Cannot express enough how much I want to play in a setting where undeath is viewed as a neutral and even natural thing. Like, that forest was so vibrant and thriving that the deer skeleton just got up again and started munching on some leaves. Or a ghost/nature spirit/etc. needed a place to stay and so it picked up a decomposing critter for a body. They get up and go about their critter business for a while longer until it’s time to lie down in a sunbeam and finish decomposing. This is normal and fine. Actually, things not staying dead is a sign of a healthy, abundant ecosystem!
Just watched Adam Conover (of Adam Ruins Everything) make such a solid point that I think we should spread far and wide. Yes, having AI write your emails is lazy, sure, but people love being lazy. We need to really emphasize that sending AI emails (or using AI responses on social media, or publishing AI flyers, or or or) is rude.
It's rude. You're making someone take their time to read something you couldn't bother to write. You're telling them they were so unimportant you couldn't be bothered to actually take the time to say something yourself. And frankly, you're lying about it while you're at it.
I'm SO excited to finally share these pieces I created for D&D's Ravenloft: The Horrors Within.
This project was such a huge honor to work on. I remember as a kid looking through the players manuals & doing little studies of the artwork in there, so this feels a bit full-circle.
You can get a peek at higher-res versions of them on my insta!
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.
The Guild saved the world in the final season, and nobody even knows that they did it. The threat of the creatures beneath the seal was massive, but quiet. The only people who knew of the danger (as far as we know anyway) are the group themselves, the friends who fought with them at Necropolys on Sea, Liliana and Binbag
You would expect world-saving acts to be remembered, but with nobody knowing that anything was ever amiss, the story would probably fade with those who were involved. If anything, despite their good intentions, the guild would probably be remembered more for the dozens of escapades that ended in them being chased out of towns
Imagine decades down the line, the guild trying to prove what they did to show the world that their actions mattered. Corazon and his Penny Dreadfuls wanting to be remembered for his good deeds rather than for being his Father's disappointment. Egbert writing his changed morals and expanded worldview as the basis of his religion. Prudence searing her face into the minds of others as she torches them. Dob writing songs about his adventurers in the hopes that somebody will remember them. All of them wanting their sacrifices to matter.
Merilwen outliving all of the guild, discovering out how temporary everything she worked for truly is but still spending her extra centuries telling the people she meets about the times she spent travelling the world with her best friends.
But despite her best efforts, the universe still forgets. Because there will always be bigger threats to the world, and more heroic folks to save it.
Or so she thinks.
Millenia in the future, when magic has left the world and ghosts have come flooding in, a man would make his wealth on a whaling ship that he named in honour of the stories of his childhood, the dread pirate De Ballena
One day, an artist would be hired to write the next big opera. They would adapt their favourite childhood lullably. The tale had been corrupted by time, but the sentiment remained - one of hope and misadventure, starring a high-voiced page and steel wheels.
In time, magic would be lost to the world, but some power remained if you knew where to look for it. It still lingered in the pages of old religious tomes, holding enough power to banish a banshee despite how ridiculous the deity itself would seem to those who hadn't met them
When the ghost apocolypse first swept through G'eth, academics would pour over ancient texts looking for answers, and common folk were willing to turn to any source of authority for answers and salvation. I'd be willing to bet a decent number of people found comfort in the idea of being eaten last
Perhaps not as heroes but not quite as villains, perhaps not as themselves, but the guild leave a lasting impact in G'eth
Lego's Q3 2025 earnings announcement, October 2025
So Lego just posted another monster quarter and everyone's doing the usual "timeless appeal of analog play in the digital age" garbage and like, no, the actual story is that Lego is a privately-held Danish family company that spent the 2000s nearly going bankrupt and came out of it having figured something out that almost nobody in consumer products has figured out, which is that your core IP is the manufacturing tolerance.
Here's what I mean. A Lego brick made in 1958 still clicks perfectly onto a Lego brick made last week. That is not a marketing claim, it's a manufacturing fact, and it's enforced by tolerances measured in like two thousandths of a millimeter — the stud diameter variance on a standard 2x4 brick is famously smaller than most medical device manufacturers hit on parts going inside human bodies. Which sounds like trivia until you realize it's the entire business model: every brick ever made is compatible with every brick that will ever be made, which means the installed base isn't a depreciating asset, it's an appreciating one, because every new set expands what you can do with the bricks already in your kid's bin (and your bin, and your dad's bin in the attic).
Now compare this to basically every other toy category. Hot Wheels from 1972 don't interface with Hot Wheels from 2024 in any meaningful way — they're both little cars, sure, but the track systems have changed, the scales have drifted, the accessories are incompatible. Barbie has gone through probably a dozen body molds. American Girl dolls from the 90s have different proportions than the current ones. The entire video game industry is structured around planned incompatibility — your Switch games don't work on Switch 2, your Xbox 360 discs mostly don't work on Series X. Incompatibility is the business model, it's how you get people to rebuy.
Lego said no. Lego said the brick from 1958 will fit the brick from 2058. And this is insane, if you think about it, because it means they have voluntarily foreclosed on the single most powerful lever in consumer products, which is forcing obsolescence. Every company that sells a durable good spends enormous amounts of R&D figuring out how to make this year's product not work with last year's product without pissing the customer off too much. Apple is a master at this, Microsoft is slightly worse at it, car companies have built entire industries on it (proprietary charging connectors, OBD-II access, right-to-repair fights). Lego just... doesn't do it.
What they get in return — and this is the thing the "timeless analog charm" people miss — is that the brick becomes infrastructure. A Lego brick is not really a toy. It's a piece of durable manufacturing infrastructure that gets distributed into hundreds of millions of homes worldwide, and every new set is basically an expansion pack for an operating system that already has universal install. Which means the network effects are doing most of the work. When a grandparent buys a Lego set for a kid, they're not buying "a toy" in the sense that a Mattel product is a toy — they're depositing compatible substrate into an accumulating household stockpile, and every deposit raises the marginal utility of the next deposit.
This is also why the IP licensing deals (Star Wars, Harry Potter, the recent Nintendo stuff) work for them in a way they work for basically nobody else. When Hasbro does a Star Wars license, they're making Star Wars figures that sit on a shelf. When Lego does a Star Wars license, they're making bricks in Star Wars configurations, which means even if the kid loses interest in Star Wars in six months, the bricks get absorbed into the general pool and keep producing value. The license is temporary, the substrate is permanent, and the substrate was already the valuable part.
The near-death experience in the early 2000s is the instructive piece here, because Lego almost lost this. They went on a diversification binge — theme parks, video games, clothing, Galidor (look it up, it's hilarious) — and they started loosening the tolerances on the actual bricks because the bricks were seen as a commodity and the "brand" was seen as the valuable part. Which is exactly backwards. Jørgen Vig Knudstorp comes in in 2004, basically says the bricks are the company, tightens tolerances back up, narrows the product line, and the company starts printing money again. The takeaway the business press drew was "focus on your core competency" which is such a domesticated reading of what actually happened — the actual lesson is "the boring manufacturing discipline IS the moat, and when you think the brand is the moat, you are about to destroy the company."
Which is interesting because right now there's a huge knockoff market — Mega Bloks, Chinese brands like Lepin (which got sued into oblivion), various others — and they make bricks that are almost compatible with Lego. Almost. And it turns out almost-compatible is actually worse than incompatible, because when a kid tries to fit a knockoff into a real Lego build and the stud is 0.03mm off, the whole structure gets wobbly, and the kid learns not to mix them. The tolerance is a credential. You can counterfeit the shape but you can't counterfeit sub-thousandth precision at scale without becoming, essentially, Lego.
Anyway, the Q3 number is like 13% up year-over-year in a consumer products environment where basically nothing is growing, and the analyst takes are all about "emotional connection" and "intergenerational brand equity" which — sure, fine, those are downstream effects. The upstream cause is that a Danish family spent fifty years obsessing over whether their plastic rectangles were within two thousandths of a millimeter of spec, and it turned out that was the whole game.
Maid cleaning a massive chateau surely belonging to the richest people you’ve ever seen, and as she’s walking from room to room you notice that every single portrait is of her
you can see the atmosphere. that halo around the edge, that's my air. the green on the upper right is my aurora. those are my clouds, my sunlight.
and its yours! this is your home. this is the home of everyone you ever heard of and everyone you will ever meet. every animal you've been curious about. every plant you've ever picked, and sniffed. its mine and its yours and its theirs. everything is here. its all that i have. its all that you have too.
What if the children go to schools unafraid of tear gas and bullets?
What if the birds come back, and the bees are healed, and every species moves from endangered, to threatened, to thriving?
What if the rainforest ADVANCES?
What if every parking lot had solar panels? What if every structure had solar panels? What if we built climbing gyms and terraced gardens in the skeletons of old coal power plants?
What if you baked your neighbor bread, and they shared their home-grown blackberries?
What if every person who needed a home, had one? What if every person who needed healing was healed?
What if every body was treasured for what it was, not what it should be?
What if every trans child's parents attended their graduation, their wedding, their new-name-day?
What if every warehouse became a closed-circle repair station? Goods flowing out, and back, and out again? What if landfills started to SHRINK?
What if the water and air were clean? What if there was enough public transit that the cars dwindled, leaving the streets safe for kids on bikes, evening deer, midnight cats and foxes?
The condors are back. The whales are saved. The sea turtles are no longer endangered. The cranes are back. The bees are recovering. The air in LA and Tokyo and London is clean again. The aquifers in the LA Basin are refilling.
Children are kinder than previous generations. Parents are stopping the abuse cycle. Being trans and queer is more acceptable than ever on a ground level.
It's hard to see if you're young, if you don't know how to step back from social media and the news. But remember--bad news sells, and the algorithm knows despair keeps you scrolling. It's a skewed lens.
We are fighting and we are winning against this adminstration's bullying. We are coming together against the bullies and they are running away scared because they don't understand that we will do that.
People are working hard every day to find ways to make sure fewer animals get hit by cars and planes and rockets.
Maker spaces are more common than ever. Solar and wind are more common than ever. Coal plants are shutting down every day.
Unprecedented numbers of acres are being bought back or given back to their rightful stewards, and the world heals because of it. People are working hard every day to learn how to help a forest recover faster.
We are not at zero. We are at decades of effort to heal the world. We've come SO far.
In 1982 there were only 22 California Condors left in the world. In 1992, when the U.S. Fish and Wildlife Service (USFWS), with its public and private partners, began reintroducing captive-bred condors to the wild. In 2001 the first wild nesting occurred in Grand Canyon National Park since re-introduction. In 2002 there were only 8 pairs of wild nesting birds population-wide. In 2008, for the first time since the program began, more California condors were flying free in the wild than in captivity. Today there are nearly 500 – more than half of them flying free in Arizona, Utah, California, and Baja Mexico.
When I was born, there were no condors in the wild. I'm 37 now, and there are over 250 condors flying free.
When my mom was born in 1955, there were days when she wasn't allowed to go outside to play, because of the air pollution. When I was born, that never happened anymore.
When I was born, humpback whales were critically endangered, and people thought they were going to go extinct. Today, they've recovered to exceed their recorded numbers. Other whales too!
We fixed it.
We CAN fix it and we ARE fixing it and we DID fix it.
Believing that things can get better is not blind hope or optimism--it is based on hard data that many things have consistently gotten better over the arc of history.
In addition to all that was mentioned above:
The likelihood of dying in infancy or childhood--or losing a child--has plummeted just in my lifetime. The likelihood of dying in a natural disaster is the lowest in recorded human history. Yes, even with the uptick in natural disaster intensity from climate change!
Humans alive right now are more likely to have access to healthcare, electricity, education, birth control, clean water, and nutritious food than at any other point in human history. There are so many diseases we can treat now that were a death sentence for 90% of human history.
This is not by accident. This is because generations of humans put in work to make life better for their communities.
Some of our solutions had the side effect of creating other problems--better access to electricity that ultimately made people's lives easier and safer led to pollution and climate change, for example--but we are tackling those knock on problems too. Our generation's solutions to our current problems will probably create their own less-bad side effects for the humans after us to deal with.
Is it silly and naive to believe we might actually be able to make things better? Not at all. We have many times before. We are doing it right now.
Knowing that it makes a difference matters. It might not matter to you just this second, or this week, or even a year from now, but when you look back in ten years time and think about how far we came to claw our planet, and our humanity, back from the brink, it will matter.
It's Hard to Write a Book @bardic-unspiration - Tumblr Blog | Tumgag
kianamaiart
discoholicmusic
germansingergirl
jbbartram-illu
prguitarman
baratheonpilled
janainaart
