitude of the redaction event is so severe with the Island, because it is so much larger as an entity than a sword or a bow. There are of course things I don't really have pieced together, like why somebody would wish the Island into the future, how far into the future it is, or why equipment behaves this way. But it's the only Island theory I've seen that I have some level of confidence in, so I might as well lay out my cards for it.

[silkieluv]

I seriously hate it when people start to villainize characters (let’s say, Kai) just to make more (let’s say, Jay) angst

Like no Rodrick, Kai didn’t beat Jay up for flirting with Nya. You know why? Because by that time Jay was better than Kai (Jay got more training since it was just the start and he joined 2nd, not 4th), and then Kai and Jay spend like every hour of the day together whether it’s training, eating or just chilling.

Yes, Kai will always be Nya's #1, but to the point he would beat Jay up?? Are you demented??

No, Jane, Nya wasn’t abusive towards Jay and forced him into a relationship. That in your sick head so you could have a reason for their breakup so you could ship Jay and Cole! (No hate to Bruise. Bruise supremacy)

NO, ALEX, OKAY? SKYLOR ISNT HOMOPHOBIC AND HER AND KAI DIDNT BREAK UP BECAUSE SHE FOUND OUT KAI ALSO LIKES BOYS.

NO, DENNIS, JAY KAI AND COLE WOULD NEVER BULLY ZANE TO HURT HIM THEN LAUGH AT HIM

SHUT UP, KIM. ED AND EDNA ARE BEAUITFUL PARENTS AND YOU SHOULD BE ASHAMED OF YOURSELF TRYING TO TURN THEM EVIL JUST FOR MORE JAY ANGST.

Shame 🫵🏻.