to anyone in the areas impacted by the wildfire smoke, my #1 biggest piece of advice as someone whos been dealing with wildfire smoke in the NW united states for years, is build yourself a Corsi-Rosenthal Cube
they perform as well as expensive HEPA air cleaners, and are comparatively VERY inexpensive. all you need is a box fan, 4 air filters, a piece of cardboard, and some duct tape!!!!
i think it took us maybe a half hour to put ours together, if that, and we replace the filters every 3 months. it's really made a HUGE difference, both when the air quality is bad, but also with our allergies
I'm fascinated by how the animators for TADC handle Caine's impossible anatomy so well. There's so much to balance and they've somehow engineered a solution for his expressions to hold during speech.
I also traced Caine answering the phone from episode 4 to demonstrate his phonemes and mouth shapes.
A WASP FLEW INTO MY MOUTH! And I was so brave about it! I couldn't panic because I am allergic and would have died literally. So I just exhaled really strongly and blew it right back out again! To anyone nearby it probably looked like I simply exhale live wasps. Like that's a thing I can just do. There's that.
I can still feel all of its flailing legs in my mouth. Cannot recommend. You'd think the wings of a live wasp, in your mouth, would be the worst part but you'd be wrong. It's the legs.
I never thought I was going to hate a story more than the time my mom was surprised by a spider, gasped, and sucked it right into her mouth and down her throat and I hate that story and I do NOT like this one either.
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.
I’ve said “I hate this” so many times on this website, and never actually meant it, because “I hate this” is just shorthand for ‘this is an example of a meme given a twist I wasn’t expecting with intent to surprise’. Which is, in of itself, a meme on this site. God damn it.
But this… This is something else.
The rapidity of a meme’s introduction to its zenith to its decline is so rapid that in ten years, you’ll need a damn twenty-page manual to explain this. It’ll be as unfunny and hard to explain as jokes in Shakespeare plays, except even more inexplicable because fuck, at least Shakespeare’s jokes are usually about anal or fucking your mother, good wholesome sex jokes we can all get behind.
For the love of fuck, how do you explain loss.jpg? How do you explain gun?
it is a YMCA reference - that’s one of the 6 memes being represented here
ok let me see if i can break this down easily. YMCA is the easiest place to start - the song itself has become a meme over time with people changing the lyrics to reference other pop cultural events. so YMCA is meme one (1)
this first lyric replacement (”take the breadsticks and run”) is a reference to the tumblr meme ‘stuffing breadsticks into my purse’. i think everyone remembers that one so i wont bother to explain it. that’s meme two (2)
“man door hand hook car door” is a meme of its own, a creepypasta from i dont remember when. it was a terrible stupid retelling of the generic ‘stuck in a car while hook handed man tries to kill us’ story so the stupid title caught on for memorability. that in and of itself is meme three (3)
‘gun’ is… yeah i dont know how to explain gun. long story short you add gun to the end of a phrase instead of what you expect the last word to be. its shock funny. its everywhere but its popular to add to “man door hand hook car door” for.. some reason? gun is meme four (4)
and the thing is, this four meme combo is something thats gone around before. meme combos are, itself, a meme. which means taking this meme combo and mixing in another meme actually becomes meme five (5)
which leaves us at loss.jpg. loss.jpg was a terrible bad comic supposed to be about some tragic event, but it was presented so poorly literally no one takes it seriously, and for some reason recreating the four-panel setup has become popular. so thats meme six (6)
(but i need to add that this is the greatest version of loss.jpg i think i’ve ever seen. the initial ‘young man’ lines up with the guy bursting through the door, and the shock meme ‘gun’ matches the shock scene of the woman in the hospital and idk if OP even thought about that but it makes this just so much better)
it might also just be a coincidence due to loss.jpg’s format but the whole white minimalist four-panel setup is also suspiciously reminiscent of those early 2000’s rage comics
I’m not reblogging this because I want to. I’m reblogging this because sometimes you’re a witness to history whether you want to be or not, and you have to embrace that.
Ok but I need someone who can actually make a meme to add on but do the Thor ragnorak meme where tumble is Hela and she’s like you can’t defeat me and Thor is that first original plane white meme and he says I know but he can and the fire guy is the most recent one. Thanks
New illustrations of kid!Ruggie and kid!Leona by Yana!! I believe it is in the third volume of the Episode of Savanaclaw manga as a bonus at the end of it.
According to the notes, Ruggie and other hyenas have entirely black hair in their childhood. I assume this is because real life hyenas have entirely black fur in their youth as well. I guess when they grow up, their hair/fur lightens, hence the gradient we see in Ruggie's current hairstyle. It looks like beastmen have more traits in common with their animal ancestors than we thought! (We've discussed how animal traits might manifest in beastfolk infants here.)
The other note on Ruggie's illustration says that he wears someone else's hand-me-down clothes.
As for little Leona! It says he aspires to be the King of Beasts awww 🥺he used to be so full of hope... and the necklace he wears bears a scar-like pattern. It is apparently imbued with a protective spell, which was cast by Kifaji.
We previously discussed what the purpose of the pendant the young royals of Sunset Savanna seem to wear (here!). According to the Magical Archives, Cheka's seems to have a tracking device but no protective spell is mentioned. Leona's meanwhile, has the protective spell but no tracking device mentioned. Are these customized based on the personality (and/or troublemaking level) of the royal in question? Is Cheka's pendant being more tech than magic an indication of the Sunset Savanna's advancements as they slowly industrialize while keeping mindful of nature? Or do the pendants just have both anyway and the design notes fail to convey that to us?
Cheka seems to sneak away from his retainers a lot, so I get why the GPS tracker might be needed for him. I have to wonder why Leona has a protective spell on his, though... Are there more people after him than Cheka, despite him being the second born prince and not the crown prince? And isn't Leona a mage powerful enough to be able to defend himself whereas Cheka doesn't seem to be able to defend himself and/or cast spells yet? Maybe Kifaji cast a protective spell on Leona's pendant to help Leona feel more like a normal kid? Like someone is out there hoping for his safety and worrying about him...
#excuse me but are you telling me that the Apollo pic is made with the help of the SUN and the Artemis one with the help of the MOON??? #that's actually so poetic i want to cry
@gorandomshesaid wait i need to sit with this one. wait.
Please keep interacting with this post because when I come to tumblr to procrastinate, this shows up again in my notifications and guilts me into writing again