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.
imagine beethoven, but with a sick drop. but be careful because i started doing that once and now i can’t stop imagining beethoven but with a sick drop
Does no one realize how racist this assumption can be? Most LLMs are trained heavily on Commonwealth and other standardized English corpora, yet now when people from Commonwealth countries naturally write in polished English, others immediately say it “sounds AI-generated.”
I fear this is the beginning of a really awful trend that will make it even harder for non-white writers to get published.
Got curious, so I went and read it myself. The AI accusation is completely absurd to me. The story is small scale, personal, laden with metaphor, and clearly draws heavily from the writer's cultural history. It's not conventionally told, but the ideas set up in the beginning are woven throughout the narrative nicely - nothing is extraneous, no threads are dropped, and it ends on a thoughtful and somewhat poetic note that explores its core themes. Unless I'm sorely mistaken, this is not kind of writing AI generally produces (at least not without significant human intervention - at which point who cares?)
The idea that it's AI generated because of a couple difficult-to-parse similes (in a piece that employs flowery simile multiple times per paragraph) is so insidious. Oh I'm sorry, this Trinidadian writer's piece exploring the cultural intersections of the Carribean and Indian diasporas on the island wasn't instantly understandable to me, an ignorant anglophone reader - so therefore he must be a fraud? Ridiculous, and in my opinion clearly racist.
How very depressing that Neil Gaiman had trended not even a tiny bit for demonstrating what a fucking horrific person he is.
As a reminder, he's suing Caroline Wallner, one of his accusers, for breaking her NDA. Not for libel. He's saying she shouldn't have told anyone about it, not that she lied.
The author says Wallner broke her NDA by sharing her story with the media, including with New York Magazine.
He doesn't need the money. He's risking the Streisand effect. He is punishing Caroline, he's trying to intimidate other victims who have signed NDAs to scare them into continued silence.
He is no friend to women, to the LGBTQIA+ community, to anyone quite frankly unless he thinks they are of value to him.
Share the story. Put it on Facebook and bluesky and whatever else you're on. Make it clear what a horrifying person he is. Tell your friends. He's paying Edendale a fortune to try and cover this up. Make this hard for him. Make it cost him money.
"This week I discovered the same pattern, executed by Google. Google Chrome is reaching into users' machines and writing a 4 GB on-device AI model file to disk without asking."
Google Chrome is downloading a 4 GB Gemini Nano model onto users' machines without consent, with no opt-in, no opt-out short of enterprise t
"There has always been a nagging discrepancy between the promise and the reality of white maleness. White men have often had the sneaking suspicion that the American dream is a fiction. At the same time, white men have always feared the potential of losing that one great superiority- the better than. If all you have is better than- better than women, better than people of color, no more and no less than that- why would you willingly give up the one prize you never had to earn?
And so these white men often end up clinging to a disturbing construct and blaming those with less power for its shortcomings. This seems for some like a better option than questioning the promise great reward ahead of them. After the wealthiest white men take their cut, there is never enough left for the average white man to have his crown. But we, as a society, continue to tell white men that their coronation is just around the corner."
Scam accusations get thrown around far too easily. Simply asking for money is not the same as scamming. A scam involves deception, lying, misrepresenting one’s identity, or failing to deliver on promised exchanges, such as unfulfilled orders, or identity theft. That’s what defines a scam.
When someone is transparent about their situation, clearly explains what they need, and is honest about how the money will be used, there is no deception involved. You may choose not to give, but calling it a scam is inaccurate. It’s not fraud. It is simply a plea for help.
Would you call someone a scammer for sharing a registry or a wishlist? Of course not because there’s no deception involved. They’re simply stating what they want or need and giving others the option to contribute. That isn’t scamming; it’s just a request.
Today is Armenian genocide remembrance day. On april 24, 1915 started mass deportations of hundreds of Armenian intelectuals and community leaders, who were (most of the time) eventually killed. Armenian women and children were systematically r//ed and forcibly converted into islam. There were more than 2 milion Armenians in ottoman empire prior to ww1, 1,5 milion of them were viciously killed. Three millennia of Armenian civilaziation in eastern Anatolis was fully destroyed. Turkey today refuses to acknowledge genocides of christian minorities in early 20th century.
Do you know that mass ethnic cleansing of Armenians in ottoman empire inspired Lemkin to coin the term 'genocide'?
Last year in september azerbaijan allied with turkey initiated a war against Armenia. More that 5000 Armenians were murdered, thousands of Armenia families had to live their ancestrial land to not get murdered. There are hundreds of vids on internet where armenian p.o.w.s are tortured. Recently azerbaijan opened a "museum" displayind dead or dying Armenians and kids were allowed to visit it.
Please educate yourself on Armenian genocide. You can also donate here to help Armenia. Thanks for reading!!
Last September, Artsakh was completely cleansed of its indigenous Armenian population. 3000 years of history were violently disturbed. Now the anti-Armenian propaganda and hatred is increasing and they are openly talking about taking over all of Armenia. Please take your time to read about these issues.
An abandoned theater in Rochester, New York. Formerly a pornographic theater, then just a facade after the Walgreens next door gutted it to use as a warehouse. The Walgreens with it's shampoo and baby formula and half of the store locked in cages; the contradictions stare into you. You go because it is the nearest pharmacy, close in walking distance for you and your disabled loved ones.
The Walgreens shuts down and it too, is now just a vacant facade, next to a vacant facade. You stare into the large windows meant to advertise it's contents, now only showcasing absence. A wide open space with torn up floors. You think about how perfect the location could be for low income housing, for squatting, for anything other than standing as a constant reminder of the city's failures to her people. The parking lot is empty, save for the occasional cop car, to ensure no one uses the lot or the building for anything at all. The cop ensures the space will stay as useless as himself.
They add letters to the theater's marquee; a reminder and encouragement of surveillance. "If you see something, say something. In progress call 911, over and done call 311." They board up the doors, ensuring that the space will stay in it's intended form: empty, useless, an eyesore that reminds everyone of their place. The freshly boarded up door has been emblazoned with a message from those who lack.
"A man would shelter, if he could / in the nook behind this new plywood / the building, abandoned / the man is too / how I wish you'd imagine / that it were you."
"Military-aged" means civilian. Anytime you read it, or hear it, you should think "civilian". When Bush and Obama and Biden were killing "military-aged" men they were murdering civilians and they knew it and everyone should recognize that. When your brother or sister or cousin or mom or dad or friend in the military killed "military aged" people, your loved one was committing murder.
gummygunk
vaporwave
yousharknotpass
pixelskylines
prguitarman
dailybridgerton
kianamaiart
