slow and steady

At the 1985 Labour party meet, the motion to support LGBT rights as a party was passed due to a block vote from mining unions.

literally just saying 'he spoke with a Welsh accent'; sure, it's a bit blunt, but it gets the job done in a pinch. "He's completely drunk," he said, his southern drawl lingering on the final syllable as if to highlight the extent of the offence. Y'know, something of that ilk, but not as shit.

learning the specific vocabulary and syntax that someone with that accent might use. Sticking with the Welsh theme, because it's objectively the best accent*, there's a bunch of things that differentiate a colloquial South Walean accent, outside of our famed tendency to elongate a vowel to the point of death. The way we use prepositions (where to by is he?), the vocabulary borrowed from Welsh - saying that someone daft is twp, or something small is dwty - can easily signpost our speech as being from that specific area, without needing to type something like "'e's absolutely 'angin', man, pissed as a faaht 'e is!" Something less jarring, such as "He's absolutely hanging, he is." is just as clear. A character who says "Do you want a cuppa?" is coded or located very differently to one who says "You'll have a cup of tea, so you will."

ditto if there are specific ways that someone from a certain area might refer to a well-known concept. Regional words for mother and father, for example, or words that are class-specific; your character who calls his parents 'mater and pater' is likely inhabiting a different socioeconomic strata than your character who calls them 'mam and dad'. See if there's a colloquial way of saying 'yes' and 'no'; a lot can be signposted if your character says 'nah' rather than 'no', or 'aye' rather than 'yes'. A character saying 'couch' is inherently coded differently to one who says 'sofa'.

The reasons that writing accents phonetically is Generally Ill-Advised, In My Opinion are as follows:

quite simply, you're probably not being as clear in conveying the sounds of the accent as you think you are. Taking JK Rowling's work as the best possible example of this, her attempts at writing a Cockney accent phonetically come across like someone is chewing a mouthful of cheese curds and struggling to contain them. There's no consistency, no proper understanding of how to transcribe syllables into writing in a way that coherently conveys the accent she's trying to portray. I mean this so seriously, but what the flying fuck is: 'Well, 'e 'ad these 'ead pains and 'e was def'nitley nervous. Depressed maybe.' It's a crime, is what it is.

it's just plain hard to read. Trying to wade through sentences full of apostrophes and elision, parsing what's actually being said, gets tiresome. It asks the reader to do work that you're actively making harder for them. And that's not always a bad thing! Making readers Put Some Fucking Effort In can be very fruitful! But do you really want them to be struggling to understand every single thing that your Character B is saying for 350 pages?

which leads me onto the last point, and the most important in my mind: writing out accents like this always, always affects accents that are already in some way Othered. They're either racialised or working class, or associated with certain local regions that have negative stereotypes - think the deep South of the US, or the Welsh Valleys. They're never the 'default'. And this raises thorny questions about what the default is, what the standardised accent is, the accents that do and do not merit differentiation from the norm. You're relegating Character B to being hard to read because he's from, idk, Sunderland. You've decided that he isn't speaking 'properly', and therefore the reader needs to understand that other people think he's speaking weirdly. That, to me, is the principle issue. Because returning to JK Rowling (a sentence I hoped never to type), the only characters who speak like this in her work are working class, or they're from other countries. They're never from, you know, Surrey. Wonder why that is. And it's easy to be glib about it, but I do think it reifies class and regional boundaries in a way that's ultimately harmful.

This isn't to say that there's never a place for eye dialect in writing - Trainspotting (edit to respond to some legitimate comments in the reblogs: I bring up Trainspotting because it's written in Scots and Scottish English, not just Scots, but I agree that this isn't the best example as the Scots portions are not part of this conversation in the same way; consider Their Eyes Were Watching God by Zora Neale Hurston as a better example, and apologies for the confusion!) wouldn't be what it is without it, and there's definitely a different conversation to be had when it's your own accent and you're making a deliberate point about identity by differentiating through eye dialect - but I think that the blanket assumption of 'oh shit, my character is from Ireland, I'd better type that out phonetically!' can actually be both damaging to your writing and to your character representation, and I think that instead doing the work to really understand the vocabulary, speech patterns and unique aspects of a language or dialect always makes a work feel more authentic and lived-in.

To wit, less of this shite:

There’s mony a slip, an’ I’m no losin’ sight o’ any o’ my suspectit pairsons, juist yet awhile. (One of the Lord Peter Wimsey novels by the very English Dorothy L. Sayers, if you were wondering, and yes, that's supposed to be a Scottish accent; I'd not be bringing it up if it were a Scottish author writing in Scots)

and more of this:

"Are we straight so?"

"Aye, we're straight," said Jim.

"Straight as a rush, so we are." (Jamie O'Neill, Irish, from At Swim, Two Boys)

june days of productivity, day 3

strawberries at work & strawberry delivery. certified clever guy. sunshine during lunch break.

activities of today:

started the redo of the time reality map

small tasks & communication for recruitment committee

cat chores

unloading the dishwasher

prepped my sauce for dinner

work

the keychains we get at work make me feel very cool and dyke-y >:)

plan for the evening: dinner, then book in the reading nook (& no phone!)

Darren O'Connor

"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."

+

Rebecca Solnit 

[Extreme breach of scientific ethics]

[Violent abuse of power]

[Method that actually doesn't obtain any information]

[Controlled double-blind studies of phenomena where that is literally impossible]

[Seeking empirical proof that a word has the meaning that it's defined as]

[Study that would have a dropoff rate of 100%]

june days of productivity, day 2

snacking strawberries at work. more solo-trips to the 2nd floor. 25min of sun basking.

activities today:

laundry bedsheets

groceries & cat food

cat chores

endurance leg workout 30min

work (late shift)

my bike has a flat tire again and it happened 2km before arriving at work and now i can't bike to work tmr :( the ubahn connection is okay, but especially biking home after 8 hours inside really helps so much.. fresh air and wind beloved

june days of productivity, day 1

welxoming smiles. lunch in chatty company. doing things on my own. iced coffee in bed. sunshine bike ride.

things I did:

first day at the new job!

groceries

mobility

video games

29.05.2026 [😴]

26|06|2026