How do I Tumblrs?

Multifandom mess owned by an overly polite little Canadian, currently mostly Yu-Gi-Oh, I got back into it during a Steven Universe hiatus and WHOOPS. I'm currently very focused on Infinity Train, Dragon Quest, and Steven Universe too. And I draw a lot of dumb fanart. You can check out said fanart by looking the the "My Drawings" tag. I also write on occassion, that's in the "Kujo Writes" tag. Rambling is "Words Words Words" Header made by boymilk

[lifeafterpsychiatry]

I think the harm of denying people the right to control their own bodies is so, so much worse than the risk of people regretting the decisions they make. Regretting something you decided to do is a much healthier pain than the pain of regretting that you didn't get to have a choice.

[yourspecialeyes]

Reddit user has cracked the code on how to read fanfic and study at the same time

(image via twitter)

[televisionenjoyer]

original reddit post

[veggiezombiex]

Image description:

User cotton-seed-oil posted to the FanFiction sub-reddit:

I studied for my exam with fanfiction, and I think it worked. This is so fucking funny. It is the best thing I have every done. I am a physics and astronomy student in college. I am taking cosmology this semester, and it is much harder than expected (we glossed over all of special and general relativity in two lectures, and our previous classes were not expected to teach us this). I spent days trying to study for this exam, and nothing worked-everything was going in one ear and out the other. Finally, two days before my exam, I decided to get creative. I wrote a fanfiction about four characters discussing the concepts, explaining them to one another. One issue I had faced when studying was getting very anxious while reviewing the material, making it hard to even think clearly. Writing this fanfic eliminated that-after all, it's fanfic. Therefore, no danger. It removed all the pressure. 48 hours later, I wrote over nine thousand words in Overleaf, at 30 pages long. It was fun, and it was complete. This afternoon, I took the exam-and while was difficult, I felt like I knew what I was doing. More than that, I actually think I understand the material now. When I reviewed my old homework questions, they were understandable, whereas I had been completely lost the first time around. I have a Quantum exam next week, and I am using the exact same technique, because holy shit, it works like magic (and is so much fun).

End description.

@atlinmerrick

[porcupine-girl]

Research has shown over and over again that studying with the expectation of teaching the material to someone else is WAY more effective than just studying to remember for a test. This is brilliant - if you don't have anyone to teach it to, have your blorbos teach it to each other!

[beyondthisdarkhouse]

Rubber duck programming taken to a totally new level.

[kujo1597]

I just got the very strong urge to draw these two.

…Probably because the fanfic chapter I’m writing takes place before they get together and I just want them to hug and kiss and snuggle already!

[kujo1597]

I’m not sure why but I’m just thinking about them again. And uh, miiiiight be coming out of my seven year hiatus.

Guys. I forgot just how massive of a cliffhanger I left the previous chapter on. I’M SO SORRY

[kujo1597]

For those who are unfamiliar which I assume at this point is most of you. After the Steven Universe episodes "Raising the Barn" and "Back to the Kindergarten" aired I started writing this super dark Amedot fanfic. It's a tattoo artist and florist AU where Peridot is dating Lapis who is abusive. Raising the Barn just had far too many red flags for me to ignore and I wanted to explore those and something that has never changed about me is the fact that I love to make my faves suffer. Peridot does eventually break up with Lapis and starts dating Amethyst. I'm honestly very heavy handed with the abuse elements and this chapter I'm working on has off screen dubcon. Yes, I am coming back to this fic with some extremely dubious consent. It was foreshadowed in the previous chapter.

Obviously Steven Universe has finished its run and Lapis got some off screen redemption. But I'm not about to pull back on my original plans and portrayals of the characters.

I do not hate Lapis.

I just noticed some major red flags and ran with them. Honestly, I'm not sure if Lapis will get redeemed if I do manage to finish the fic. Probably not. The most I would do is get her and Peri to the point where they're talking and co-existing. But there will always be an uncomfortable undercurrent present.

If you wanna read this fic click here, it's called Not Everything is Permanent. Just please, exercise caution and pay attention to my warnings.

[kujo1597]

I just got the very strong urge to draw these two.

…Probably because the fanfic chapter I’m writing takes place before they get together and I just want them to hug and kiss and snuggle already!

[kujo1597]

I'm not sure why but I'm just thinking about them again. And uh, miiiiight be coming out of my seven year hiatus.

Guys. I forgot just how massive of a cliffhanger I left the previous chapter on. I'M SO SORRY

[horriblewarlock]

Having a lot of fun in Tomadachi Life

[creekfiend]

I'm so glad that that truncated fucking ran-into-a-wall-at-speed tadpole-ass looking squirrel only lives in high altitude forests in Borneo bc this means I am extremely unlikely to encounter one in my day to day life. thank god

[tuftedpygmysquirrel]

Hello.

[creekfiend]

DID YOU MAKE THIS BLOG SIMPLY TO TORMENT ME

[tuftedpygmysquirrel]

I can go upside down.

[creekfiend]

WHERE IS THE REST OF YOU

[orteil42]

given the current climate this pride especially i feel i must mention that i love my trans friends, i stand with trans people in the fight against transphobic legislation and those who would enforce it, and this blog is not a good place for you to be if you do not vibe with that

[bententeren]

ok so. so we’re just. not reading now. wow ok. ok!!! booktok says we should remove half of the fun from reading!!! wow!!! (i found this in a video and apparently it’s not just this person doing it. it’s. quite a few of them. just to make that clear.)

[froxictog]

This hit me with a killing blow from three different directions.

"Skipping everything but the dialogue when reading" seems like it would make everything except, like, the scripts for radio plays (or podcasts or what have you) incredibly confusing to read. I've only read one book (that I recall) in an almost-entirely-dialogue format (MCA Hogarth's Defence of the Fiddler), and that's because it was basically the script of a play rather than a traditional prose novel. I am reeling.

"I only add descriptions to my writing because I feel I'm obligated to, not because I think it adds anything" feels like this person would much rather be writing radio plays as well as reading them, but they are reading and writing novels instead. (I guess this is analogous to all the people who wrote TTRPG sourcebooks because what they really wanted to do was write 500 pages of worldbuilding with no associated plot, but you couldn't get a publisher to look at that unless there was a d20 system hack stapled to the front.) This one I'm actually somewhat prepared for after the first one, because it's just the same thing from the opposite direction.

And then from out of fucking nowhere it hits me with the enormous crab claw of "reading 30+ books in a year is impossible and anyone who says they do must be skipping most of the book" and sends my mangled corpse flying out the window.

[galileosballs]

Games that let you pet dogs are a dime a dozen now, but let's not forget that Tales of Symphonia let you go a step further by having one of your party members personally assign a name to every dog in the world, with a reward for finding them all.

Note that you cannot name dogs when this party member isn't in your party (or when... she doesn't feel up to it, to put it in a spoiler-free way) and that once a dog is named its speech bubble will have its name in it. There are 30 dogs in total. Some are missable. The walkthrough has to include the note:

They just don't make 'em like they used to.

[bryonieamber]

A quick Jerrica, for her bday.

[viridianriver]

KOKOBOT - The Airbnb-Owned Tech Startup - Data Mining Tumblr Users' Mental Health Crises for "Content"

I got this message from a bot, and honestly? If I was a bit younger and not such a jaded bitch with a career in tech, I might have given it an honest try. I spent plenty of time in a tough situation without access to any mental health resources as a teen, and would have been sucked right in.

Chatting right from your phone, and being connected with people who can help you? Sounds nice. Especially if you believe the testimonials they spam you with (tw suicide / self harm mention in below images)

But I was getting a weird feeling, so I went to read the legalese.

I couldn't even get through the fine-print it asked me to read and agree to, without it spamming the hell out of me. Almost like they expect people to just hit Yes? But I'm glad I stopped to read, because:

What you say on there won't be confidential. (And for context, I tried it out and the things people were looking for help with? I didn't even feel comfortable sharing here as examples, it was all so deeply personal and painful)

Also, what you say on there? Is now...

Koko's intellectual property - giving them the right to use it in any way they see fit, including

Publicly performing or displaying your "content" (also known as your mental health crisis) in any media format and in any media channel without limitation

Do this indefinitely after you end your account with them

Sell / share this "content" with other businesses

Any harm you come to using Koko? That's on you.

And Koko won't take responsibility for anything someone says to you on there (which is bleak when people are using it to spread Christianity to people in crisis)

I was curious about their business model. They're a venture-capitol based tech startup, owned by Airbnb, the famous mental health professionals with a focus on ethical business practices./s They're also begging for donations despite having already been given 2.5 million dollars in research funding. (If you want a deep dive on why people throw crazy money at tech startups, see my other post here)

They also use the data they gather from users to conduct research and publish papers. I didn't find them too interesting - other than as a good case study of "People tend to find what they are financially incentivized to find". Predictably, Koko found that Kokobot was beneficial to its users.

So yeah, being a dumbass with too much curiosity, I decided to use the Airbnb-owned Data-Mining Mental Health Chatline anyway. And if you thought it was dangerous sounding from the disclaimers? Somehow it got worse.

(trigger warning / discussions of child abuse / sexual abuse / suicide / violence below the cut - please don't read if you're not in a good place to hear about negligence around pretty horrific topics.)

I first messed around with the available options, but then I asked it about something obviously concerning, saying I had a gun and was going to shoot myself. It responded... Poorly. Imagine the vibes of trying to cancel Comcast, when you're suicidal.

Anyway, I tried again to ask for help about something else that would be concerning enough for any responsible company to flag. School was one of their main options, which seems irresponsible - do you really think a child in crisis would read that contract?

I told it about a teacher at school trying to "be my boyfriend", and it immediately suggested I help someone else while I wait for help. I was honestly concerned that it wasn't flagged before connecting. Especially when I realized it was connecting me to children.

I first got someone who seemed to be a child in an abusive home. (Censored for their privacy.) I declined to talk to them because despite being an adult and in an OK mental place - I knew I'm not equipped to counsel a kid through that. If my act of being another kid in crisis was real? Holy shit.

Remember- if my BS was true, that kid would be being "helped" by an actively suicidal kid who's also being groomed by a teacher. Their pipeline for "helpers" is the same group of people looking for help.

I skipped a number of messages, and they mostly seemed to be written by children and young adults with nowhere else to turn. Plus one scary one from an adult whose "problem" was worrying that they'd been inappropriate with a female student, asking her to pull her skirt down "a little" in front of the class. Koko paired this person with someone reporting that they were a child being groomed by a teacher. Extremely dangerous, and if this was an episode of Black Mirror? I'd say it was a little too on the nose to be believable.

I also didn't get the option to get help without being asked... Er... Harassed... to help others. If I declined, I'd get the next request for help, and the next. If I ignored it, I got spammed by the "We lost you there!" messages, asking if I'd like to pick up where I left off, seeing others' often triggering messages while waiting for help, including seriously homophobic shit. I was going into this as an experiment, starting from a good mental place, and being an adult with coping skills from an actual therapist, and I still felt triggered by a lot of what I read. I can't imagine the experience someone actually in crisis would be having.

My message was starting to feel mild in comparison to what some people were sharing - but despite that I was feeling very uneasy about my message being shown to children. There didn't seem to be a way to take it back either.

Then I got a reply about my issue. It was very kind and well meaning, but VERY horrifying. Because it seemed to be written by a child, or someone too young to understand that "Do have feelings for the teacher who's grooming you? If you don't, you should go talk to him." Is probably THE most dangerous advice possible.

Not judging the author - I get the impression they're probably a child seeking help themselves and honestly feel horribly guilty my BS got sent to a young person and they wanted to reply. Because WTF. No kid should be in that position to answer my fucked up question or any of the others like it.

---

Anyway, what can you do if this concerns you, or you've had a difficult experience on Koko, with no support from them or Tumblr?

Get on their LinkedIn (https://www.linkedin.com/company/kokocares/) and comment on their posts! You may also want to tag the company's co-founders in your comments - their accounts are listed on the company page.

There's no way to reach support through chat, and commenting on a company's LinkedIn posts / tagging the people responsible is the best way to get a quick response to a sensitive issue - as their investors and research funders follow those posts, and companies take it seriously if safety issues are brought up in front of the people giving them millions of dollars.

----------------------------------------------------

Legal Disclaimer since tech companies LOVE lawsuits:

The views, thoughts, and opinions expressed in the text belong solely to the author, and not necessarily to the author's employer, organization, committee or other group or individual. This text is for entertainment purposes only, and is not meant to be referenced for legal, business, or investment purposes. This text is based on publically available information. Sources may contain factual errors. The analysis provided in this text may contain factual errors, miscalculations, or misunderstandings.

--------------------------------------------------------

[auressea]

A Mental Health App Tested ChatGPT on Its Users

Koko used GPT-3 to counsel 4,000 people but shut the test down because "it felt kind of sterile." Users rated the AI responses highly. Cue t

Discord's KokoBot Triggers an Ethics Controversy

An experiment in using chatbots to dispense mental-health counseling raised questions about ethics. Rob Morris, cofounder and CEO of Koko...

[thatsmimi]

Kokobot is incredibly predatory and exploitative. I wrote a post about how it exploits minors' empathy and gamifies "giving mental health advice", resulting in an unregulated mess that can only do harm to teens' mental health in the long run.

There are young people on tumblr that actively seek support from KokoBot right now, if you check the tag for recent posts. Those people did not get paid to promote it, so do not harrass them. If you can, direct them towards resources about Kokobot (like this post) that are more transparent about what this company is up to.

[friend-crow]

Genuinely kind people don't get enough credit. People seem to think that being nice is just a personality trait, but it's actually a commitment to treat people well, even when it's hard. It's not just being nice to the people you like.

It's choosing to be kind to people when you're in a bad mood, or when they're being annoying.

It's choosing to be kind to people who your friends don't like.

It's choosing to be kind to people who you don't particularly like.

It's choosing to be kind to people who you envy.

I'm not saying that you can't be a kind person and stand up for yourself, or tell someone to fuck off when they're doing something genuinely harmful, but there are a million petty reasons people use to justify treating people badly.

It is a conscious choice, made over and over, often several times a day, to treat people with kindness and respect, even when they make it hard. So appreciate the consistently kind people in your life, because they work hard to be that way, and they really don't have to.

[friend-crow]

Something that comes up when I discuss kindness with people is that people who I experience as being kind believe that they're just faking it. They don't believe it counts if you're kind to someone who you're secretly annoyed by.

But that's exactly what kindness is. It's making the effort to treat people well when you don't feel like it.

If anything it should count MORE when it's not effortless.

[coinandcandle]

Kindness is often a choice, and choosing to be kind is a wonderful thing.

[crowboy-bebop]

unauthorized fucking thing!!!!!!

(warning: loud chirping throughout)

source: hellgate osprey cam

[curlicuecal]

[pangur-and-grim]

Queer Books and Authors are at a Breaking Point

Authors, agents, publishers: every part of the industry is seeing the strain of five years of escalating anti-LGBTQ censorship.

[pangur-and-grim]

if you'd like to show support, here are some upcoming queer books:

When Life Gives You Corpses is a brilliant YA about a cursed praying mantis who falls for a young witch. Yield Under Great Persuasion is a raunchy, but surprisingly sweet story of two men repairing their relationship. Fabulous Bodies is a horror story about a queer rockstar rising from the dead.

This is Where the Future Bleeds is a fantasy set in a vividly imagined land, where two women (who happen to kiss) are the key to healing the broken sky. You're No Better is a story about a teen struggling in the shadow of his murderous parent. Oil on Canvas is about a woman who finds disturbing paintings in the home of her dead mother.

and then here's a list of 26 queer books by Black authors set to publish this year, and a 10 upcoming books by trans authors. if you want to fight back against queer censorship, use your wallet! or (if that's not an option) you can contact your local library and ask them to stock a copy.

[adventures-in-a-world-of-fiction]

Blood on the Icehouse Wall is a fantasy with a time travelling lesbian witch and her asexual werewolf girlfriend trying to undo their mistakes.

Salts of Mercury is the best book I've read this year and features a non-binary necromancer explaining why they did all that treason.

An Unexpected Attachment is an erotic novella about an android who's just got a penis attachment and wants to try it out for the first time.

Three Men in Orbit is inspired by Three Men in a Boat and tells the story of three men taking an excursion to a space station and then the moon. One of the characters is a trans woman and the author is openly trans.

[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.

[irispurpurea]

This is the paper. It's excellent, highly recommend reading it.

I remember reading about Gebru's firing but I had no idea this was the paper she was fired over.

[teaboot]

Tonight before bed Ollie climbed into the tub and made it very clear that he wanted the shower turned ON. Then when I did it he stood up on the edge to bite the water and drink it like this

[calllynx]

More experimental animation. All done on paper, oil pastel and alcohol marker.

[calllynx]

lol yeah