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.
Scientists invented a fake disease. AI told people it was real: Nature.com
I'm a bit frightened for the time when someone less ethical than the person that did this decides to repeat the experiment but leave out the part where they come in later and announce that it was fake and people wind up diagnosed with the fake condition and all kinds of wacky hi jinks ensues.
for those wondering why they're free to take now, it's because the company that made those "chicken soup for the soul" books bought them a few years ago and then completely collapsed so bad they couldn't afford to dispose of or even take the blu rays and dvds out of their kiosks all over.
so any of them is free game because they're all located on other business' property and they usually don't want to have to pay to get rid of them either. so asking the store manager usually gets you the ok to pull it out and keep it.
there was a period of time right after their bankruptcy where you could put in any debit or credit card and it would spit out movies without charging you. you could even put in like an expired or deactivated card, or a visa gift card with a $0 balance, didnt matter, they'd just start spitting discs out. a lotta people raided redboxes for movies for a couple months, with some people doing what me and my brother and my dad did here, taking the whole box and signs and marquees as well. because managers sure as hell don't want a big abandoned piece of trash on their sidewalk disappointing customers. BUT they're also often too cheap to pay someone to remove it. so they just sit there.
luckily there are no shortage of freaks like us who will just take them away on our own volition. we did it all "by the book", too: we set up cones and caution tape, disconnected electricity properly, used an angle grinder to grind down the bolts in the concrete so nobody would trip on them, then cleaned everything up afterward and sealed off the electrical panel so the store would know everything is safe and tidy. though they were hesitant when we were first contacting them, they were honestly very relieved and grateful when we finally took it away, especially once they saw that we "knew what we were doing" (we don't) and look like we've "done this before" (we haven't).
the fun part: the reason why this redbox, in particular, was completely full and unraided is because the computer hardware inside had failed some months before the bankruptcy, and a failing company sure as hell wasn't gonna send a tech out to our podunk dipshit city to fix it, so it was impossible to rent movies or take any discs out. plus, for who knows how long, people were returning old redbox discs to this machine and not taking any out, leading to a much higher variety of movies than your average redbox.
there is a thriving community of redbox hackers and modders out there, as well, creating open-source software for repurposing the machines and not letting their very interesting and robust disc-management hardware go to waste. this one belongs to my brother (who was very annoying persistent and did all the legwork of contacting managers and securing permission) who is a programmer by trade and will be hacking it into a family-access movie library, with whatever discs we want. i mean the machine is completely weatherproof and has a built-in AC unit, it would be such a waste to not try to turn it into something cool.
if we get another one, i'm gonna try to mod it into some sort of art or zine vending machine. the disc boxes are just the right size for small print art or stickers. would make a great "little free library" too.
remember: the rules are made up. act like you belong there and you can get away with anything. this applies to your own life
this messed up vintage cat sewing pattern has tormented me since i saw it & like some other folks have done in that post - i tried my hand at tweaking the pattern to resemble the illustration (and my personal tastes) a little more. i've ended up with this, which i have only tested at a small scale and not this final version exactly (where i have done such things as further widening the cheeks and finalizing the leg shapes.) i bestow it upon you nice folks now 👐
go forth and make weird little beanbag kittens! pls show me if you do!
My first attempt! I made the pattern a bit smaller as I wanted it to be able to fit in a pocket, but then (accidentally but perhaps unavoidably) sewed it with a wider seam allowance than the resized pattern indicated, so the face is proportionally a bit too big and I lost some detail in the ear shape. I'm pleased with it though! It was fun to make something and to do some handsewing.
Harry Belafonte and Odetta inventing acting performing “A Hole in the Bucket,” which i am only uploading here because the source on youtube (x) has incredibly low volume.
Hello, I want to use my blog here to boost someone in need.
Dile and their mother have been homelessness for some time now. They're in need of funds for food, travel, medical care, and hotel stay. Dile just got of the clinic after being very ill and it would be ideal if they were able to be indoors to recover. Currently, at least $700 would be enough to move things to a more favorable state.
Here is a link to Dile's paypal.
Dile is a friend of a friend. They used to post amazing artworks onto tumblr before their current circumstances. You may remember that I've reblogged their fundraising posts including a gofundme before. Unfortunately, if you know anything about homelessness, it's incredibly hard to get out of once you're in it. Its just not something that is resolved with one instance of aid. Its an ongoing campaign.
I know from experience that reach and spread is so important to fundraising. All it takes is getting to the right people. If you don't personally have the means to help (understandable in this current time), you still have the power to help by getting word out.
The iron hook slid free from his shoulder with a wet metallic shriek. Something black and arterial splashed across the stones between them.
The torturer stepped back instinctively. Not out of mercy. Out of surprise. The prisoner laughed. Not loudly. Worse than loudly. Softly. Like he had just remembered a private joke older than civilization.
“You still think pain is a language,” he said.
Another blow. This time across the mouth. Teeth cracked. Blood sheeted down his chin in long ribbons.
The interrogator hissed through clenched teeth. “Tell me where God went.”
The prisoner turned his head slowly. There was blood in his smile now.
“There are organisms,” he said, “living beneath Antarctic ice that have never seen the sun and have still learned how to eat.”
The room had gone very still. Somewhere in the dark, machinery groaned.
The interrogator grabbed him by the jaw hard enough to bruise bone.
“You think this makes you immortal?”
The prisoner spat a clot of red onto the floor between them.
“No,” he whispered.
“I think it makes you temporary.”
The torches flickered.
For one impossible second, the interrogator became aware of his own pulse. The heat in his veins. The soft wetness of his eyes. The damp animal electricity inside every living thing. The prisoner watched realization bloom across his face and smiled wider, blood running between his teeth.
“You cannot threaten a creature from the dirt,” he said, “with returning to the dirt.”
— excerpt from Shit I Just Made Up To Exemplify How All This Tumblr Prose Sounds
blakekathryn
kianamaiart
theartofmadeline
