Hidden Butterfly

Will | she/her | 30-something | czech republic | science nerd | fandom nerd | queer. Enjoy my fandom of the week.

[chronicallyunemployed]

in tongue tied, the cat is obviously fabulous and eye catching, but i love watching rimmer and lister cause the actors did not wanna do that lol. Chris barrie is just aggressively looking ahead while craig charles keeps glancing at the camera like he's trying to mentally persuade them to stop this nonsense

[hidden-but]

I needed to rewatch it, and omg lol yeah

[hidden-but]

by mě zajímalo, kde se dá v Ově nejlíp trsat, trsat ...

[becausegoodheroesdeservekidneys]

Bank of England are letting you vote for what animals you want on their new bank notes: https://www.bankofengland.co.uk/banknotes/help-us-design-our-next-series-of-banknotes

Pine martens are an option!

PINE MARTENS??!?

[becausegoodheroesdeservekidneys]

Oh my god, you can choose up to two from each category:

HOW CAN I NARROW THIS DOWN

[v-spicata]

I chose the fox as one of mine, it's an obvious choice but it'd be nice to celebrate an animal so commonly denigrated. Not that old 'foul mart' has had much of a fun time of it historically either.

Some interesting options here in general, they've not just gone with the obvious animals.

[becausegoodheroesdeservekidneys]

I ended up not choosing the fox, purely because I actually reckon it's going to romp home - for all the controversy, it's the most common wild mammal people see in urban centres, and it's charismatic

I went pine marten, as I've been involved in helping their reintroduction to Wales, and then I wrestled with myself for an Age before finally going hedgehog.

Birds: puffins were the easiest choice. The UK - and Pembrokeshire Coast National Park in west Wales specifically - has a significant portion of the global breeding population of puffins, thanks to Skomer and Grassholm islands. In a country with the biodiversity depletion we have (bottom 10% of countries globally for biodiversity), the islands of Pembrokeshire are almost obscene in how high their biodiversity is, and it's for breeding specifically. We can be justly proud of those. Plus, puffins are fun clowns.

And then I agonised about the others until I finally went for the Great Spotted Woodpecker, a bird I do periodically see and get excited about every time

The Lumped-Together-Others: the bumblebee, you have to. I adore bumblebees.

And then I went for the marsh fritillary, because it's super endangered and I'm an environmentalist with a specialism in habitat management and ecology, and therefore spend a non-trivial amount of my time explaining how to manage for the little assholes.

But MY GOD it took me a while

[hidden-but]

I don't know some of them, so Let's Learn News Animals With Bank Of England, while I'm wikipediaing them myself:

bottlenose dolphin (image by NASA, lol), brown hare (european), European hedgehog

grey seal (male, pic actually taken in England), pine marten (european), red fox (apparently there are two different subspecies in Europe with some fuzzy range difference, but apparently this one is present in Great Britain)

Atlantic puffin (pictured in Wales) , (wastern) barn owl, common kingfisher (i always forget how beautiful they are in the real life)

Eurasian curlew (my brain keeps reading the name as curfew), great spotted woodpecker (female, because i wanted a picture on a treeside), white-tailed eagle

Atlantic salmon (apparently also known as landlocked salmon, gods, i found so many rabbit holes of animal life to follow today) , basking shark, buff-tailed bumblebee (apparently this one is used for industrial tomato pollination in greenhouses as well)

common frog (european), emperor dragonfly (the original x-wing), marsh fritillary butterfly (has a wide variability of the patterning and color scheme of the wings, which even might land them in a "species complex" category, because humans like boxes and biology doesn't care)

aaand that's all.

It was great to have an excuse to have a look at some "common" wild animals in a greater detail and learn some new things.

[pwhl-mybeloved]

oh ok

[iconuk01]

Nice try Zamboni Rangers, but we know that...

BLACK was the Mammoth

BLUE was the Triceratops

[thefingerfuckingfemalefury]

A shameful mistake!

[thestargayzingheroine]

I always liked that the Triceratops Zord had tank treads instead of legs.

[thefingerfuckingfemalefury]

Lets them nyoom real fast! :D

[hidden-but]

one thing i didn't expect when I agreed to have an online talk at a summer school in Kiyv, is that I will be reminded to write and send them my fucking abstract every time I hear international news.

[hidden-but]

and yeah, it always feels like a miracle to meet this particular colleague in person at a conference, because he not only has to travel by train to Poland before he can fly anywhere, but also he has to get a special permit to leave the country first.

[hidden-but]

also, fuck Russia, in case that wasn't obvious

[hidden-but]

one thing i didn't expect when I agreed to have an online talk at a summer school in Kiyv, is that I will be reminded to write and send them my fucking abstract every time I hear international news.

[hidden-but]

and yeah, it always feels like a miracle to meet this particular colleague in person at a conference, because he not only has to travel by train to Poland before he can fly anywhere, but also he has to get a special permit to leave the country first.

[hidden-but]

one thing i didn't expect when I agreed to have an online talk at a summer school in Kiyv, is that I will be reminded to write and send them my fucking abstract every time I hear international news.

[justalurkr]

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.

[m3ll1sa]

I remember watching an interview with Timnit and Emily about their paper back then. They seemed to be in the minority of experts that spoke about how the narrative for stochastic parrots is dictated by the people supposedly concerned about everyone’s future. But this future is millions of years away and what actually happens to real humans doesn’t matter to them in the slightest. Timnit even suggested a neologism for bastards like that:

[banchie]

i've been phasing the phrase 'google it' out of my vocabulary and going back to 'look it up'. fuck you youve lost your generic trademark privileges

[hidden-but]

oh, look, people complained about the copilot key being such a bloody fucking nuisance that is impossible to feasibly remap, that Microsoft actually added this option to Settings

Like, the options are still Copilot, Search, or Copilot (a different version), but it's still better than nothing.

[hidden-but]

I love when the author's voice in nonfiction books comes through so strongly that you can make an educated guess if you'd like to have a chat with them at a conference coffee break, or rather absolutely fucking not

[hidden-but]

Shoutout to a 2002 Attenborough documentary about mammals for straight up showing whale dicks in the middle of copulation.

[hidden-but]

Zrovna běží na Prima ZOOM Válka za polárním kruhem, což je navzdory generickému názvu celkem zajímavý finský dokument, který ukazuje skandinávský pohled na Studenou válku - a na rozdíl od amerických dokumentů v tom není těch 400 tun patosu jako obvykle.

[twofingerswhiskey]

"oh food now has so much added to it, past food was so pure and untainted" victorians used to cut bread with chalk and aluminum powder. romans put lead in the wine, which was made from dirty feet mushing unwashed grapes covered in horse shit and road dust. i think our species will survive a few additives in food. our food systems have never been cleaner and safer. it has room for improvement, but we're not putting fucking plaster of paris in the milk

[hidden-but]

Mindfulness cvičení jakože sedím u otevřeného okna, poslouchám, co se děje venku a čekám až přijede Rohlík a doveze 20 kilo rašeliny pro mou drahou rodičku, která se s nima nechtěla tahat autobusem.

[hidden-but]

National Geographic is now Disney?

wtf have I missed ?

Also fuck you and your 1007 partners.

[hidden-but]

yesterday i was reading a headline about a movie featuring AI Pacino ... it took me second to realise that mister Pacino is abbreviated Alfred, not artificial intelligence.