Reading, Writing, no Rithmetic
Some weeks back one of my hiking buddies sent me and some others this article: Who is Blake Whiting?
It’s from something called American Scholar, of which I was previously unaware. You can click the link above and read it all yourself, but the gist is this: Blake Whiting is a ‘historian’, who, as the article says –
‘In one week alone last fall, he published 13 books on a host of complex archaeological and historical subjects, ranging from the collapse of Near Eastern civilizations in 1177 BCE to the recent discovery of a huge Silk Road–era city in Central Asia.’
To get right to the punch line, no human can do that, and indeed, BW is a LLM. Moreover, it is an LLM which produces books that are pretty straightforward rip-offs of the work of actual people, and then are sold on Amazon. What the LLM appears to do is appropriate but re-word the stolen material, never using quotation marks but also never citing the stolen work. Plagiarism, by any other name.
The books are not copyrighted, because copyright belongs to a person, and there is no Blake Whiting. Amazon was contacted by the author of the piece, and here’s a quote from the Big A’s spokeswoman:
“We have content guidelines governing which books can be listed for sale and remove books that do not adhere to those guidelines, whether AI–generated or not,” Amazon spokesperson Jennie Bryant said. She declined to comment on Whiting’s works.
Of course she did. What could she say? I mean, the fact that this ‘person’ published 13 books in a week, that didn’t violate any of Amazon’s guidelines, apparently.
So, LLMs are now being used to rip off actual authors to produce fundamentally plagiarized books for sale at a speed that no actual human researcher can match. And Amazon has no hesitation in selling these rip-offs at a profit.
As another quote from the AS article puts it:
If an AI program accesses your just-completed dissertation and salts it with data and text from other sources, then that book you planned to write for a general audience, based on years of research, might be available online before you can get your proposal to a potential publisher.
But wait, there is more good news about LLMs and publishing, this time from my own personal experience.
I read a lot of history, and one such book I read recently was titled ‘The Last Days of the Ottoman Empire’, by Ryan Gingeras, published in 2022. The Ottoman Empire lasted for hundreds of years, and at its height included the Christian and Muslim Holy Lands, Greece and the Balkans in Europe, a large chunk of North Africa and Egypt. It was centred in what is now Turkey, and by the time of the First World War, it had been in decline for some time. The book details those last years, from WWI into the 1920s, when it finally broke apart completely, leading ultimately to the formation of the modern state of Turkey. The Ottomans made the (perhaps understandable) mistake of siding with Germany and Austro-Hungary in WWI, so things fell apart fast from 1918 on.
It’s a fascinating read. The many languages, cultures and ethnicities involved, with all the conflicting motivations those engendered, make it a highly complex narrative to follow, but Gingras does a pretty good job, and for an academic, he is a decent writer.
Consider this passage, detailing what happened in Istanbul in 1920, after a French military force was forced to abandon its garrison at Marash by a Turkish military force fighting to create an independent Turkish state.
Three days later, British and French troops fanned out across Istanbul. By the afternoon, leading members of The Chamber of Deputies were placed under arrest. With the aid of British troops, many of them Muslims from South Asia, key ministries were stormed and occupied. In the days preceding the takeover, the Allies had given no public hint of this coup. Yet in moving against each of the principal arms of government, British and French forces brazenly asserted that Istanbul now fell under their authority…In making their presence felt, British officials arrested eleven of the most outspoken Nationalists in the capital, including Rauf Orban and Kara Vasif, co-founder of the Karakol organization. All were then deported to Malta for internment.
The Empire had been ruled by a religious figure, the Sultan. Gingeras further writes:
How the Sultan greeted the occupation of his capital is not clear. Mehmed VI, according to British intelligence sources, was delighted at the apprehension of so many prominent Nationalists. The Allied seizure of power likely saved the Sultan from attempting ‘a similar “soft” coup with inferior forces and a considerable risk of failure’.
Even in this one episode we see many players, and complicated motives and moves. But it is clear from the passages quoted what happened here, right? The French defeat at Marash induced the British and French to respond by staging a coup in the capital of Istanbul, arresting Deputies, taking over Ministries and deporting Nationalist leaders. The Sultan, wanting to maintain his own position, is perhaps not too unhappy at having these Nationalist leaders out of the way. The Nationalist movement was a secular one, after all, and not likely to keep the Sultan in charge if they prevailed.
Ok, decent book, not up to the standard of John Keegan or even Victor Davis Hanson, but a good read, on a topic of which I was largely innocent. It taught me stuff, and there is no higher praise than that.
Some while after finishing this book, I saw a review of a book in the TLS called Mafia: A Global History. The review is written by one Ian Thomson, and although I did not notice it at the time, the author of this book is none other than Ryan Gingeras. This book got a pretty good review from Thomson, his last words being: ‘….Mafia is an excellent history of all things murderous and mobster.’ His one complaint was about some ‘clunky generalizations’ in the book.
Now, I pay little attention to reviews, but I bought a copy of this very recent (2026) book because the topic seemed fascinating. It was not until I started on it that I actually noticed the author was Gingeras. Bonus, I thought – this should be good.
Wrong. The writing is miserable. It resembles the kind of writing I used to get from a mediocre university student whose idea of doing research is to look up a bunch of hopefully relevant things in Wikipedia, then string them into a barely cohesive list of sentences.
Here is Gingeras writing about San Francisco:
Since its heyday as a mining hub, the town had developed a reputation for crime, corruption, and loose morals. Single men seeking to strike it rich in the nearby Sierras gave rise to hundreds of gambling establishments in the city. Local card halls and betting parlors attracted gamblers as well as legions of thieves and ruffians. Casual violence was commonplace at the many saloons and dance halls of the city’s core. By the 1870s, several neighborhoods fell under the influence of gangs comprising young miscreants and petty crooks. San Francisco’s prostitutes attracted the greatest attention.
Um….attracted the greatest attention from who? The press? The city government? The public? Those gangs of miscreants and crooks? And I know what a thief is, someone who steals things, but just what the heck is a ‘ruffian’?
And, by the way, the first sentence clearly should read ‘During its heyday as a mining hub….’
Then comes this sentence, sitting all alone and friendless in the middle of a paragraph:
Xenophobia and other anxieties, such as widespread alcoholism and opium use, colored local impressions of those involved in “trade”.
Why did xenophobia color anything? Were most of the prostitutes from foreign countries? Nothing is said about that. What is meant by calling alcoholism and opium use ‘anxieties’? Is he suggesting they were not really all that widespread, yet people were anxious about them? Hell, could he even be bothered to tell us how widespread they were in San Francisco in those days?
From slightly later in the book, here is a passage, which seems like it ought to be quite central to the book’s thesis. In 1909 a fellow named Theodore A Bingham wrote about criminal organizations in New York City, and he stressed that these were not just a disconnected bunch of criminals, but were ‘thoroughly organized’.
Here is what Gingeras writes about this:
Bingham’s conclusions correspond to a critical moment in the formation of modern-day mafias. Since the beginning of modern Western efforts to control or eradicate vice, many presumed that the forces of law could prevent the masses from indulging in prostitution and gambling. By the early twentieth century, it appeared in many places, first perhaps in the United States, that this was wrong. The state itself was part of the problem. But why petty officials and local policemen allowed vice to go unchecked spoke to something more than political corruption. As “organized crimes”, illicit gambling and prostitution drew upon the support and complicity of a range of conspirators, including representatives of the state. The development of the white slave trade showed that the laws of supply and demand were making prostitution an increasingly globalized trade. What appeared to drive vice was not simply venality, but the profits it could generate. Vice had become an industry.
That paragraph is a string of ideas with no attempt to develop any connection between them. Further, the paragraph has no apparent connection to Bingham’s central point – crime in NYC is ‘organized’. One might expect a discussion of what Bingham meant by that term, and what were the facts that at the time suggested he was right – or maybe wrong. Instead we get this disjointed paragraph.
Here are the separate ideas laid out in the paragraph –
- Something changed in the early 20th century regarding the control of ‘vice’, first in the US.
- Officials of the government could be bribed to look the other way when crimes were committed.
- supply and demand globalized prostitution.
- vice was driven by both venality and profit.
Let me start with 3. First, is it the case that prostitution became globalized in the early 20th century? The only thing I can imagine that meaning is that there were women from country A practicing prostitution in country B. Was that true? If so, did that really start in the early 20th century? Gingeras seems uninterested in offering up any evidence or explanation for this claim.
As to 4, does Gingeras really wish to assert that vice being driven by the profits it could generate first became apparent in the early 20th century? You can say the demand for prostitution and gambling is driven by venality if you wish, but surely since the beginning of time, their supply was motivated by profit. There are not and never have been any non-profit brothels or gambling houses.
Similarly, regarding 2, I am willing to bet a lot that officials have been looking the other way during the commission of crimes since money was invented. No, actually, since before that.
There is a connection one could make between vice being profitable and officials being willing to overlook it, but Gingeras cannot be bothered to be explicit about it, so I will take up the task.
If vice is profitable, then it is also surely profitable for criminals to bribe officials to leave your operation alone, since that increases the likelihood that you can go on making those profits, rather than ending up in jail. The two things are connected, indeed. If Gingeras realized the logical link, he did not bother to lay it out for his readers.
Again, a paragraph like the one quoted above I would have categorized as Wikipedia writing back in the day. A string of assertions, with no attempt to provide a coherent logical argument.
But this is 2026, and although I ran into no LLM writing back in my days as an instructor, I have seen it since. My strong suspicion is that Gingeras used an AI chatbot to do much of the writing of this Mafia book. I can see no other explanation for the steep drop in the quality of writing since he wrote the Ottoman book, as that was back when chatbots were neither very good nor very available. A key reason for my believing this is actually my first remark above, noting that Gingeras wrote ‘since’ when logic indicates he should have written ‘during’.
It is a hallmark of LLM writing to get it just ‘a little wrong’ in that way. To use words that are ‘close’ to correct, but not really. The book is bristling with that kind of prose.
If I’m right, then the Blake Whiting example shows that he could do even better if he just got out of the way and let the chatbot do all the work. 13 books in a week. Way more impressive than two in four years.
Now I know almost nothing about the TLS reviewer of this book, Ian Thomson, but he is supposed to be an author himself. Given that, I also suspect that he had a chatbot write his review. How anyone could read the entire Mafia book and then write that the book is ‘excellent’ is beyond me. I quit reading maybe halfway through. The continuing stream of barely connected and not-quite-right sentences felt like someone was throwing warm oatmeal at my face, and I was not learning a damned thing.
Postscript: Just before I posted this, some days after writing it, I had another thought. Perhaps another explanation for the drop in quality in Gingeras’s writing from 2022 to 2026 is that in the early 2020s, his work was edited by someone at his publishing house. By 2025 or so, Avid Reader Press was not willing to pay for a (human) editor, so Gingeras was on his own, or, worse – he or his publisher employed an LLM editor.