A song to read by: “Jump Into the Fire,” by Harry Nilsson

What I’m reading: “The Viking Heart,” by Arthur Herman 

The week that was

Last week was a relatively sleepy set of days, which offered a welcome reset from the last two enjoyably taxing months. 

The highlight of the week was, by far, the New York Marathon today, as several of my best friends from Austin came into the city to support another friend of ours, Hannah, as she ran the race. 

The opportunity to spend some time, in particular, with my friends John and Winston was a welcome one, and it gave us an excuse to stop by the newly opened Wythe brick and mortar in the Lower East Side. That, and watching the University of Texas Longhorns pull an ugly victory from the jaws of a near (self-inflicted) defeat, were the clear peaks of the week.

Also, I saw “Killers of the Flower Moon” Friday, and despite eagerly looking forward to the film for months, I was unimpressed and honestly quite bored. Am I alone on this?

On my mind

This week I spent a lot of time thinking about generative artificial intelligence for a magazine story I am working on.

GenAI technology seemed to burst out of the gate at the beginning of the year, but its momentum has certainly slowed — whether that is intentional or not is another story. But for the first six months of the year, I and many others in the industry legitimately wondered if GenAI might be soon putting us out of our jobs.

I covered its slow creep into the media industry, charting and then extrapolating the ways in which it might transform the landscape. For a while, I had the morbid feeling that I was chronicling the rise of the very technology that would soon replace me. 

At the time, one publisher told me enthusiastically that they were experimenting with a GenAI tool that ingested its writers’ work, then could produce articles in their distinct styles. What, I wondered, did the writers think about that? At times it has felt like we were being handed a shovel and being told we had been given the opportunity to dig our own graves.

But in the course of reporting out this latest story, the mood seems to have shifted. Speaking with copyright lawyers and academics threw some cold water on my mounting hysteria. 

“This is just the latest iteration of an ongoing tug-of-war between content creators and content distributors,” one attorney told me. Law, which is grounded in precedent, offers the same kind of cool comfort history does — that there is nothing new under the sun. While the details may change, the themes remain the same.

It also dawned on me just how reliant GenAI is on the datasets it ingests. It is unable to offer cogent output on the events of tomorrow until after it has ingested the data of tomorrow. It can extrapolate, with relative precision, based on past patterns of behavior, but that is not novel, and it is certainly not foolproof.

In essence, the technology is just a new way to index preexisting information. Much like its search product predecessors, it cannot think or legitimately create; it can only comb through and recontextualize what it already knows, and everything that it knows it learned by ingesting information produced by someone else.

GenAI will certainly keep evolving, but I have far fewer concerns about its scope than I did just a few months ago. Like so much modern technology, it is a tool that makes sense of information. But with no real ability to generate the information itself, it will necessarily rely on the people who do.

Published this week

— Publishers have long used ad networks to boost their putative reach, but the publishers that make up those networks have typically been on the wrong side of an asymmetrical power dynamic — especially when it comes to multicultural publishers. For the Adweek magazine, I wrote about a handful of new ad networks, all formed since 2020, that are reimagining the collective model. (Link)

— Part of a larger effort of mine to follow-up on my own reporting, I looked at The Verge and its redesigned website one year later to see how it fared. Editor in chief Nilay Patel famously told me that the redesign resembled Twitter because he anticipated, last year, that the platform might soon begin to shed users. How right he was! (Link)

— Finally, a juicy scoop about the private equity-backed media company Recurrent Ventures. After a blockbuster entrée into the media world, the publisher has run into a host of challenges, many of which I have detailed previously. Now, it has named its third CEO in three years. (Link) 

One good rumor

Did Vice quietly kill their print magazine product? I heard a tip recently that they were bringing it back, which made me realize it was gone. Anyone with any more details on this, feel free to respond to this email!

Some good readin’

— The amazing story of how “The Texas Chainsaw Massacre” movie was made. (Texas Monthly)

— Great scoop from Max here. If your new ownership is scaring away sources, that might be a sign you went wrong somewhere along the way. (Semafor)

— I am very conflicted over farmed salmon! In America, the solution can never be “buy less.” It is always has to be “buy something different.” (The New York Times)

— If you have to read anything, read this! A great piece exploring how, empirically, Covid lockdowns were a failed experiment. (New York Magazine)

— Finally, I love Tim Heidecker. His “no more bullshit” character is one of the funniest things I have ever seen. (New York Magazine)

Cover image: "Consecration of the Emperor Napoleon I and Coronation of the Empress Josephine,” by Jacques-Louis David