there are places in the world today that are experiencing 40°C for the first time in recorded history. of course there's no way to know whether chucking billionaires into volcanos will appease the sun god but i feel we're doing the scientific method a disservice if we don't at least try
The fact that I've gotta be out here cold calling pharmacies, not even chains but each individual location, to figure out if they have the stock to fill my prescription is insane. like a doctor sent you a prescription that said give this person medication and somehow it becomes my personal grail quest to hunt the fucker in the wild
"Racialised" is much better than PoC but I've been leaning a lot on the concept of racial markedness. Because that allows us to make statements like "the name Jamal is racially marked in USA". Rather than saying something like "Jamal is a PoC name", a nonsense statement, saying it's racially marked in USA allows us to contrast with societies like Albania or the Arab countries where the name Jamal is ordinary, thus unmarked.
It's a concept I've kind of imported from linguistic analysis; saying a speech pattern is more or less marked does not really allow us to avoid the subject of who's doing the marking. A statement like "womens' speech is more marked in Lakota" necessitates that we understand that it's the Lakota who are marking womens' speech. A foreigner can't tell the difference and probably doesn't understand why it would thus be weird to see a man using speech patterns associated with women, in the same way an Albanian wouldn't understand why USA people would think Jamal is a Black name.
You! You get it. In my view, if someone is saying "racialised" or "racially marked" without acknowledgement of context, they are doing it in a way that is gramatically incorrect.
[Start ID: tumblr tags that read "#ohhh fuck that's a really good way of lookking at it #it forces the relative nature of it all to the forefront #it *makes* the listener pay attention to the fact that their context isn't THE context #and removes the assumption of Default]
“If I have one message to give to the secular American people, it’s that the world is not divided into countries. The world is not divided between East and West. You are American, I am Iranian, we don’t know each other, but we talk together and we understand each other perfectly. The difference between you and your government is much bigger than the difference between you and me. And the difference between me and my government is much bigger than the difference between me and you. And our governments are very much the same.”
1. The court holds Google responsible for statements made by its AI, considering them Google's statements (search engines have limited liability for results in their engine as they're the words of other sites/companies/people), meaning when their AI lies/hallucinates they're liable for the defamation/harm resulting from those statements.
2. Google's defense that customers are generally aware of the lack of reliability and are responsible for fact checking was dismissed. As the court pointed out, that would "significantly diminish" AI Search's stated purpose and it can't be distinguished from Google's business practices/statements as a search tool.
3. Studies have found about 91% of Google's everyday AI responses are accurate, leaving millions of searches per HOUR with potential liability for falsehoods. 56% of correct responses weren't supported by the sources the AI listed. Both of which mean Google is now liable for a LOT more AI "errors."
4. Google was held liable for 80% of court costs in this case and this precedent is expected to reverberate around the world. This is a massive shift from the 3rd-party search provider role Google has previously played and it comes right as they've tied ALL searches to their AI search.
Every day I handle more money than I will ever make. Every day.
At the start of my employment, my boss showed me videos of people stealing, and we both had a chuckle about it. How silly they were! There was a camera overhead, and it’s not to watch the shoppers. See, we can’t actually stop shoplifters. They get away with it maybe nine out of ten times. But we, who are watched and tallied and witnessed? We are always caught.
At first it was hard to hold one hundred dollars bills. An amount I had never seen before. An amount that didn’t exist in my household. It’s normal now. Here is something that is not for me.
“What the hell, I’ll take another,” says the man, pondering our 200 dollar watches. What the hell. Total comes to 580 and not even a flinch in his face. I have been working for 11 hours today and made only 110 dollars. It will go to my rent. Today I work for free, it feels. When I get my check, I will have 35 dollars left for food and saving.
The six hundreds he hands me go into the cash register. For a moment, I imagine having money. Then I put it away, counting out his change.
I know for a fact we sell our products for double what they are worth. That I could be making commission. That they could hand me those 580 dollars and change my life and not even mark the difference in their checkbooks. He’s not the only sale they make today, but I am the reason they made it. He’s not the only one spending 600 dollars, but if I hadn’t spent two hours with him telling me about his life, he wouldn’t have spent any. I go home. I don’t own a watch.
I have watched and rewatched a video on how to make salmon four ways. My shopping list is always the same. Pasta. Rice. Tuna. If I can afford butter it was a good week. I dream of the world I will never walk in, where I can throw the best fish fillet in the cart with a shrug. I hold hundreds in my hand and look up at the camera. I put them under the cash drawer.
I go to work. I scrap together my savings. I eat my bowl of rice slowly. My manager takes a paid week off from work just for his birthday. He owns a yacht.
i wrote this while i was working at orlando’s walt disney world parks.
i was part of their college program. i moved to the state for it. they legally owned the building i was living in and still charged me rent. i ostensibly was being charged to work for them. it was a 2 bedroom apartment and they placed 6 adult women in it in forced triples.
as many as one in ten disney employees have experienced homelessness while working for the company. despite huge efforts to unionize, strike, or otherwise demand fair treatment; disney has refused to increase employee quality of life.
disney admits publicly that a good portion of their success is because the employees (“cast members”) are dedicated, passionate, and selfless. this is never reflected in pay. even “face” characters (ie those that are princesses etc) make barely above a minimum wage.
at the time that i worked there, i made $8.50 an hour. at one point i was asked to create a human shield around a bag because a bomb dog had alerted to it. for eight fucking dollars an hour.
i now work a very cushy office job. i have bought the salmon and cooked it all four ways.
i go to the store. i am nice to the person behind the counter. she looks up at the camera while she counts out my change. there is nothing fundamentally different about her and i.
good art is when something looks like real life, the more real it looks the more better the art. abstracted figures give my trad children nightmares, one time they were exposed to cubism and couldn't go outside for a week
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.
The concept that married people live longer is interesting. I'm sure there is some merit to the idea that if you're married there is someone there to nag you about going to the doctor, but I think much larger factors are having the finances of dual incomes and access to an immediate support person.
Surgeries require having a designated person to look after you. Many injuries require driving to somewhere like an emergency room which can be hard to do if you are the one injured. If you're home with the flu, it's hard to tell when it's bad enough to go to the hospital without another person checking on you. And if you pass out it requires another person to find you like that to get medical aid.
You can prop it up as the benefits of marriage, but I think there's a much deeper discussion to be had about how we've built society around marriage as an inevitable conclusion and neglected to build support systems that function outside of romantic pairings.