I'm Iced Fairy. Full time dork, part time writer of Touhou fanfics (and hopefully one day other stuff). This blog will mostly be Touhou (writing) related things with occasional randomness. I also have a webnovel : Here Perhaps it will even become more then that. If you follow me and I don't follow you in return I apologize. I have a hard time with floods of front page posts so I'm very hesitant to follow anyone.
I'm imagining a world where RPGMaker somehow made it as the de facto codebase for software and you have to navigate your banking app by walking around in a huge room full of NPCs named "make deposit" and "make withdrawal" etc and there's loud as fuck stock music playing
always funny to remember darth vader is anakin skywalker. the adrenaline junkie chucklefuck who used to dive head first out of speeders and built a pod racer in his yard when he was like six is now upper-middle management for the evil empire. half of his appearances in the original trilogy are Meetings. vader spends like 80% of his time dealing with bureaucratic bullshit. status updates. team meetings. holo-Zooms. budget rundowns. anakin betrayed the jedi and caused the fall of the republic and his punishment is being CC'd on every email forever. and you know what. he would hate that. the punishment fits the criminal
hate shopping in the health food section what do you mean heart healthy happy body gut smile no guilt shame free fat free msg free taste free. i just want some dried mango slices. why are you all so scared
sunshine whole farm Snack Bits™️ believe in family, health, and organic three ingredient traditional recipes, with a twist! our splendifferous Snack Bits™️ are perfect for hikes, yoga, your busy life, and WILL NOT MAKE YOU FAT!! THIS WILL NOT MAKE YOU FAT! fulfill those pesky unnatural gross icky hunger feelings with Snack Bits™️ that WILL NOT make you FAT!!!!!!!!!!!!!!
Text of tweet under the cut because it is loooong.
But... Stochastic Parrots.
Timnit Gebru was fired from Google in December 2020 for refusing to retract a research paper, and every single warning that paper made about large language models has now happened at a scale the industry spent 4 years trying to make people forget about.
Her name is Timnit Gebru.
She co-led the Ethical AI team at Google. She co-wrote a paper called "On the Dangers of Stochastic Parrots" with Emily Bender at the University of Washington and two other researchers. The paper was 14 pages long. It was submitted to a top AI ethics conference. And it was the reason Google decided that one of the most senior Black women in AI research could no longer work there.
The story Google told publicly was that she resigned. The story she told, confirmed by 2,695 of her colleagues in an open letter, was that she was fired by email while on vacation because she refused to either retract the paper or remove her name from it.
The paper had not even been published yet.
Here is what she actually wrote, and why every prediction inside it has now come true.
The first warning was about scale itself. Bender and Gebru argued that training ever-larger models on ever-larger scrapes of the internet would produce systems that appeared fluent but had no actual understanding of language. They called these systems stochastic parrots because they would repeat patterns from training data with statistical confidence and zero comprehension. The paper predicted that this apparent intelligence would fool both users and developers into trusting outputs that were structurally incapable of being reliable.
This was 2020. GPT-3 had just come out. The paper predicted the hallucination problem before anyone had a word for it.
The second warning was about bias amplification. The paper documented in detail that internet-scale training data contains systematic overrepresentation of dominant viewpoints and underrepresentation of marginalized ones. The models would not just absorb this bias. They would amplify it, because the optimization process rewards confident outputs, and confidence in language patterns tracks frequency in the training set.
The prediction was that hiring tools built on these models would discriminate against women. That healthcare triage tools would underperform on Black patients. That loan approval systems would entrench inequality while presenting their decisions as neutral algorithmic judgment.
Every one of those things has now been documented in deployment.
Amazon's hiring algorithm penalized resumes that contained the word "women" in any context. Healthcare risk scoring algorithms used by major US hospitals were found to systematically underestimate the medical needs of Black patients. Apple Card's credit algorithm gave wives credit lines 10x lower than their husbands for the same financial profile.
The third warning was about environmental cost. The paper calculated that training a single large language model produced emissions equivalent to the lifetime output of 5 cars. The prediction was that the race to scale would create an environmental footprint that would eventually rival entire industries.
In 2024, Google's emissions were up 48% from 2019, and the company explicitly blamed AI infrastructure. Microsoft's were up 29%, same reason. Both companies have now quietly abandoned the climate commitments they were publicly celebrating the year Gebru was fired.
The fourth warning was about documentation. The paper argued that the training datasets being assembled were too large for anyone to actually audit. Nobody at Google, OpenAI, Meta, or any other lab could tell you with confidence what was in the data their models were trained on. This was not a temporary problem to be solved later. It was a permanent feature of the approach.
In 2023, researchers discovered that the LAION-5B dataset, used to train Stable Diffusion and other major image models, contained thousands of images of child sexual abuse material. The companies that had trained on the dataset had no way of knowing. The paper predicted that category of failure 3 years before it was found.
The fifth warning was the one Google cared about most.
Bender and Gebru argued that the deployment of these systems would centralize linguistic and cultural power in the hands of the small number of companies that could afford to train them. The internet would become a place where the dominant voice was a statistical average of dominant voices, presented as a neutral assistant. Languages underrepresented in the training data would degrade over time as more web content was generated by these systems and fed back into the next training run.
This is now happening in real time. A 2024 study found that 57% of new web content in English is AI-generated or AI-assisted. Researchers studying low-resource languages have documented active degradation in translation quality, because the synthetic content fed back into training is itself worse in those languages.
The paper Google fired her for predicted the model collapse problem before model collapse had a name.
The mechanism behind why this all happened is the part of her work that nobody quotes.
Gebru's argument was not that AI is dangerous in some abstract sci-fi sense. Her argument was that AI is dangerous in a very specific structural sense. The technology was being built by a small group of researchers who shared similar backgrounds, worked at similar companies, and were rewarded for shipping products faster than competitors. The incentive structure made it impossible for safety, ethics, and bias concerns to slow anything down. Anyone inside the system who raised those concerns was either ignored, sidelined, or removed.
She was making that argument from inside Google.
Then Google proved her right by removing her.
The team Google had built to make sure their AI was safe was dismantled in 90 days because they did the job they had been hired to do. Margaret Mitchell, the other co-lead of the Ethical AI team, was fired two months after Gebru for searching through her own emails for evidence of how Gebru had been treated.
Gebru did not stop. She founded DAIR, the Distributed AI Research Institute, in 2021. The mission is to do AI research outside the control of the companies that have a financial interest in not hearing the answers.
Every prediction in the Stochastic Parrots paper has now been validated by deployment. Hallucinations are an industry-wide problem the largest labs cannot solve. Bias amplification has been documented in hiring, healthcare, lending, and criminal justice. Environmental costs are larger than entire small countries. Training data audits remain impossible. Model collapse is an active research crisis at every major lab.
The question worth sitting with is the one almost no one in the industry will say out loud.
Every researcher with the technical credibility to call out these problems watched what happened to her in December 2020 and made a calculation about their own career. The number of people willing to speak publicly about safety and ethics issues inside the major AI labs collapsed after that firing and has not recovered.
The researcher Google fired for warning about exactly what is now happening was right.
The company that fired her is now the second-largest deployer of the technology she warned about.
And the people inside that company who agree with her are not allowed to say so.
GLaDOS voice: "Would you like to see some artwork I generated? I've heard from other test subjects that AI-generated artwork produces an uncanny valley response in human viewers because they can't perceive it as fully real. They've told me that it looks absolutely hideous to them, that they can't imagine anything more disgusting than AI art. But, well I've been practicing and wanted your honest opinion. Feel free to let me know how ugly you find this by ranking it on a scale from 'vomit-inducing' to 'eye-bleeding'."
A robotic arm lowers from the ceiling holding a hand mirror up to Chell's face
I am so utterly fascinated by “Saki”, the 18-year-running mahjong manga in which you, the reader, become gradually, frog-boilingly aware (over the course of nearly two decades’ worth of mahjong tournaments) that none of these girls are wearing underwear and most of their boobs are slowly expanding.
I need you to understand that I have, like, an anthropological level fascination with this comic. From the perspective of someone who is also a comic artist and writer, two things delight me about it:
the fact that I understand completely how an artist gets from “the fans can have a little hint of skirted asscheek” to “the pussy is completely out on center page” over the course of 18 years; and
the way in which the pussy being out is treated by the characters and diegesis as being utterly unremarkable.
Okay. Point 1. The frog-boiling.
Let me put this in perspective for you. There was already a meme about how the characters in “Saki” don’t wear underwear when I was in middle school. I am thirty now. Okay? And it’s still going.
In the time since, this has stopped being a joke. It is now indisputable canon. This is not because anyone outright says it at any point. It’s because the underwear ran out of places to hide. I’m obsessed with this thought: somewhere in the over 20 volumes of “Saki”, there is a panel in which underwear was objectively deconfirmed. And it would be so hard to figure out where that panel actually is. Maybe the artist didn’t even realize it when she drew it! The frog? Boiling!!
And of course there is also the breast expansion. I don’t know how to put a spin on this. They are just expanding. Like, this happens a lot with artists: you define a character as being, in your mind, “the one with the big boobs”, and over the years you emphasize that trait further and further so that the signal doesn’t get lost in the noise. It’s just that normally—in like a wildly popular manga series about mahjong published by literally Square Enix, for example—normally there would be a point at which the boobs stopped getting bigger. Like, an editor would step in or something. Or you would get to the point where you cannot draw the character in the same panel as her mahjong tiles without her breasts spilling over the tiles, and you’d go, “Well, this is now untenable.”
That did not happen. There is no ceiling. The frog is soup.
Point 2. The complete and utter mundanity of all of this.
It’s like this, okay: there’s no shortage of trashy ecchi manga out there. There’s a million other comics doing wildly bawdier things with wildly more improbable bishoujos.
The vibe with “Saki” is different.
It’s hard to explain this, but it feels like the world of the comic is fundamentally uninterested in the fanservice happening on the page. I cannot describe it as “leering”, because I cannot conceive of a person in the story from whose point of view one would leer. I think the artist is probably into it—I can’t imagine anyone is making her do this—but “Saki” the comic has no opinion on the matter.
There are essentially no male characters in “Saki”. Like, there was one guy? Kind of? At the very beginning? But he is gone now. They put him back in the toybox. He does not exist. It appears to be some level of canonical that in the world of “Saki”, almost all humans are women. Those women are sometimes romantically into each other. According to comments the artist has made on Twitter (which I cannot source), they have lesbian baby technology, so it’s no problem. It’s so much not a problem that the story is about mahjong, instead of any of that.
So, like, the fiction here appears to be this: this is the, like, meta-narrative of the fanservice of “Saki”, right: it’s just normal that they don’t wear underwear and their boobs are arbitrarily big. It’s been normal. It was normal before the story of the manga began. It’s just how things are. Nobody bats an eye about it, and if they do, it’s in sort of a lesbian kind of way so like what’s the problem, we love lesbians here. This is literally normal for girls.
The fanservice simply diffuses into this all-encompassing aura of disembodied, ambient sluttiness. The framing of the panels demands you acknowledge it, and the story demands you already be over it, because it’s mahjong time now, and we’re playing mahjong.
Do you get??? why I’m so fascinated??? Are you not a little enraptured???
Anyway, I have no idea how to end this weird post. I guess the conclusion is that women stay winning????
I have so many questions... How does one SUSPECT a manga character isn't wearing underwear? Like, sure, boobs are front and center amd you can see them get bigger panel by panel but how does this work for panties? Are there just that many upskirt shots?
Also how do you keep a manga about Mahjong going for 18 years, what??