I can't make up my mind

꧁ 𝕢𝕦𝕖𝕖𝕣 │ 𝕙𝕖/𝕙𝕚𝕞 │ 𝕞𝕠𝕣𝕖 ꧂ I block transphobes, zionists, fascists and bots on sight.

[suodelic]

sex with a guy who wants to hunt you for sport for 6-12 hours immediately beforehand

[suodelic]

sex with a guy who gave you a big head start and mocks you about it after he’s caught you anyway

[suodelic]

sex with a guy who wants to watch your brain kick down a few notches when he pins you down like you’re a prey animal that knows it’s done running

[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.

[blackjame]

[neuroqueerings]

So I've now been on tumblr for 5 months. First there were The Ides, which I think I took in my stride. But then there came Neil banging out the tunes, which I, in my innocence, thought meant you all shared my passion for Neil Sedaka. Alas no. And now you're asking me to eat ramen, simply because it's going to be May? Where am I? What is this place?

[depsidase]

[afusionoffandoms]

One dead, 16 injured at Berlin Pride event after vehicle crashes into crowd, police say

Police say an intensive manhunt is under way after the incident in Berlin's city centre.

https://www.reuters.com/world/berlin-pride-parade-called-off-after-car-hits-crowds-police-say-2026-07-25/

Hundreds of thousands of people had gathered peacefully in the German capital for the annual Christopher Street Day celebrations, one of Europe's largest LGBTQ ​events, before the apparent attack brought it to an abrupt end.

"There are many injured, among whom multiple people have ​life-threatening injuries," police said on X, adding that an urgent manhunt was underway for ⁠one or more possible perpetrators.

[niknak79]

[odysseusyes]

even with those four numbers there are countless possible combinations good luck with figuring out which one is the right one you punk

[eatsleepcrap]

*straightens calculator*

It’s pretty likely that it’s a four digit number, and as there are four digits chosen there, that means that there cannot be any repetition. This mean that there are:

n!/(n-4)! possible orders. As ‘n’ is 4 (number of digits available). 4!/0! which becomes 4x3x2x1/1 which simplifies to 24. That means that there are 24 possible combinations of codes. This would take you about two or three minutes to input all possible codes.

[crashingsunlight]

Unless an alarm goes off if you don’t get it right in 3 tries

[eatsleepcrap]

*straightens calculator again*

Kick the fucking door in

[the-brave-little-blogger]

well ‘technically’ the code is most likley 1970. statistically, a majority of people, when told to choose a 4 digit code will choose their birth year. and this key pad is obviously a few years old to put it nicely, thats most likley it. 

[everyonesfavoriteging-blog]

some sherlock holmes shit just went down over here

[the-neon-pineapple]

No, no, no. Don’t base your deductions of psychology. Let’s talk chemistry. When you first press a button, there’s more of the natural oils on your skin, and therefore it wears down the numbers on the keys faster. Obviously 0 is the first one, then. Try 0791 first.

Sherlock out.

[perks-of-being-chinese]

woah.

[panicrobot]

it got better

[twistedthicket1-deactivated2022]

and this is why the sherlock fandom could either rule the world or end it….

[ofthefantastic]

Close, but not quite, I think. People will almost always choose a number they can remember. What’s memorable about 0791? Try 0719 - a birthday, 19th of July. That is more likely.

[venatrix26]

Those deductions are great and all, but unnecessary.

The light is green.

The door is already open.

[party-gilmore]

And that’s why we have a John Watson.

[lotrlocked]

This is “top 10 favorite posts” level.

[moriartyfortheevening]

Omg, it’s actually on my dash! This post is like a fossil!

[oriocookie]

@hellsite-hall-of-fame

[scramratz]

For someone who's been rejected my entire life, you'd think I'd handle it better

[fairycosmos]

fucked that you can’t fix other people especially when you really care about them. Oh so im just supposed to be there for you while you suffer. like a useless cunt gargoyle

[creatingblackcharacters]

Now, a question for my Black followers ONLY:

The rest of you can (and really ought to!) reblog, but it's not your turn to talk. Just listen and reflect!

What DOES a good apology for antiblackness look like, to you?

Not a shitty twitter apology that dodges accountability. Not a genuine apology made to white fans to soothe their embarrassment and egos and maintain white solidarity.

What does an ACTUAL apology for antiblackness consist of, to you?

What actions do you need to see from people who claim to be better, to deserve your support despite a history of being unsafe?

Why do you think nonblack people feel they're entitled to grace, and what would actually EARN them your grace? What makes you not want to offer it at all?

I think there's a disconnect between what the group being harmed needs, and what people think the group being harmed should settle for. We've been forced to know the latter. How do Y'ALL feel?

[ihazyourkitty]

Friendly reminder:

No, you don't want that otter. No, that is not a pet. No, that up close encounter where you can pay to cuddle with an otter is probably not ethical. No, I'm pretty sure that channel with a guy and his pet otters is not a legitimate "animal rescue", stop recommending it to me.

Yes, the otter is cute. Please admire them either from a distance, or behind a barrier at an accredited facility. Thank you.

.....no I'm not going insane seeing how flooded social media is with unethical otter content.

[loubatas]

The recent popularity of otter cafés in Japan caused a rise in animal trafficking & poaching of these otter species in Asia. Asia isn't an exception: the same will happen anywhere. The more popular an exotic wild species is as a "pet", the more it encourages poaching & trafficking. Even if you can give your exotic animal the best living conditions (which is always more complicated than with a domestic species), this factor will always remain

If you want a pet, you want a domesticated species. If you want a domesticated mustelid (the family otters belongs to), we have one: it's called a "ferret"

Yes, they do require specific care you need to learn about (like any animal, wild or domestic), no they don't look exactly like otters (it's the closest you'll get in domestic animals though) But they're a lot more adapted to live in human houses, and they're not at risk of being poached in the wild. Mostly because like dogs or domestic cats, they don't exist in the wild

[sevibun]

[persephoneme]

you can't say "hey has anyone noticed that M/M fic outnumbers F/F like 100:1” or “it feels racist that only 3/202 characters on the ao3 top 100 ships list are Black and two of them are Alastor HazbinHotel” bc some ppl will start going like “oh so you think we should FORCE people to write about things they DON’T CARE ABOUT for WOKE????” and you’ll be like “no, i’m pointing out that the conditions that created this disparity are informed by racism & misogyny” and ppl will say “it’s not BIGOTED to only care about WHITE MEN” and then the gargoyle king appears

[infectiouspiss]

trans people don't need to hear your justifications for still enjoying harry potter in 2026 stop looking for forgiveness where you're not going to get it

[infectiouspiss]

if i walk into your house and see harry potter stuff, i'm just gonna assume you're a transphobe. no amount of "i got it from a charity shop" can save you from the fact that you value those books enough to display them in your home.

you can't just put the work of someone who's currently trying to eradicate trans people on display without being questioned about it, especially if you claim to care about us.

same goes for your social media, if you're willing to keep putting jkr's work on a podium and proudly talking about it, then it's simply logical to believe that you don't care about other people. it's selfish if i'm being honest, to think that your comfort matters more than the lives of millions.

[infectiouspiss]

and to the trans people saying that they're "reclaiming harry potter for the trans community"

no you're not, you can't do that while jkr and her neo nazi friends are currently trying to kill you. you need to let go or else your own community won't be happy to help you when our collective rights are stripped further away.

[infectiouspiss]

trans children in the uk are killing themselves because of the things jkr has done.

children are dead. children who should be happy and free to be themselves are fucking dead because jkr has used her immense wealth and social power to change laws specifically to harm trans people. if you really think that your fucking "comfort characters" are more important than the lives of children then you should say that with your full chest. go on. tell the entire world that you don't give a fuck about dead kids. that you don't have an ounce of empathy in your disgusting shrivelled heart, because you still have your shitty fucking books and your garbage movies. your comfort matters more than anything, doesn't it? maybe jkr will tell you you're one of the good ones, wouldn't that be nice?

[crabussy]

hi trans kiwis and friends. if you haven't seen the news, they're trying to pass a frankly insidious bill in aotearoa to define the terms 'man' and 'woman' based on biological sex. this unsurprisingly reflects a lot of similar cruel efforts happening overseas at the moment. IT HASN'T PASSED YET, but I figured I should speak up about it because this is happening as we speak.

(screenshot from the linked RNZ article)

it seems very fucking bleak!!!! please don't lose hope! it hasn't passed yet and a lot of the shoddy bills suggested by the coalition have been shot down already. it's still worth knowing about. you don't have to share this post if you don't want to. I just know that a lot of my followers are kiwi. if there are any updates as to what we can do to push back against this, I'll make a relevant addition. kia kaha, okay? love you all.

[crabussy]

UPDATE: you can now very quickly and very easily submit a comment on the bill! these will be counted and considered in parliament.

it took me less than three minutes, and that's only because I wrote a whole paragraphs on why this is a stupid idea. all you really need to put is "don't pass this bill" if you want. it'll only take 30 seconds.

if you're able to and you want to, please consider reblogging! the more reach this gets, the better. thank you all so much [:

[lnfini-deactivated20230930]

disclaimer: I am east asian. if anyone who is not white sees anything wrong with my phrasing, inaccuracies, or insensitivity, or something I missed, please feel free to add on. I’m just one person with one perspective; none of what I say should be taken as The Singular way to draw an Asian character. if you havent done so already, please take the effort to expand your view of Asian culture outside this one tutorial.

if a white person reblogs this and adds something stupid I’m going to bite and kick you like a wild animal

[6480n]

I LOVE this post and wanted to add some additional info, cause I see a lot of people who assume that drawing asian hair is the same as drawing white hair. This is not the case! There’s more to it than just the color.

Image description for the original post and my addition are underneath the read more at the end.

Thank you for reading! Once again, image description for my images and OP’s images are under the cut.

Just like OP, I’m only one single person, so if anyone wants to correct me or add something I missed, then go for it. And further disclaimer, there are exceptions to everything I’ve said in this post, and it only applies to East and Southeast asian people who are either not mixed or white-mixed, as those are the hair types I’m most familiar with.

Edit: AUGH I FORGOT TO mention this but @ everyone in the comments talking about blue eyes on biracial asians, here’s another guide I made that goes over color inheritance for biracial people: https://6480n.tumblr.com/post/633074808569069568/making-this-guide-because-i-see-this-question-time

Keep reading

[sometiktoksarevalid]