The acrobat seal from Safonovo
In the village of Safonovo, a charming sea guest has taken a liking to a resting spot near the famous Be-6 aircraft monument.
The pinniped performer regularly swims to the memorial to bask and show off his agile form against the backdrop of the historical landmark.
The seal so accurately imitates the curve of the fuselage and the takeoff direction of the winged machine that it seems to be intentionally parodying the Soviet seaplane for the sake of spectacular shots.
So guess what? The composer of this legendary JRPG-esque song for The Weather Channel actually came forward recently, having learned about the song’s quasi-meme status. His name is Chris Kennedy.
After looking through the evidence, both, but moreso the second.
Ea-Nasir's tablet is dated to 1750 BC, which is coincidentally aligns to the death of Hammurabi. For context, he lived at the end of the Isin-Larsa Period, a time in Babylonia's history where it was a collection of warring city-states. Ur and Larsa were the most powerful of these, since they were farthest south and controlled most of the trade coming up the Persian Gulf. (Isin, near where Hammurabi was from, was in the North and had lost power about 200 years before.)
Right after Hammurabi's death, all the city-states he'd conquered, including Larsa and Ur, decided that they didn't give two squats what the people in the North thought, and started a rebellion.
The tablets in Ea-Nasir's house have been translated. It's very difficult to find them, but the book is called Foreign Trade in the Old Babylonian Period, Leemans 1960, and he makes a series of interpretations that still align with our understanding of the culture today:
Ea-Nasir was hot-headed. 3 tablets note him talking rudely to messengers and traders.
Ea-Nasir sold copper to private merchants AND the temple, which was the government of Ur. The receipt we found is in such a large quantity we can assume the government was likely his primary buyer. The complaint tablets are from notably from private merchants.
Ea-Nasir was an alik-Tilmun; or 'one who travels to Dilmun'.
Where is Dilmun? Good question! Archaeologists spent the next 40 years figuring it out! At this point, they're fairly certain it's in present-day Qatar. The city was used as a midpoint port to bring in copper from Magan and Meluhha (current-day UAE/Oman and India respectively.)
The reason we know this, is because Oman is an old, old copper-producing region. It's an ophiolite (rock from the seafloor that's been uplifted to the surface) that contained a spreading center (think Mid-Atlantic ridge) which forms deposits of copper and other metals as sulfides from the black smoker vents (copper-iron sulfur, lead sulfur, zinc sulfur, etc.)
To produce copper, you have to remove the iron and the sulfur. To remove the iron, you add "flux", which essentially bonds iron to silica, because it likes silica more than copper does. And to remove the sulfur, you add oxygen, which burns off the sulfur as gaseous SO2.
The copper is heavier than the iron and silica, and sinks to the bottom of the furnace. The iron and silica, slag, flow out the side. The resulting ingot looks like the bowl below. And a lot of times, holes remained from gas getting trapped at the bottom.
They measured copper by weight though, so this wasn't too much of a problem. However, if there weren't enough flux, or the fire wasn't hot enough, iron would also get trapped in the copper ingot, making "black copper"; if a merchant wanted the 97% pure copper that could be made using this process, a lot of iron would definitely be considered 'bad copper'.
Switching back to the culture!
Around 1800 BC, the same time as this was going on, the culture of Oman underwent a noticeable decline. Many of their coastal mines stopped producing copper and people moved inland. They also stopped making bronze with tin. This is notable, because tin was scarce in the Bronze Age and insinuates they might've been left out of the trade route. At the very least, they had stopped being Mesopotamia's primary supplier and started doing their best to keep up with the times.
(At this point, I'll point a finger to Cyprus, which was firing up its smelters at the same time. Cyprus is very interesting, but it pertains less to Ea-Nasir, so I'll just wave enthusiastically at their oxhide ingot copper and tin trade domination.)
So we can't know if Ea-Nasir wasn't a chronic scammer, but I think all the evidence outlines a different story.
Ur, a powerful city-state rebelling against a conqueror within Ea-Nasir's lifetime. Ea-Nasir, selling large amounts of copper to the government, and smaller sales to private merchants who complained about being given scraps; a man who was still traveling to trade copper in a state that had lost their monopoly on the copper trade and was possibly producing some less-than-ideal quality.
He mostly sounds like a person with strong ties to his city and culture. Maybe not the best copper merchant, but certainly a passionate one.
References below the cut:
Velde, C. (2003). Wadi Suq and Late Bronze Age in the Oman Peninsula. Archaeology of the United Arab Emirates, 102–113.
Bibby, G. (1970). Looking for Dilmun (Second Edi). Alfred A. Knoff, INC.
Olijdam, E. (2014). From rags to riches: three crucial steps in Dilmun’s rise to fame (poster). Proceedings of the Seminar for Arabian Studies, 44(July 2013), 277–286.
Forbes, R. J. (1950). Metallurgy in Antiquity. EJB. https://doi.org/10.1126/science.207.4426.50
Leemans, W. F. (1960). Foreign Trade in the Old Babylonia Period. Leiden, E.J. Brill.
Kassianidou, V. (2013). The Production and Trade of Cypriot Copper in the Late Bronze Age. An Analysis of the Evidence. Pasiphae. Rivista Di Filologia e Antichità Egee, VII, 133–146. www.libraweb.net
Khol, P., & Bertille, L. (2008). By Land and By Sea: The Circulation of Materials and Peoples, ca. 3500 - 1800 B.C. Intercultural Relations Between South and Southwest Asia: Studies in Commemoration of E.C.L. During Caspers (1934 - 1996) BAR International Series 1826, 1826, 29–42.
Giardino C., (2019). Magan - The Land of Copper: Prehistoric Metallurgy of Oman. Ministry of Heritage and Culture - Sultanate of Oman (2019).
I must say, Steven Universe is a lot more enjoyable when you don't need to read through a heaping helping of Tumblr discourse about who's secretly abusing whom this week.
theist accelerationism: the world needs to get as bad as possible as fast as possible so God is forced to intervene
atheist accelerationism: the world needs to get as bad as possible as fast as possible in order to trigger a collapse of the current economic and socio-political structure
agnostic accelerationism: nobody knows why the world needs to get as bad as possible as fast as possible
Copyright governs who has the "right" to produce and distribute "copies" of books/music/movies/creative works. This is where fair use doctrine applies, because most creative works are referential by nature.
Weird Al is allowed to parody everything because he's operating under copyright law, not trademark law.
Trademark governs who can "trade" under what "mark" i.e. the brand identity of a company. Companies don't own their trademarked word forever, but they maintain the exclusive right to sell things under that brand in their specific market sector. Patagonia doesn't own the name of a geographical region, they just own the right to be the only company using that name to sell clothing and outdoor gear.
A drag queen name can be a parody of a clothing and outdoor gear company.
A company's trademarked logo can be used in parody creative works, with more leeway if it's not for commercial purposes. Trademark parody is allowed! Patagonia has been aware of and allowed Pattie Gonia's trademark parody for years.
Trademarks are specific to market sector. Actress Chase Infiniti could start a makeup line named after herself and her trademark would not infringe on the Infiniti car brand because they are different markets and there is no risk of confusion. Pattie Gonia could probably trademark her name to sell frozen veggie burgers and Patagonia would not care.
Drag queen Jan Sport did a collab with JanSport bags. What Jan Sport almost certainly did not do is independently apply to register "Jan Sport" as a trademark in order to sell bags on her own, because that would infringe on JanSport's own trademark in the bag market sector.
What Pattie Gonia is not allowed to do -- the thing that Pattie Gonia actually did do and is being sued for -- is apply to register "Pattie Gonia" as a trademark to sell clothing, because apparently Pattie is in talks with North Face and HydroFlask to sell "Pattie Gonia"-branded gear. These companies probably won't finalize anything unless Pattie shows that she actually owns the trademark. Unfortunately, "Patagonia" is already a registered trademark in the clothing market sector, and these two names are too similar to exist in the same sector (see: "likelihood of confusion" legal standard).
Your drag queen name can parody a clothing company. You can parody the trademarked logo of a clothing company. But you cannot use the same name to then go on to also become a clothing company.
In order to maintain their own trademark, Patagonia must sue for trademark infringement. If they don't sue, and Pattie Gonia gets her own trademark, Pattie could sue Patagonia for infringement on her trademark. You can see why Patagonia won't be dropping this suit no matter how much you harass them.
Yes, Pattie's legal fees to fight this will cost more than the $1 she's being sued for. Pattie could also not fight this, withdraw her trademark application, not spend any money, and carry on being an environmental activist drag queen named Pattie Gonia. She would probably be better off making nice with Patagonia in the hopes of a Jan Sport-esque deal where Pattie designs an exclusive fabric and Patagonia maintains the trademark, but apparently Pattie's legal team has been sassing off to Patagonia in their communications for years, has applied for a trademark they should 100% know they'll never get, and has now decided to play the victim on social media just in time for Pride month, so I don't know how likely that is. I guess we'll see!
This is mostly correct, but I’d like to offer a small correction. The product deal with Hydroflask and North Face apparently occurred in 2022, and HydroFlask got Patagonia involved to make sure everything was in the clear. It seems like Patagonia was very agreeable about everything at the time, and only asked that Pattie Gonia and her partners avoid using the Patagonia logo and font or similar images, and to avoid putting the words “Pattie Gonia” on any products. This is the email exchange from 2022, from the recent Patagonia trademark complaint, including Pattie Gonia apparently agreeing to the limitations.
The new conflict is from Pattie Gonia using the Patagonia imagery and the Pattie Gonia name on her own merchandise. This is the email Patagonia sent, with the images they feel conflict with the 2022 agreement.
Pattie responded to that by disagreeing that she had broken any agreement, and also obliquely threatening to expose Patagonia for making tactical gear for the US military?
It’s possible that Patagonia understood the terms from 2022 to be a good-faith ongoing agreement about keeping the brands separate, and Pattie interpreted it as an agreement limited to the now-ended North Face and Hydroflask collaboration. It’s also possible that Pattie Gonia didn’t believe she was actually agreeing to anything at all, since her responses were very neutral, though positive in tone, up until 2025. The email chain does, however, show what I think is a very clear effort on Patagonia’s part to protect their trademark while also showing support and goodwill towards Pattie in her use of the Pattie Gonia stage persona.
I seriously think that in the next couple years you're going to see so many people experience paranoia re: whether any given text/image they're looking at is AI-generated that it will be a clinically-recognized condition. It's happening to me to some extent, and I'm watching it happen on art sites like DA/Pixiv.
Within the last couple days both Sam Kriss and Alexander Wales have made “the AI text is everywhere man” posts, the former spitting mad about it and the latter just kind of exhausted and defeated. And I definitely share a bunch of this feeling, but at the same time, sometimes humans do in fact use em-dashes and say “it’s not X, it’s Y”. I don’t think being hypervigilant for these kinds of questionably-accurate tics is a healthy way to engage with the internet. I think over time it will drive you insane, even more assuredly than the psychosis of the people who use LLMs too much.
Yeah, the conversation over "AI tells" has gotten really heated just over the past few days. I suppose the proximal cause is the Commonwealth story, but that feels more like an pressure release than a reason.
Obviously I am biased, but the valence of the conversation seems so weird to me. Like no one on either side are treating the works of art (text or visual art) as an object you want for itself and its own qualities. If a story is bad, say it's bad. If it's good, enjoy it. Commentators seem much less willing to say "this writing is simply bad."
Yeah, the advertisements directed at you are annoyingly bad. Does anyone remember the discourse over restaurant websites a decade ago?
Instead, you get a gnostic puzzle of "these phrases and this way of using a metaphor and this particular vocab tell you its provenance, and that drives me into a rage."
Of course if you see the same tics repeated several times you will get sick of them, but what you find out is that only 1% of people are reading enough to get triggered by it, and at least 90% of people respond blankly, completely unaware of the Not X It's Y patterns and if you have to describe the patterns to them, they don't get why it's bad or you're so upset about it.
But putting it that way sounds like "you should like my bot's output and if you don't you're being too picky." And that way lies just as much error.
The Too Hot for Tumblr Take is that Sam's (who I really adore) line "But you still keep letting it do your writing for you, as if I won’t be able to tell. Listen: I can tell. I can always tell. You think I won’t notice, but I will. There’s no hiding from me." sounds hauntingly like someone swearing there is no such thing as a passing trans person and they can always tell.
I understand where this is coming from, but I don't think this is getting the key points? I have no doubt that the people you are talking about are real! But what Sam Kriss is noticing, alongside many others, is this very real "collapse" in the quality of writing around them. It honestly would not matter that much if all of the pieces were written by humans, right? If everyone just organically started writing like an overworked LinkedIn consultancy pitch, a lot people would also be pretty flustered and annoyed.
From this lens the story is about a cascade that only starts with LLMs. We live in a world of editors & gatekeepers - you don't go into the Library of Babel and pull out books to read, you are directed to content by algorithms, magazines, popularity, etc. And those institutions used to halt, at least somewhat, the spread of drivel. They are not halting LLM content - particularly the social media algs & popularity contests. Your "feed" every day now has something "eveyone" is discussing that has the same repetitive, overwrought prose as the last thing that made the rounds. And you can say "just don't read it", but opting out of society like that is not the easy choice it is framed to be! If you wanna play the game of society now you have to put up with it when you didn't before.
And you *really* have to put up with it based on your field, because a lot of people are not writers because they want to be; they write because they have to. They are tech folk, politicians, business types, whatever, writing factual pieces about useful information. When some insider posted about the latest DoD wargaming in 2019 or whatever, they had to put in the work to make the post passable because they had to write it themselves. Now these busy people are increasingly just skipping that by throwing their notes file in Claude. "Good enough!" And it is good enough for the median person, it goes viral or whatever.
But it is absolutely 20% worse than the posts from the past, and also has identical prose to the five other viral posts about trade policy and such you just read. Not everyone can eat the same exact meal every day, even if it is an incredible dish (and these are not) - the human brain isn't built for that.
LLM writing is an inadequate equilibrium. It is *juuuust* good enough to pass all our previous social bullshit filters, but not good enough to the fake the quality those filters were meant to promote. And of course, today is probably the least amount of LLM use writers will ever do. You cope, you follow the good writers, its ~fine. But it is very annoying.
From this lens the story is about a cascade that only starts with LLMs
But the cascade started well before LLM's, right? Sloppification in websites and reality shows and clickbait was a very widespread problem already. LLM's have accelerated it, but I'm not sure LLM's would be used the same way if they hadn't emerged right when "more content for the content god" was the economic modality.
I certainly don't think, uh, grilling caterer websites were any better before AI.
Obviously the specific thing Sam was describing is very, very annoying. The question is whether that is about AI or about the algorithmic, context-free trend of internet content already. It's like blaming CGI for why there are all these badly written Marvel movies, and luridly describing the skin texture of some alien battle suit as your punch. The phenomenon are related, but there's confounders of both sides. CGI in good movies, mindless trash without CGI, bad writing slop before AI, and theoretically AI writing that is enjoyable in itself.
Which makes the question of "how do we differentiate" then very important, because are we talking about AI tells or bad writing tells? Which is the folly of only using tremendously bad examples for your polemical; it doesn't tell us where the line is.
Don't even get me started on Litrpg's.
The real horror of the Commonwealth scandal is: what would have happened if the story was not noticed as a technological problem. Would anyone care that a prestigious cash prize was given to such a terrible story, or would everyone just shrug about that particular genre.
https://www.theargumentmag.com/p/the-literary-world-is-sleepwalking?publication_id=5247799&post_id=199248169 Gives a great differentiation between the sort of metaphors bad AI uses, and the sort of metaphors a good author uses. So you gotta ask - why weren't you already calling out the bad metaphors, without needing AI as an excuse?
But the cascade started well before LLM's, right? Sloppification in websites and reality shows and clickbait was a very widespread problem already
Yes, and it sucked! And it sucks that it is getting worse.
I certainly don't think, uh, grilling caterer websites were any better before AI.
Here I actually disagree - it wasn't by much but they were better. Obviously I don't have a dataset of grilling websites specifically, most restaurant websites haven't chnaged in a decade, but I have seen a lot of "everyday prose" go from "boring but utilitarian" or "a little cheesy but it has its own flair" to "oh this was so obviously written by Claude in 3 seconds from a bullet point list". And those last ones are worse - they communicate the information inefficiently and have meaningless turns of phrase. And of course they are boring.
This is a lot more blatant when it comes to images/art for a lot of places, for example - lots of cute art and stylistic imagery on everything from menu ads to flyers for school festivals has been replaced by a standardized image with the identical shadings and font styles of someone's first attempt at making AI art. In the past did the median school flyer probably pull some random art from the bowels of the internet? Yep! And that resulted in more diverse art than today. It is a loss.
The question is whether that is about AI or about the algorithmic, context-free trend of internet content already.
I won't like defend Sam's piece, I am not him, so don't take my own replies as that. If your objections are pointed that way I won't stop you. For my own take though, both are "equally" to blame and the distinction doesn't matter too much? Because it probably won't self correct. I do not think algorithmic systems will adjust to clamp down on this. After all, it works for the median person, right? I don't see any reason this would really shift. LLMs are the proximate cause, and there certainly isn't any moral arguments around blame to be had here. So we can just note the observation.
So you gotta ask - why weren't you already calling out the bad metaphors, without needing AI as an excuse?
We totally were! We absolutely, 100% were, a lot. People have been writing all about the shitty prose of YA novels, the bland repetitiveness of LitRPGs as you mention, the insane stock formula of LinkedIn spam, the crude heartstring-pulls of modern litfic, the works! And it all just got twice as bad. So we are writing about it twice as often.
(Ofc there are good versions of every genre mentioned above - every medium has its overused tropes, it doesn't make it all bad. Except maybe the LinkedIn spam, but I'll keep an open mind?)
germansingergirl
janainaart
annaxmalina
discoholicmusic
gummygunk
yousharknotpass
theartofmadeline
andulkaphoto
