#AI and history | Explore Tumblr posts and blogs

reasonsforhope · 1 month ago

Text

"In the 1750s, an Italian farmer digging a well stumbled upon a lavish villa in the ruins of Herculaneum. Inside was a sprawling library with hundreds of scrolls, untouched since Mount Vesuvius’ eruption in 79 C.E. Some of them were still neatly tucked away on the shelves.

This staggering discovery was the only complete library from antiquity ever found. But when 18th-century scholars tried to unroll the charred papyrus, the scrolls crumbled to pieces. They became resigned to the fact that the text hidden inside wouldn’t be revealed during their lifetimes.

In recent years, however, researchers realized that they were living in the generation that would finally solve the puzzle. Using artificial intelligence, they’ve developed methods to peer inside the Herculaneum scrolls without damaging them, revealing short passages of ancient text.

This month, researchers announced a new breakthrough. While analyzing a scroll known as PHerc. 172, they determined its title: On Vices. Based on other works, they think the full title is On Vices and Their Opposite Virtues and in Whom They Are and About What.

“We are thrilled to share that the written title of this scroll has been recovered from deep inside its carbonized folds of papyrus,” the Vesuvius Challenge, which is leading efforts to decipher the scrolls, says in a statement. “This is the first time the title of a still-rolled Herculaneum scroll has ever been recovered noninvasively.”

On Vices was written by Philodemus, a Greek philosopher who lived in Herculaneum more than a century before Vesuvius’ eruption. Born around 110 B.C.E., Philodemus studied at a school in Athens founded several centuries earlier by the influential philosopher Epicurus, who believed in achieving happiness by pursuing certain specific forms of pleasure.

“This will be a great opportunity to learn more about Philodemus’ ethical views and to get a better view of the On Vices as a whole,” Michael McOsker, a papyrologist at University College London who is working with the Vesuvius Challenge, tells CNN’s Catherine Nicholls.

When it launched in 2023, the Vesuvius Challenge offered more than $1 million in prize money to citizen scientists around the world who could use A.I. to help decipher scans of the Herculaneum scrolls.

Spearheaded by Brent Seales, a computer scientist at the University of Kentucky, the team scanned several of the scrolls and uploaded the data for anyone to use. To earn the prize money, participants competed to be the first to reach a series of milestones.

Reading the papyrus involves solving several difficult problems. After the rolled-up scrolls are scanned, their many layers need to be separated out and flattened into two-dimensional segments. At that point, the carbon-based ink usually isn’t visible in the scans, so machine-learning models are necessary to identify the inked sections.

In late 2023, a computer science student revealed the first word on an unopened scroll: “porphyras,” an ancient Greek term for “purple.” Months later, participants worked out 2,000 characters of text, which discussed pleasures such as music and food.

But PHerc. 172 is different from these earlier scrolls. When researchers scanned it last summer, they realized that some of the ink was visible in the images. They aren’t sure why this scroll is so much more legible, though they hypothesize it’s because the ink contains a denser contaminant such as lead, according to the University of Oxford’s Bodleian Libraries, which houses the scroll.

In early May, the Vesuvius Challenge announced that contestants Marcel Roth and Micha Nowak, computer scientists at Germany’s University of Würzburg, would receive $60,000 for deciphering the title. Sean Johnson, a researcher with the Vesuvius Challenge, had independently identified the title around the same time.

Researchers are anticipating many more breakthroughs on the horizon. In the past three months alone, they’ve already scanned dozens of new scrolls.

“The pace is ramping up very quickly,” McOsker tells the Guardian’s Ian Sample. “All of the technological progress that’s been made on this has been in the last three to five years—and on the timescales of classicists, that’s unbelievable.”"

-via Smithsonian, May 16, 2025

604 notes · View notes

ellenjewettsculpture · 5 months ago

Text

Happy Year of the Snake! on venom and vigilance - 2025

460 notes · View notes

foxsketch6543 · 1 year ago

Text

Is it me or is history is repeating itself?

2K notes · View notes

its-not-a-pen · 9 months ago

Text

i love the hilarious eunuch ranking system by @welcometothejianghu so i decided to make one based on (mostly) REAL historical chinese enunchs!

in chronological order:

Warring States Period long story short, the PM is sleeping with the Empress Dowager, and he wants to extracate himself before her son (future Emperor Qin Shi Huang) gets old enough to find out. the PM finds her a suitable replacement, and the replacement is attached to a guy named Lao Ai. They pluck his beard and pass him off as a eunuch so he can sneak into the palace. Bing bang boom everyone's happy. This goes terribly wrong later, since Lao Ai tries to replace the emperor with his own kids and stages a failed coup. rest in pieces buddy...

it's very likely that Lao Ai was a ficticious character invented by Sima Qian, who will be appearing on this list later.

Qin Dynasty

Zhao Gao helped the first emperor of China conquer an empire, and administer it efficiently with his legal knowledge, but he also made the second emperor into a puppet, and weakened the empire for his own political gain. Max points of complexity, but you'd get more loyalty out of a coffee club punch card.

Before launching his soft coup, he decided to test the waters by bring in a deer and gaslighted the emperor by calling it a horse. the officials who were loyal to him called it a horse, and he executed the rest.

Han Dynasty Jiru, male favourite of Emperor Gaozu (Liu Bang), the peasant scoundrel who became the founder of the Han Dynasty. look, if the emperor has a harem of hundreds of women and you manage to catch his attention, you max out in style points. simple as. for most of chinese history it was fairly common for high-ranking men, especially eunuchs, to wear make up like powder and rouge, but i decided to give Jiru some women's huadia as well, cause he's a baddie.

Jiru gets a bad rap for alledgely distracting the emperor from his duties, but lets be real, history is written by civil officials who have no shortage of professional jealousy and gender/sexuality related prejudice towards eunuchs, since they were the personal attendants of royalty and could exert a lot of influence. plus Liu Bang was already pissing in the hats of confucian scholars, most of the poor work ethic is on HIM. Jiru should get credit for making him marginally less of a troglodite.

all in all he didn't try any court intrigue so extra points for loyalty and complexity. free my man >:( he's just a Han dynasty Monica Lewinsky who got slutshamed by jealous coworkers >:(

Han Dynasty

meet the father of east asian history, sima qian. half the people on this list can owe their placement here thanks to his extremely though history books "records of the grand historian"

history at this time was mostly "creative writing" and sima qian attempted to give the practise more academic intergrety, he went out and personally interviewed people, tried to get primary sources, and got rid of most of the more fanstastical aspects. however, he was not without his biases and some texts can be seen as allegorical/veiled insults towards the Han Dynasty, especially towards Emperor Wu. unlike most of the people on this list, sima qian was from the gentry and castated later in life as a punishment for treason. he was implicated after trying to defend a friend, and could not pay the fine to commute his sentence. the gentlemen at the time were expected to die by suicide rather than live with such ignimony, but sima qian chose to live so he could finish writing the history book his father started. the "giant conspiracy" joke explained: the chinese word for penis is a homophone for "conspiracy".

#chinese history #eunuchs #qin dynasty #han dynasty #warring states #zhao gao #lao ai #jiru #sima qian

836 notes · View notes

disease · 10 months ago

Text

Frank Rosenblatt, often cited as the Father of Machine Learning, photographed in 1960 alongside his most-notable invention: the Mark I Perceptron machine — a hardware implementation for the perceptron algorithm, the earliest example of an artificial neural network, est. 1943.

#frank rosenblatt #tech history #machine learning #neural network #artificial intelligence #AI #perceptron #60s #black and white #monochrome #technology #u

820 notes · View notes

dresshistorynerd · 3 months ago

Text

That ChatGPT ask about The Great Famine made me think about how generative AI is so devastating for historical understanding, but especially the understanding of dress history. There's a lot of completely inaccurate history AI slop, both generated images and writing. I'm always searching primary source images, but even then I've started to come across the occasional AI generated image, which is always very bad. Google image search is at this point almost useless.

The issue is that there's so much outdated information and frankly misinformation about history in the internet and in other source materials the AI scraps. This is of course an issue with many other fields as well, but history definitely belong to the fields which have most misinformation spread about them. When a human armed with a bit of critical thinking comes across a source they can evaluate it's validity. But AI doesn't have critical thinking. It can't think. So if there's a lot of sources, which peddle the same common misconception, AI weights it above what the few rigorous academic sources say. And then the AI becomes basically a misinformation peddling machine. And there's no way of finding our where it got it's information. Even the AI doesn't really know because of the way it works, and that's why, when you try to ask it, it tries to scours for sources that say similar things it was saying rather than giving the actual source it got the information from. And if it doesn't find a source? Well, it might hallucinate one. This is of course an issue in multiple level not just because it's filling the internet with misinformation and drowning out reliable information. It's also basically a plagiarism machine.

This is even worse with dress history than most historical fields, because dress history is such a new field and there's comparatively very little quality research. Even many historians of different fields peddle a lot of old misconceptions of dress history (sometimes you even come across a fairly new paper on dress history and there's some misconception repeated as a truth). So of course AI is very unlikely to produce anything close to accurate information on dress history.

So to see for myself how bad it actually is (and to have a bit of laugh) I decided to try out ChatGPT for the first time. I decided I would ask it a question I have answered in a blog post myself before to see how close (or far) it got.

Here's what I asked: "When did it become socially acceptable for men to wear dresses?"

I have written a very long post on this before found here.

And here's the answer ChatGPT gave with my commentary:

The idea of men wearing dresses has evolved significantly over time, with different periods in history showcasing varying degrees of acceptance. In general, men have worn clothing that could be considered "dress-like" at various points in history, though the specific social context has changed. 1. Ancient and Medieval Times: Men in many ancient cultures, including the Greeks, Romans, and Egyptians, wore tunics, robes, or other garments that we might now think of as dresses. These were considered appropriate for men, often signifying social status, wealth, or ceremonial importance.

Not bad, though this makes it sound like men only wore dresses to signify status, wealth or ceremonial importance, when in fact all men wore dresses. I also made a point in my post to say these were dresses, not just "dress-like". They were (mostly) constructed and (mostly) called the same as the equivalent clothing women wore, so of course I take issue with the framing.

2. Renaissance and Early Modern Period: During the Renaissance and up through the 17th century, men’s fashion included elaborate gowns, skirts, and dresses, often with frills, lace, and rich fabrics. These were considered fashionable and elegant for men, particularly among the aristocracy.

Here's were we're getting to the inaccurate territory. Yes still during renaissance men were wearing dresses, however it was becoming fashionable among especially young noble (and soldier class) men to wear extremely tiny skirts, until they stopped being skirts and turned into very fitted thighs (hose) and doublet. It was the period when dress became increasingly less acceptable for men to wear, but it was precisely lower class and peasant men, who were slower to follow the new fashions and masculine standards, not very surprisingly. But by Early Modern Period it had become unacceptable for men to wear dresses. Yes, there was still hems at points, but those were coat hems, which might seem like basically the same as a skirt hem, but conceptually they were different. It's the same today. No one thinks nothing of it if a man is wearing a long overcoat, with buttons closed, on top of a suit, but if they are wearing thighs and same length skirt? Well, for the next couple of weeks you'll get outrage op-eds and conservative influencers losing their minds over "the death of Western civilization".

Also in 1600s (roughly around 1630s-1660s) rich fabrics and elaborated detailing was not fashionable, not among women and certainly not among men. Men's fashion especially favoured a lot of black and other dark and neutral tones, influenced by somber religious atmosphere influenced by the all the Reformations and Counter-Reformations and wars about them going on at the time. (Of course there's some counter examples, but the general statement ChatGPT gave to me is simple incorrect.)

3. 18th and 19th Century: By the 18th century, Western fashion for men became more structured, and by the Victorian era (19th century), men were expected to wear trousers, waistcoats, and suits (my comment: you mean coat right, ChatGPT?? suit already encompases trousers and waistcoat). The shift towards more gender-specific clothing for men and women took root, and women’s dresses became more ornate and feminine, while men’s clothing became more utilitarian and restrictive.

First of all the height of structuring in Western fashion for both men and women, was during the Elizabethan Era, and additionally men's fashion got briefly a little less structured during the 18th century. As I explain in my post about the actual origins of modern men's suit (it was not Beau Brummel), the three piece suit with trousers, waistcoat and coat was popularized in 1670s. Women's dresses becoming more "feminine" means literally nothing when what was feminine was just seen differently. Of course it became more feminine in our modern eyes, when we get closer to the present day in the historical timeline, because modern femininity was being formed at the time. Men's fashion did become more utilitarian (I talk about a bit about the reasons why in the post about three piece suit above), but I wouldn't call it more restrictive necessarily.

Points for ChatGPT though for acknowledging that fashion became more gender specific around 18th and 19th centuries.

4. 20th Century: The 20th century saw men in dresses or skirts make occasional appearances in fashion, especially in avant-garde or countercultural movements. For example, in the 1920s, some male actors in Hollywood (my comment: here it's "forgetting" where it's going with the sentence, because it doesn't understand what a cohesive sentence means), such as Marlene Dietrich’s famous cross-dressing moments, challenged gender norms. In the 1960s and 1970s, figures like David Bowie, with his androgynous style, brought more attention to the concept of men wearing dresses as part of a wider push for gender fluidity in fashion.

The fact that cross-dressing became more acceptable in the early 20th century for women and people seen as women, didn't really do anything (or very little) to the make cross-dressing more acceptable for men, or those seen as such. Counter culture fashions in 70s and 80s were much more accepting of androgynous men and men wearing skirts, but the wider mainstream fashion was absolutely not accepting of that, definitely not outside the rock icons, who themselves were also side-eyed a lot.

5. Late 20th to 21st Century: By the 1990s and 2000s, with the rise of gender nonconformity and movements pushing for gender equality and expression, men wearing dresses became more visible and socially accepted. Public figures like Prince, Kurt Cobain, and later, celebrities like Harry Styles and Billy Porter, have worn dresses and skirts as part of their personal style, pushing the boundaries of traditional gendered fashion.

It's absolutely still not seen as acceptable for men to wear dresses. Yes, couple of celebrities have done that, and yes there's alternative fashion and queer people who do that, but if you take a look at how people in legacy and social media talk about them, it's pretty obvious that's not socially acceptable in the mainstream.

In essence, it became more socially acceptable for men to wear dresses as part of a broader cultural shift toward embracing non-binary gender expression and rejecting rigid gender norms, especially in the 21st century. Celebrities, designers, and activists have played key roles in challenging the old norms, making fashion more inclusive and fluid. This acceptance continues to grow in many parts of the world, although social attitudes can still vary greatly depending on cultural context.

"A broader cultural shift toward embracing non-binary gender expression and rejecting rigid gender norms." Laughing crying. Yeah, there was a cultural shift, but clearly it was not super broad. It's really laughable to suggest in the heart of the reactionary backlash moment that it's somehow broadly socially acceptable for men to wear dresses.

Also this "although social attitudes can still vary greatly depending on cultural context" makes it sound like outside the western context, which this answer is all about, the attitudes haven't changes as much, when in fact outside the western context dresses for men are much more often acceptable and part of the cultural norms.

Perhaps it was unfair of me to ask my question in a misleading way, since the answer to "when did it become acceptable for men to wear dresses" is "it was for a very long time, but it's not anymore". But this is part of the issue with ChatGPT and similar text based AI. When you ask them a question they will take the assumptions of the question and run with them, even if the assumptions in the question are completely false. It's not just a misinformation machine, but a misinformation machine designed to reinforce your preconceived notions. For example if you would ask ChatGPT "why were corsets unhealthy?" it might answer with a very long answer full of pseudoscientific misinformation straight from the mouths of Victorian eugenists. I did that and the answer was extremely predictably. I'm not going to go through all that since is was so very long and this is getting long enough, but for a taste, after all the bullshit reasons it gave, here's what it wrote at the end:

Why Corsets Were Worn Despite Health Risks: Fashion and Aesthetic Appeal: Corsets were primarily worn to achieve a specific beauty standard that emphasized an extremely small waist and a curvy figure. This was considered a symbol of femininity, refinement, and social status during certain historical periods. Cultural Norms: The societal pressure to conform to the idealized figure meant many women endured discomfort and health risks to fit into the norms of the time.

The biggest eyeroll. Of course nothing about how it distributes the weight of the boobs much more comfortably in the waist rather than the shoulders, or how the heavy dresses people wore needed support or movement would have been very difficult, or how it was an easy way to alter the body, any kind of body, to the fashionable silhouette without needing any invasive procedures or dangerous diets. The most classic, most predictable corset misinformation. This is exactly what I meant at the beginning. The internet is so full of this bullshit already, in the AI algorithm it makes it true.

The lesson is: do not use ChatCPT as your search engine and don't trust it as your source of information. There is numerous of issues with generative AI and the spread of misinformation is certainly one of the biggest, seeing how much death and destruction misinformation has caused only with vanilla social media. I'm afraid to think about the repercussions of refining that metaphorical misinformation machine with the literal misinformation machine. If you want to do research into dress history yourself, but don't know where to start, rather than asking ChatGPT, here's my tips for where to start.

Also be skeptical of the sources you find in the internet, including everything I write. In the age of generative AI it's even more important to use proper sourcing. I'm definitely not perfect about this, since I'm extremely chaotic in my research most of the time and bad at keeping my sources in good order (I don't want to do boring stuff when researching for my hobby, I do that enough for uni), but I'm making an effort to be better about it, at least for my more structured posts. But if you want to know any sources I've used for any reason, when I haven't listed them, you're always free to ask. I might have it somewhere among my 500 open tabs still open, or I might still be able to find it again. So anyway, be skeptical of writings (again, including mine) which don't have sources listed, not just because it could be AI, but also because the same issues apply outside AI - it's much harder to evaluate how trustworthy the writing is or to know weather it's plagiarized.

#anti ai #history #research #historical research #generative ai #chat gpt #dress history

310 notes · View notes