Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
✓ Live Streaming✓ Interactive Chat✓ Private Shows✓ HD Quality
Anya is LIVE right now
FREE
Free to watch • No registration required • HD streaming
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.
tl;dr: all "algorithmically" pushed stuff on a newsfeed is mostly ads. nothing that's really surprising form this vulture article, but it is dismal and makes me grateful for one website where you only see things from people you follow WITHOUT horrible short-form video content
Listen, if you think this isn't happening on Tumblr, it is. Obviously, you're being exposed to stealth advertising if you use the algorithm, like the For You page or Explore. But even if you don't and you just use the Following dashboard, stealth advertising can trickle down to you if the people you follow use the algorithm, or if they reblog from people who use the algorithm, or if they reblog from people who reblog from people who use the algorithm.
Or if anyone you follow, or anyone they or their followers follow, actually are one of these paid stealth advertising accounts without knowing it.
Odds are good that you've seen AT LEAST one ad in disguise if you've been on Tumblr for any substantial amount of time. You may have recognized it, you may not have. I barely use the algorithm, and I know I've seen a few stealth ads that I recognized as such over the years. Probably a few more that I haven't.
#genuinely: this is what further research into the famous 'marshmallow test' showed#it wasn't that kids who were able to delay gratification were more likely to be successful later on#due to intrinsic qualities#it was that kids who had a stable upbringing were more likely to be successful#and ALSO: those kids had trust that their caregivers would keep promises#which is WHY they were willing to give up one marshmallow now for the promise of two marshmallows later#kids who did NOT have trustworthy caregivers#or who were in a fundamentally unstable situation#DID NOT have that trust so they wanted their one marshmallow NOW#same deal here i think#it's not that Gen Z is bratty#it's that they have no trust in the system and no faith that promises will be fulfilled#and frankly i do not blame them -@cicerfics
What do you mean “chat” is now referring to ChatGPT and not twitch chat? What? What? What the fuck? No?
When I address chat I am speaking to a presumed Greek chorus of real human people shitposting on their lunch break, not a machine that devours lakes to covert electricity into slop.
Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
✓ Live Streaming✓ Interactive Chat✓ Private Shows✓ HD Quality
Anya is LIVE right now
FREE
Free to watch • No registration required • HD streaming
“The works of the roots of the vines, of the trees, must be destroyed to keep up the price, and this is the saddest, bitterest thing of all. Carloads of oranges dumped on the ground. The people came for miles to take the fruit, but this could not be. How would they buy oranges at twenty cents a dozen if they could drive out and pick them up? And men with hoses squirt kerosene on the oranges, and they are angry at the crime, angry at the people who have come to take the fruit. A million people hungry, needing the fruit- and kerosene sprayed over the golden mountains. And the smell of rot fills the country. Burn coffee for fuel in the ships. Burn corn to keep warm, it makes a hot fire. Dump potatoes in the rivers and place guards along the banks to keep the hungry people from fishing them out. Slaughter the pigs and bury them, and let the putrescence drip down into the earth.
There is a crime here that goes beyond denunciation. There is a sorrow here that weeping cannot symbolize. There is a failure here that topples all our success. The fertile earth, the straight tree rows, the sturdy trunks, and the ripe fruit. And children dying of pellagra must die because a profit cannot be taken from an orange. And coroners must fill in the certificate- died of malnutrition- because the food must rot, must be forced to rot. The people come with nets to fish for potatoes in the river, and the guards hold them back; they come in rattling cars to get the dumped oranges, but the kerosene is sprayed. And they stand still and watch the potatoes float by, listen to the screaming pigs being killed in a ditch and covered with quick-lime, watch the mountains of oranges slop down to a putrefying ooze; and in the eyes of the people there is the failure; and in the eyes of the hungry there is a growing wrath. In the souls of the people the grapes of wrath are filling and growing heavy, growing heavy for the vintage.”
do they think... that prologues and epilogues... which take place within the story and are very much inside of the same book as the rest of the pages... don't count as part of the book....???.....
A few weeks ago, a small but higher-than-normal number of accounts were mistakenly suspended. The suspensions were quickly reversed, but our response wasn't good enough, and we want to say more.
We're sorry it took us this long to address this. Trust and safety issues are difficult to discuss publicly, and we can’t share details about specific individuals or how our systems work without exposing Tumblr to bad actors. But caution led us to say too little, too late.
We’ve heard from members of the trans community on Tumblr that they were disproportionately impacted, and that deserves a direct response. According to 3rd party researchers, Tumblr’s userbase has the highest proportion of LGBTQIA+ folks on social media, so it makes sense that when something goes wrong, those communities might feel that disproportionate impact.
One thing we want to emphasize is that we do not moderate people's identities. We moderate behavior. We know that identity shows up across a Tumblr profile in many ways, from followed topics to the flags people put in their bios, and more. However, these signals play no role in how our moderation systems make decisions. We monitor those systems for evidence of bias and take corrective action when we find it.
We understand that the communications sent to affected users, and our broader silence to the community, didn't meet the standard people expect from us. That feedback is fair, and we apologize. We've updated the messaging sent to people impacted by these incidents. We are also overhauling our process with a goal of responding to mistaken suspension appeals within 24 hours, and have instituted an ongoing internal review of how suspensions and appeals are handled.
Going forward, we're committed to finding a better balance: being more transparent with our community about issues that matter, even when we can't share everything.
Tumblr belongs to everyone. We take that seriously, and we intend to earn back your trust. We are not afraid to have tough discussions with you or make updates based on your feedback, though on occasion, it might take some time.
To the people whose accounts were affected, and to the members of our trans community who felt targeted: we are truly sorry.
Providing my own data on this issue, since i have had the pleasure of gathering a great amount of evidence on this event:
a small but higher-than-normal number of accounts were mistakenly suspended
Two hundred within five minutes.
The suspensions were quickly reversed
Some suspensions were reversed.
We’ve heard from members of the trans community on Tumblr that they were disproportionately impacted
Correction. All members harmed but one were trans women, specifically, not just "members of the trans community".
We moderate behavior.
All members terminated were terminated after following one specific blog, created a few hours before. The "behavior" you moderated was following a trans woman's new blog.
However, these signals play no role in how our moderation systems make decisions.
Then explain why trans women are disproportionately banned on false grounds. What happens if someone gets mass reported by transphobic userrs? Why have you chosen to restore the account of a kiwifarmer, known hate group, and ban yet more trans women? Explain your decisions.
We are also overhauling our process with a goal of responding to mistaken suspension appeals within 24 hours, and have instituted an ongoing internal review of how suspensions and appeals are handled.
Going forward, we're committed to finding a better balance: being more transparent with our community about issues that matter, even when we can't share everything.
Since this event, trans women have been terminated at greatly accelerated rates, on more and more openly false grounds. Their appeals have been denied often without any human oversight, recieving a denial within seconds, or not recieving any answer.
To the people whose accounts were affected, and to the members of our trans community who felt targeted: we are truly sorry.
Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
✓ Live Streaming✓ Interactive Chat✓ Private Shows✓ HD Quality
Anya is LIVE right now
FREE
Free to watch • No registration required • HD streaming
The ground in these places is too compact for water to soak in during wet season which leads to flooding but digging these holes gives the water a place to stop and soak in. And they’re pushing back the desert with this. By just digging holes.
The new plants also help even more water soak into the ground which reduces flooding even more.
These places also give people places to grow food and graze animals like people are turning completely dry compact desert into a refuge for wildlife and plants and solving regional food insecurity just by digging holes.
The half-circles are called zaï! They're a traditional farming practice in the Sahel desert, and their introduction + reintroduction can be largely credited to Yacouba Sawadogo, the man linked above! He reintroduced and innovated on the zaï on his own farm in the 1980s, and did extensive outreach (along with scientist Mathieu Ouédraogo) to encourage other farmers to adopt them as well.
He also promoted the use of cordons pierreux, which are basically just lines of rocks to reduce erosion, preserve sediments, and increase water absorption.
Immensely cool dude. He's been a personal hero since I learned about him.
Ooooh, Mr. Sawadoga innovated the traditional zai method by adding manure and other biological matter to the holes! This put nutrients in the soil as well as helping even more with water retention and attracted termites whose tunnels helped loosen the compacted earth, all of which supported plant-growth like no zai before! Which increased water-retention even further! Oh excellent, excellent work!
slightly more serious less flippant version of my post from yesterday: whether you ar doing so from a point of pro- or anti- nintendo or video games it is so silly to talk about whether 'video games are more expensive now' 'video games have always been this expensive' when what you're talking about is exclusively big-money titles that you're buying on release day at full price. that's not "video games". that's "video games made for current-gen hardware by activision/nintendo/EA/etc". its like saying "Movies nowadays all have Iron Man in them"
many of the best video games of all time are made by one weird trans girl and available for free Free on itch.io. many more are made by two autistic finnish guys and available on steam for like $10. the "video gamse used to be even more expensive adjusting for inflation" thing is only true about usamerican kids who were playing the new nintendos at the time, to me as a kid "video games" were like ~$8 from a guy with a cd burner or when i was slightly older like $20 for three from the second-hand ps2 games bin. like its like when people make sweeping statments on Movies and they exclusively mean the MCU. get some fuckin perspective on the medium god damn
oh my god i talk a lot of shit about getting back more of daleks master plan but its actually happening there's a clip on the article and everything. oh my god it's 1am. i have work in the morning.
Peter Purves, who played the Doctor's assistant Steven Taylor, was invited to the Phoenix Cinema in Leicester on Wednesday under false pretenses to view the two episodes, and he said: "My flabber has never been so gasted."
Daleks Master Plan was broadcast once in the UK in 1965/66 and then never again bcos it was never picked up for overseas broadcast. there were 2 copies known to exist total. this is genuinely incredible news I knew we were getting a missing episode return but I can't believe it's actually DMP oh my god
every accusation is a confession with fascists but it can be hard to remember that also applies to their obsession with 'the enemy is both strong and weak.'
Reminder: don't believe this yourself. Fascists don't just feel insecure, they are insecure because they control structures people rely on to strip the security away from anyone on a whim. But this a phenomenon that is contained inside the fascists, you don't want it inside the you. Got it?
The result of this self-destructive social engineering is that fascists self-fulfilling-prophecy themselves: they fucking suck but also need to be taken seriously anyway. This is harder than it sounds. For example, a fascist administration is attempting to assemble personal information of its citizens:
Is this act of surveillance immense and terrifying? Yes.
Is their data supremacy largely bullshit? Also Yes.
Should regular people stop assuming even 'secure' solutions are private? Yes.
Will this kind of surveillance largely be used to justify 'whatever' after the fact, rather than identify dissenters proactively? Yes, Again.
Can this government act with minimal oversight to threaten its own population? Yes!
Does it know its hands from its ass? No.
Can you be identified personally for your speech by the government? Yes.
Will they need to subpoena documents from another country to do it because their own record keeping is in the shitter and you can use a VPN? It's probable if your speech is 'important' enough.
Are there trigger happy shitheads on the ground? Unfortunately Yes.
Are they so incompetent they can't handle their basic responsibilities? Yes and that's not better because that makes them more ashamed, more trigger-happy.
It's really hard to coherently communicate how capable/shitty fascists are because of this. "You need to worry about opsec, but also you don't need to worry about hypercompetent surveillance," but, like, what does that mean? Be worried a medium amount?
An empowered idiot is always one of the hardest things to impede because while they can do anything it's hard to describe when or how or how effectively they actually will until it's too late and they've already done irreversible harm.
Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
✓ Live Streaming✓ Interactive Chat✓ Private Shows✓ HD Quality
Anya is LIVE right now
FREE
Free to watch • No registration required • HD streaming
I don’t know about anyone else but these sorts of headlines frequently make me highly suspicious so I did some follow up research after reading the article. I am happy to say that Green Sea Turtles no longer being counted as endangered is not due to the IUCN or any government shifting the goalpost for what can be considered an endangered species. The only update is, as the article states, an update in surveyed population numbers. Notably the article itself states that the current Green Sea Turtle numbers are not currently at what they historically were when the species was thriving, but this change in status over a relatively short period should absolutely be considered a significant win for conservation efforts and indicates a larger more grounded hope in our ocean ecosystems ability to heal.
Fighting for a better world is worth it and it is not too late.
I technically exist @mumble-bun - Tumblr Blog | Tumlook
