-- Tom -- He/him -- I like making fractals & talking about math. On my blog, you will find a combination of math, math-related jokes and (sparsely) my original art.
hi just saw your post about a function you were doing your thesis on l? it was like from 2023,, oops. im actually intrigued if you wanna share about it !! it seems cool!
hi! thanks for your interest :) ive finished my thesis on calkin-wilf dynamics in december 2023, and youre one of many people whove asked me to learn more. thats why i did explain some stuff about my son in this post i made back in january of that year
of course that was just the beginning of my thesis period so over the months i did make some more posts explaining things about or adjecent to my thesis. for example here and here. theres also this post about continued fractions but i struggled to explain them in an understandable way so im not really proud of it.
also the goal of 2024 is to publish a paper about the same topic, so ill keep everyone posted on that!
for the record im not technially 100% anti-AI, in the sense that its a broad category of tech being lumped under one umbrella term so it feels over-zealous to say i hate all of it all the time forever. but i also think trying to discuss what it actually IS good for is difficult right now when i cant take one step without something trying to convince me to use chatgpt to summarize my life and speed up my hobbies and turn my friends into chatbots and optimize my life into oblivion. i am certain there is nuance to the topic but can we stop cramming the square peg into the round hole before you start trying to sell me on the legitimate benefits of the square peg. please.
Neural Nets have existed for decades and are genuinely useful. It's a form of AI that recognizes patterns, and can do stuff like identify cancer cells, tell whether an egg is fertilized or not, detect fraud, and optimize routes.
Those are Expert Systems, tuned to do exactly one thing. If you (say) ask a medical expert system a question about financial law, it's useless. The autopilot that flies a 787 has no idea how to drive a truck on the freeway. A Coulter Counter is excellent at identifying lymphocytes in a blood sample but can't predict the next card in a blackjack game.
And so on.
The problem with so-called generalized AI (AGI) is that we don't have that yet. It doesn't exist. It MIGHT some day, but AGI has been "10 years away" since the 1980s. The goals keep moving as we learn more about how people and machines process data.
But the current crop of AI techbros have been selling generative Large Language Model AI (LLM) as AGI because generative systems do a good job of faking it. There's no actual thought going on, merely the illusion of thought via predicting the next word in a sentence accurately.
If you let a human toddler listen to 800 hours of YouTube car influencer videos, that toddler might end up sounding like a car influencer. They'd parrot horsepower numbers and 0 to 60 times, mention EV range and MSRP numbers.
But they wouldn't understand any of it.
That's ChatGPT.
And yeah, it's worse than useless because it doesn't even know when it's lying or hallucinating. It just babbles convincingly until you stop it.
But for techbros to make money selling that as "AI"? It's the perfect scam, especially if you don't understand how it works.
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.
when you see a tumblr poll with this picture attached and you know it's time to lock the fuck in lest you get a bad grade in an impromptu absurdist pop quiz you didn't know you were about to take
Oh yea and while I'm at it: Chrome is ass. Like it exists to actively harvest your data and force unwanted AI features down your throat.
Chromium-based browsers (like Opera, Edge, but also the DuckDuckGo browser) are mid because Google still somewhat controls how these browsers function.
So take some time out of your day to install a browser that does not depend on Google, like Safari, Firefox or Librewolf!
so i havent touched any maths lately BUT last week i made a collage about mandatory vs voluntary learning w some artists and someone put on a blue paper with an aleph_0 symbol so naturally i put an aleph_1 symbol somewhere and made the infinite cardinalities into a metaphor for rigidity/freedom
You know how there's that genre of posts that's like "[screenshot of something horrid and dystopian happening with technology] hahaha, I sure hope nothing bad comes from defunding humanities while pouring boatloads of money into STEM"?
I need an inversion of that that's like "hahaha, I sure hope we don't run into any problems with masses of artists and writers deciding that it's fascist to understand the law and computer science."
That would have to become a real problem with real societal implications first. I know, your older colleagues refusing to learn how to use Moodle is a bother, but that's nothing compared to literally 1984
It seems like every humanities grad and creative on the internet has decided that the only way they will survive the future is by enshrining copyright laws more expansive than ever before and by being allowed to sue anybody for having looked at their work and had a single thought about it.
Permanent DMCA regime and an end to fair use seems pretty dystopian from where I'm sitting and the fact that the comments on this post are full of uwu smol bean creatives who are very proud to say things like "I don't care how AI works, it's evil" but can't imagine how their attempts to define model training as theft might have downstream problems that harm everyone BUT megacorporations is exactly the issue that I'm trying to articulate.
The problem is not that my elderly colleague doesn't know how to use a software program, the problem is that I'm watching liberal arts majors gleefully cheerlead attempts at implementing absolutely devastating legislation because they don't understand laws or computer science.
You know how there's that genre of posts that's like "[screenshot of something horrid and dystopian happening with technology] hahaha, I sure hope nothing bad comes from defunding humanities while pouring boatloads of money into STEM"?
I need an inversion of that that's like "hahaha, I sure hope we don't run into any problems with masses of artists and writers deciding that it's fascist to understand the law and computer science."
That would have to become a real problem with real societal implications first. I know, your older colleagues refusing to learn how to use Moodle is a bother, but that's nothing compared to literally 1984
It seems like every humanities grad and creative on the internet has decided that the only way they will survive the future is by enshrining copyright laws more expansive than ever before and by being allowed to sue anybody for having looked at their work and had a single thought about it.
Permanent DMCA regime and an end to fair use seems pretty dystopian from where I'm sitting and the fact that the comments on this post are full of uwu smol bean creatives who are very proud to say things like "I don't care how AI works, it's evil" but can't imagine how their attempts to define model training as theft might have downstream problems that harm everyone BUT megacorporations is exactly the issue that I'm trying to articulate.
The problem is not that my elderly colleague doesn't know how to use a software program, the problem is that I'm watching liberal arts majors gleefully cheerlead attempts at implementing absolutely devastating legislation because they don't understand laws or computer science.
so i havent touched any maths lately BUT last week i made a collage about mandatory vs voluntary learning w some artists and someone put on a blue paper with an aleph_0 symbol so naturally i put an aleph_1 symbol somewhere and made the infinite cardinalities into a metaphor for rigidity/freedom
