drinking beer with crazy straws and enslaved readers as the AI of antiquity

A conversation with Jackie Murray about Plato and how AI hurts students and scholars

About a year ago, I was in a bar in Manhattan when I saw a neon sign on the wall that purported to be a quote from Plato about beer. I was immediately suspicious — was there even beer in classical Athens? I texted

Sarah E. Bond

, who has given multiple talks at breweries about beer in antiquity. She told me there’s actually some scholarly debate about this very question and directed me to this excellent article by Max Nelson.

(Side note — why did I write about classical reception among horrible internet trolls when I could have written about beer and gotten invited to give lectures at breweries?! And for what other unexpected venues might that be possible? Could I write about stuffed animals in antiquity and get myself invited to lecture at Build-a-Bear?)

But not everyone is lucky enough to have a beer scholar in their contacts. These days, most people use AI for this kind of question. So I also asked ChatGPT for a fact check, and I was truly shook by how hilarious and bad and delightful its response was. I have a whole section of this newsletter dedicated to dunking on AI, but it still never fails to surprise me. It was so absurd that I decided I needed a Plato scholar to weigh in, so I wrote to Jackie Murray.

Readers may remember Jackie from my previous conversation with her about trans resonances in Odyssey 6:

I knew Jackie was working on a book about slavery in Plato — which, she then told me, has now manifested itself into three separate projects (a long book, a short book, and an article), which does have a way of happening in academia. She’s also working on a book about Apollonius and an art book. She graciously took time away from her extremely busy schedule to talk to me about the Theory of Forms and AI.

Jackie Murray is an Associate Professor of Classics at the State University at Buffalo. She grew up in Toronto. Her research areas and publications are in Hellenistic and Latin Poetry, Race and Ethnicity in Antiquity, and Black Classicisms, especially the reception of Classics in African American and Afro-Caribbean literature. Her monograph, Neikos: Apollonius’ Argonautica and the Poetics of Controversy is under contract with Harvard University Press.

This conversation has been lightly edited and condensed for clarity.

---

Thank you for making time to talk with me. So I was in New York in a bar, and they had one of those neon signs up, and it said, “he was a wise man who invented beer. – Plato.” So I put it into ChatGPT and asked it for a fact check.

Right.

And ChatGPT wrote… you're not ready for this:

The quote, ‘He was a wise man who invented beer,’ attributed to Plato, is widely regarded as misattributed. Plato didn't believe that anything could be truly invented, as he thought a form of beer always existed and could be discovered, not invented. This stance aligns with his Theory of Forms, where he posited that non-material abstract forms, and not the material world of change known to us through sensation, possess the highest and most fundamental kind of reality. Moreover, various online sources discussing misattributed quotes agree that this saying is not found in any of Plato's known works.

So as a scholar of Plato writing an article, a small book, and a large book on Plato, I'd like to hear what you think of this use of Plato's Theory of Forms by ChatGPT to critique and fact check this sign in a bar in New York.

Okay, we can agree with the fact that he never said that. It's hilarious. Well, I think it's an ingenious approach to the Theory of Forms. But just because something is a form doesn't mean it can't be invented.

If you're literally talking about like Latin invenire, it's to find, right? Almost like to discover. Which is completely consistent with this idea that invention is just like the first person to realize, to see, to find something that is recognizable as beer in a “Theory of Forms” sense.

This is why, because it assumes then that you can have invention without humans. Which I guess AI would think is possible. Because they're trying to create a world without humans, right?

A post-human world or inhuman world.

We could give ChatGPT the benefit of the doubt. Sure, something that always existed cannot be invented. Okay, that's step one. However, that's kind of not what's being asked. Because something can always exist, but we never have access to it. And so someone has to bring some kind of copy of it into the world. And that's what we mean by inventing in English.

Does Plato ever explain how invention works with the Theory of Forms? I mean… the Theory of Forms is not a big part of Platonic philosophy, right?

No, he's not interested in invention.

I mean, he might be interested in inventors and people who invent and craftsmen.

I think maybe what they're doing is mixing up how he understands learning. Because learning is remembering the forms. So if inventing is a form of learning…

All right. So he was a wise man who first remembered that beer was a thing.

Yeah, but in terms of, like bringing a beer to the table, making it possible that somebody could drink it.

To be fair. That was GPT-4, and I asked GPT-4o today and it responded much more succinctly:

No, Plato did NOT say, ‘He was a wise man who invented beer.’ That quote is widely misattributed to him. There is no record of it in any of Plato's dialogues or writings.

I like that it distinguished there between dialogues and writings. Like, okay, sure, ChatGPT. It continues,

The quote is more of a modern, humorous saying that people have falsely credited to various historical figures like Plato or Benjamin Franklin. So while Plato had plenty to say about wisdom and moderation, praising beer wasn't one of them.

Not a very grammatical sentence as I look at it. So the bigger problem with this, of course, is that there is no record of beer being widely drunk in Greece.

But in Egypt, right?

I mean, it had been invented in Egypt and Thrace. There's a reference to drinking something that sounds like extremely strong beer in Xenophon, at Anabasis Book 4.5.24-36.

Okay, so… what does he say here? So he says, “there was also meat, barley and beans, and barley wine in large bowls. Floating on the top of this drink were the barley grains. And in it were straws, some larger and some smaller, without joints. And when one was thirsty, he had to take these straws into his mouth and suck. It was an extremely strong drink unless one diluted it with water. And extremely good once one got used to it.” This does sound like beer. It probably is.

It sounds sort of like a Belgian tripel ale. One of those really strong beers. I do like how he describes that the straw is not jointed, which immediately made me think of those straws that you get with the bendy thing at the top. Or some kind of like fantastical twisty straw. Just Xenophon drinking beer out of a krater with like one of those crazy straws.

unlike understanding the Theory of Forms, creating this image was right in ChatGPT’s wheelhouse

Well, it might be because you're not filtering it, right? So the barley and stuff is at the top.

Exactly. So the straw helps you get the stuff that's at the bottom, so you don't get a mouthful of barley.

I'm trying to see where the hell they are in this.

Great question. They are in Armenia. They're making some progress, but they've got a ways ahead. Miles to go before they sleep.

This is very cool though.

But it does seem like the degree of explanation required here suggests that it was not something that people would be familiar with.

Exactly. Precisely the point is that they didn't drink beer. They drank wine. These other folks over here in this wheat culture.

And Xenophon and Plato, you're really talking very, very close to contemporaneous.

They're contemporary. So much so that they're probably in the school together with Socrates. Even if Plato could have said something about drinking something, it wouldn't have been that. It wouldn't have been beer because he didn't know what it was, because Xenophon is surprised about this interesting drink.

The bar might as well have said, “‘In vino veritas.’ -Plato.” Relatively speaking, that would be as accurate. In Latin. At least we know Plato drank wine, unlike beer.

But Xenophon, we might be able to give him that credit. Probably after having that first drink he could have said, “hmm, the man who invented beer was very wise.”

There you go. He also gave really detailed descriptions for how to use a straw in case that is a thing that you need. At any rate, ChatGPT did a thoroughly mediocre job assessing this.

Right. But folks that are working on digital translation and all of that stuff, working on using AI in Homeric reconstruction of papyri and things like that, were actually talking about how they can get the AI to generate a reasonable Homeric line. And that was a year ago.

I'm sure it's only gotten better. It does raise a question of Lucian's Homer in the underworld. So, for context for our readers, in the Dialogues of the Dead Lucian goes down to the underworld, and there he meets Homer. And he tries to ask him about some controversies in Homeric textual criticism, because scholars have been athetizing lines, claiming that some of them weren’t authentic Homeric lines but instead later interpolations. But when he gets to the underworld, Homer says, “No, I wrote all the lines, even the athetized ones.”

So my question is, did he write all the lines, including the ones that were made up by ChatGPT?

That’s a great question. I'm actually using that passage in my Apollonius book to point out that this was controversial. The new Lucian would say, “Did you write those?” “I totally wrote that. I wrote them from here in the underworld.”

Exactly. “I am the Homer bot.” Incredible. So, you're at the Center for Hellenic Studies. You don't have any students right now, right? So you haven't been getting just piles of essays that were entirely written by ChatGPT recently?

Not yet, no. I expect next semester I'm going to have to come up with a strategy where that isn't the case. What I have been doing is getting them to write the essay in class. I wouldn't have any problem with them making ChatGPT write the essay and then memorizing the essay and then spitting it out in class. It's okay because at least they've done some brain work there. They sat there and memorized the essay. I'm okay with that.

It's really interesting that you say that, because my first Latin teacher used to tell us, “just go home, translate this passage from Cicero,” and then nobody would do it. And he would translate it in class, and we write down his translation and we would memorize his translation and we would regurgitate it on the test. So it was kind of like what you're talking about.

Only that they would be getting it to write the essay, memorizing the essay and then spitting it back. But at least there's some mental energy being spent on memorizing.

And then writing it by hand. When you have to write something by hand, you learn a lot.

So that's my way of thinking about it. Like, okay, fine. If you're going to do it, do it that way.

Yeah, although… let’s go back to the Homer bot for a moment. It does feel like you should be able to make like a GPT Homer. Because essentially what you're talking about is… well, one of the things classicists romanticize about, as you know, is “O for the lost 19th century classicists who had memorized all of Homer.”

The golden age.

“There was a time when classicists were much better than we are.” And obviously, you know, decline and fall narratives are part of how classics is taught. It’s baked right into Homer, with Nestor complaining that heroes used to be so much stronger than they are in the Trojan War era. But it is true to a certain degree, classicists used to know Greek and Latin much better than we do now.

They could definitely spit out all of Homer, but they couldn't think creatively though. There's some losses and some gains.

I mean, I think that our version is much more interesting in a lot of ways. But it seems like with our technology, it should be relatively easy to create a bot that's as familiar with Homer as the best 19th century Homeric scholar and is so steeped in Homer and nothing else that it could fill in those papyri. You know, there are LLMs that are really trained on just the bible. Really specialized ones.

We should try that. We should try to see it.

I think that there are a lot of potential uses for that kind of tool and unfortunately, it feels like most people in academia when they think about AI go to sort of an existential dread place.

Oh, yeah. Well, mainly because, well, there's a loss of some kind of mental faculty that's happening, right? Because the students aren’t thinking for themselves. In the same way that back in the day, people could memorize all their friends’ phone numbers, and nobody can do that today. I mean, that may not have been a great faculty, but the point is that you've lost it.

Which brings us back to Plato, right? Because Plato wrote about this concern that writing as a technology was going to hurt people's memories.

So that's the first real worry about it, and like the ancestor to this particular technology. You have the technology and then you immediately have the worry of but… we're losing something. And indeed, we are.

I'm sure that when Plato learned how to write he stopped remembering all of his friend's phone numbers.

Yeah. Once you get a smart tablet. Yeah, it's kind of interesting. He's also using it, right? Because he is obviously writing.

He's writing quite a lot. He's writing a huge amount.

What's interesting is that you wonder to what extent they are offloading a lot. Because, well, who's doing the real writing? Is it really him, or is it an enslaved scribe that's doing that? Certainly the people reading it out loud are enslaved readers. We know that. There is a kind of, a using of enslaved human beings as your tools. As your smartphones, as it were. That's going on at that time.

One other interesting thing that nobody's really looking at is the fact that a lot of the searching for things and reading the text and writing the text is actually being done by the slave. And so there's an offloading of that kind of thinking onto the slave.

So the slaves were really elite knowledge workers in a way. They were the search engine.

I'm convinced of that because I can't see these guys doing it. They're too busy thinking.

Yeah, exactly. Drinking beer with crazy straws.

But given the glimpses we get of this… say you’re Callimachus, how would you run the Alexandrian library and write 800 books? Really? Did he really? Or did he have a bunch of slaves running around getting the data for him, writing the stuff down, which I think is the case. We could think about it that way where it's like, there's always scenarios where not all of your actual production or creation is actually done by Michelangelo, another example where half of the stuff is being produced by his assistants.

His workshop.

But he gets all the credit. So the new thing is that this thing we are co-producing with is not a human.

Right. Although there is a lot of unwaged human labor that goes into creating it.

Oh, yeah. That's the other thing that people are even more upset about than worrying about what's happening to our brains.

The theft of material.

The folks who are profiting from this have stolen work from all of us, academics and creators. Nobody's being paid for that. You couldn't have this thing if you didn't steal half of the information. But I think it all goes back to this idea of how we value, how we look at work, which has completely been a slavery model. Every kind of work is always in this slavery model and somehow we can't think outside that box.

Exactly. I feel a little uncomfortable equating slavery with theft of intellectual property, but there is that unwaged labor similarity, right? When you steal somebody's labor. Then that is sort of like a form of slavery.

Well, yeah, if you define slavery in terms of unpaid labor. And in that case, I'm free to go wherever I want. They're just taking my labor. Whereas the physical body is not necessarily involved. But the point, though, is that there's still a violation happening to my humanity, your humanity, everybody's humanity that is contributing to this unwillingly. They did not consent to that.

---

Jackie and I had that conversation in early April, and since then the conversation around students using AI has only gotten grimmer. But I hope that there are possibilities here to make students think about AI, and the kinds of knowledge it’s replacing — and also the long, rich history of the outsourcing of knowledge work onto unpaid labor, which has been happening for literally thousands of years.

If you’re an academic who has a creative approach for getting students to actually, you know, do work instead of just using ChatGPT to write all of their homework for them, please share in the comments! We could all use the advice.

One thing is for certain: that bar should rethink its neon sign. I wonder how fast I could get myself kicked out if I tried to give them an impromptu lecture on the Theory of Forms?