Please be aware of common scams! If someone reaches out to you with a scam, please report and block them. Outside of our support email and official [tumblr] accounts ONLY on [tumblr] (when in doubt, look for the same Staff Member badge that you see on this blog), Staff will never message you or reply to your posts to notify you about changes to your account. Also, [tumblr] will never ask for money besides for Premium subscriptions, domains, paid themes, and what's sold in TumblrMart.
🌟 New
You can now sort your likes from oldest to newest on web and iOS. Do you remember what your first liked post was?
You can now search for posts within Communities on web.
Redesigned blog headers on iOS — Same as Android, now following and subscribing to a blog are front and center.
Tapping "Not interested in this post" now temporarily filters that blog and their posts from recommendations. This is an additional improvement to how "Not interested in this post" works. We had announced that tapping this option will down-weight the tags from that post in your interest profile, and filter out recommended posts similar to the ones you have dismissed.
A new Media auto-play setting on web lets you choose whether GIFs and videos auto-play, independent of your system's reduced-motion preference. Find it in Settings → Dashboard.
The reblog graph explorer on web now shows at a glance which reblogs have commentary.
🛠 Fixed
We fixed an issue on web where high-resolution image thumbnails appeared noticeably degraded in quality.
We fixed an issue on iOS that was causing crashes related to image decoding, particularly with GIF-heavy content.
We fixed an issue in the iOS and Android apps where users with unverified age in mature-labeled communities couldn't see the "Leave community" button.
We fixed an issue in the iOS app where the Unblock button in the Blocked Blogs list stopped responding to taps after scrolling.
We fixed an issue where blogs previously marked as "not mature" by our team couldn't add content labels to their posts.
We fixed an issue where blocking anonymous ask senders from a sideblog would fail with an error.
We fixed an issue where users on older operating systems encountered SSL errors when clicking password reset or email verification links.
🚧 Ongoing
We're aware that some blog exports are failing. If you're affected, please file a Support Request so we can track your case.
We're investigating reports of some image thumbnails appearing pixelated in the mobile apps, despite looking fine on web.
We're investigating reports from some iOS users who are unable to view mature content despite having their settings correctly configured.
We're aware that some users are seeing an empty For You feed. We're investigating.
We're aware that some Firefox users are encountering errors when editing and publishing posts or sending messages. Sometimes, trying a second time works. We're investigating.
We're also aware that the Home and Activity notification badges are getting stuck for some Firefox users. We're investigating.
🌱 Upcoming
We're trying out a redesigned Trending tab on web, with topic blocks showing follower counts, preview posts, and related tags for each trending topic.
We're trying out a new post editor entry at the top of the dashboard on web.
As shared previously, we're building some new features for replies, including a new blog tab to see a blog's comments!
We're working on some updates to account deletion. In the future, when you delete your account, we will keep your data for 30 days. This will allow you to restore your account if you regret your decision, deleted your account by accident, or your account is hacked and deleted.
Experiencing an issue? Check for Known Issues and file a Support Request if you have something new. We'll get back to you as soon as we can!
Wanna go ad-free and get a bundle of other perks? Check out Tumblr Premium!
its probably a normal sign for the economy that all of my adulthood fantasies are like "imagine having your own kitchen living room and bathroom to decorate" "what if i could get on a train" "maybe one day i could purchase a sturdy pair of shoes" "i should save and invest in a single bicycle"
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.
this shit from the comments is unironically so fucking sad to me. this is where we’re at now? “aren’t books supposed to have morals?” genuinely let’s all just pack it in and go home, we tried the whole “experiencing life and art at a greater complexity level than an eight year old can handle” thing and it didn’t work out, somebody break the news to oscar wilde, we’re done here. “books are meant to teach you something.” christ.
Ok extra context from your friendly neighborhood Master's Degree holding archivist:
Archives are "curated" according to what's called a "collection policy". That means a policy that dictates what you collect; usually along a theme. For example at the university I used to work at the collection policy was to collect archival material pertaining to the local area and its history/important figures. At the historic home I currently work at, the collection policy covers materials pertaining to the person who lived at the house.
AO3 collects fan works and original writing. That's it. That's how they "curate". They don't allow a copy paste of published works (for legal reasons) and they don't allow commissioned works (for legal reasons) . Everything else that falls under fan works and original writing is part of their collection policy. Everything. That's how archives work.
Slacktivism has made me hate the "privilege" argument and even the word because 90% of the time it's used to shame people. For having social advantages, for not knowing/remembering social things, for having a different opinion about...well, anything, for accidental insensitivity, etc.
And it sucks because privilege is technically a very real thing and people do have it. But at some point "check your privilege" went from "remember that experiences and world views beyond your own exist and don't dismiss or belittle them" to "feel guilty and self-conscious for how your cards stacked up".
Also it would be nice if we stopped throwing "privilege" around like a condemnation and started focusing on how some people DO use their privilege for the betterment of society. Not everyone who has money or clout or connections is a big bad billionaire pissing in the water supply.
“why are you, as someone in their 30s, still on tumblr” oh so you think you’re gonna be normal when you’re my age? you think you’re gonna be CURED?? you think the witches’ curse will have been lifted by then?? cmon now
One hour and seventeen minutes into the first episode, Mahidevran is already crying with them sad violins in the background. How anyone came to stan a crybaby who's genuinely also politically stupid is beyond me.
The sad part is there’s a decent chance a large proportion of them can be blamed on one spider.
The tarantula that bit JRR Tolkien as a child.
He swore he didn’t have a spider phobia and the experience had nothing to do with the man-eating giant spiders in The Hobbit, the even more giant and even more man-eating spider in Lord of the Rings, or the unholy eldritch spider from outside creation that plunged the world into darkness and made literal Satan scream like a little kid in the Silmarillion. Very few people believe him.
Given LotR’s influence in the fantasy genre, there is a high probability that tarantula is the progenitor of even more fictional spiders than Ungoliant was.
“fantasy universes have too many spiders” factoid actually just statistical error. Georgs Spider, who bit JRR Tolkein & is to blame for menacing over 10,000 fantasy universes, is an outlier adn should not have been counted
I absolutely despise ambiguous endings to stories, especially in shows that ran long and had constant influx of drama and angst and horror and violence. I feel the same about depressive and negative endings.
baratheonpilled
snowghoul
vaporwave
gummygunk
kaledo-art
andulkaphoto
