hello welcome to my account. I like science, especially biology! I'm into many fandoms so there's a not so low chance that I might just be interested about what you talk about! let's see.. what else? I'm an adult. And I like the theoretical magenta.
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.
from what i hear, we have 'till about thursday morning in the US to make our complaints and pressure them into not doing this. given there's been pressure on them for the better part of the year and they're still trying to give it a push, we gotta redouble our efforts, like, tenfold. if these laws go into effect, then you're gonna have to kiss the internet goodbye. everything that you valued about it will be gone, and it's going to be a long, long, LONG time (if at all) for it any of it to be reversed.
this cannot happen. if it does, then any sort of criticism of the US government can easily be shut down one way or another. coverage of the attacks on foreign nations as anything other than an absolute atrocity? gone. your creations? your art? your fan art? your smut? all gone and heavily restricted.
this is all a massive ploy to force you out of your anonymity and control the speech and content of the internet to such a degree that it becomes pointless to use. i repeat, THIS MUST NOT HAPPEN. if you hated the integration of discord's age verification, then you gotta scream loudly against the laws that would force them to implement it. same goes for other services that would do the same if these laws went into effect.
currently, these are the best ways to contact your lawmakers, congresspeople and whatnot to get them to shut this shit down. a lot of congresspeople who made sure this passed already know plenty of the risks involved, so if you call, in addition to you telling them what these bills will actually do, let them know that this is a dealbreaker for you. put on the pressure. make them make the right decision, even if they don't want to. and if they're conservative, pull out the buzzwords. freedom and justice and whatnot. pack on the good ol' guilt.
All of the bad internet bills. One website.
Section 230 of the Communications Decency Act has protected free speech online for decades, but now there is a bipartisan effort in Congress
Congress has again put forward legislation aimed at controlling what content children see online, known as the Kids Online Safety Act (KOSA)
Find your members of Congress by typing in your address on Congress.gov.
but whatever you do, please do it. this is our big chance to shoot this down, and if we don't, then i hate to fearmonger, but it's pretty much over. as of this post, we have just shy of 48 hours to flood their inboxes over phone or email.
DO NOT LET UP. DO NOT GIVE IN. PROTECT YOUR RIGHT TO FREEDOM OF EXPRESSION.
from what i hear, we have 'till about thursday morning in the US to make our complaints and pressure them into not doing this. given there's been pressure on them for the better part of the year and they're still trying to give it a push, we gotta redouble our efforts, like, tenfold. if these laws go into effect, then you're gonna have to kiss the internet goodbye. everything that you valued about it will be gone, and it's going to be a long, long, LONG time (if at all) for it any of it to be reversed.
this cannot happen. if it does, then any sort of criticism of the US government can easily be shut down one way or another. coverage of the attacks on foreign nations as anything other than an absolute atrocity? gone. your creations? your art? your fan art? your smut? all gone and heavily restricted.
this is all a massive ploy to force you out of your anonymity and control the speech and content of the internet to such a degree that it becomes pointless to use. i repeat, THIS MUST NOT HAPPEN. if you hated the integration of discord's age verification, then you gotta scream loudly against the laws that would force them to implement it. same goes for other services that would do the same if these laws went into effect.
currently, these are the best ways to contact your lawmakers, congresspeople and whatnot to get them to shut this shit down. a lot of congresspeople who made sure this passed already know plenty of the risks involved, so if you call, in addition to you telling them what these bills will actually do, let them know that this is a dealbreaker for you. put on the pressure. make them make the right decision, even if they don't want to. and if they're conservative, pull out the buzzwords. freedom and justice and whatnot. pack on the good ol' guilt.
All of the bad internet bills. One website.
Section 230 of the Communications Decency Act has protected free speech online for decades, but now there is a bipartisan effort in Congress
Congress has again put forward legislation aimed at controlling what content children see online, known as the Kids Online Safety Act (KOSA)
Find your members of Congress by typing in your address on Congress.gov.
but whatever you do, please do it. this is our big chance to shoot this down, and if we don't, then i hate to fearmonger, but it's pretty much over. as of this post, we have just shy of 48 hours to flood their inboxes over phone or email.
DO NOT LET UP. DO NOT GIVE IN. PROTECT YOUR RIGHT TO FREEDOM OF EXPRESSION.
tony's chocolonely mildly pisses me off every time i eat it. like yeah i get it ur doing a whole symbolism thing about unfairness and whatever, but its actually SO DIFFICULT to eat fucking. gerrymandered chocolate. your symbolism is ruining my chocolate experience.
like. ur already more expensive. i understand this. i am willing to pay more money to have slave-free delicious chocolate. why must you punish me further by making it a goddamn puzzle to break a piece off.
Tony's Chocolonely does not ensure a living wage for cocoa workers in West Africa (see 2020 report by Voice Network).
Food Empowerment Project maintains an updated list of chocolate brands they recommend based on the following standards:
transparency about the country of origin for cocoa
if sourced from Western Africa, the workers must own the companies/be in charge of the profits from their labor (this standard is because child labor and slavery have been widely documented for decades in this area)
if sourced from Brazil, the workers must own the company and/or be in charge of the profits from their labor, or the company must be going above and beyond to support the workers and their families (this standard is because "child labor and slavery only recently have been documented in cocoa in Brazil" and they are "giving Brazilian companies that are trying to address the issue by supporting the workers and their families the benefit of the doubt.")
this isn't perfect, obviously, but it is much more grounded in the rights of workers than the various certifications (e.g. Fair Trade, Rainforest Alliance, etc.) you'll see on product packaging
lot of terfs have been reblogging this so I may as well publicly state that the woman on the right is modeled with permission after my transfemme friend. if you relate to it as strongly as many of you claim in the tags I urge you to reflect upon that with empathy and compassion about the depth of experiences you truly do share with trans women.
otherwise fuck off I guess. my art is not fuel for your hatred.
There are a million and one correct reasons why "corporate sponsored tumblr influencer making money" would never work. But one reason which I rarely see mentioned is that, for every Popular Post By Popular Tumblr User, only a small percentage of its notes came directly from source. The rest came from spread.
Like I genuinely think there is no singular Tumblr user with enough active followers to produce a Popular Post without reblogs that breach well outside their circle. There's a few who could probably get a ~1,000 note likes-only post on a post with reblogs disabled. Which is really not doing numbers compared to some standard-fare tumblr post like "enjoying my penits and crackerjacks" with 10,000 notes from some OP with 300 followers.
Posts cannot do numbers without the broader tumblr community's say-so. And the vast majority of those people do not give a shit about any single OP. And a single OP, even with top-tier follower count, could never generate post escape-velocity by shilling bullshit no one wants to people who are actively hostile to advertising.
janainaart
jacobvanloon
origamiaround
jbbartram-illu
