The hardest calling card there is to fake

I read a lot of business books and LinkedIn posts now that were written by ChatGPT, and I can usually tell. They’re always competent and usually lifeless, with the same three-part structure, the same tidy takeaways, and replete with a voice that belongs to no one. AI made writing mindless, and in doing so it made a real human voice the rarest thing in the feed.

That is, partly, why a good book still matters, and why it’s worth more now, not less. A book is the hardest calling card there is to fake. Anyone can generate a thread in ten seconds. Almost no one can produce two hundred pages that are worth the read.

A book pays off beyond the sales figures. Most don’t sell many copies, and that’s fine. The return is credibility, and the authority that comes from having actually written the thing. A book is a door. It gets a founder onto the stage, into the interview, and into rooms that were closed the week before. As Groucho Marx wrote in his memoir, he could have written a classic if he’d wanted to; he’d just “rather write for the little people.”

Two things usually go wrong on the way there. Most founders decide they’ll just write it themselves, all three hundred pages, on weekends, between board meetings. Chapter one goes great. Chapter one always goes great. A year later there’s still only chapter one, and they’ve quietly stopped bringing it up. The ones who don’t stall hand themselves to the industry that’s grown up around all this instead, which is its own kind of trouble: hybrid publishers charging five figures for what an author could do alone, ghostwriting mills that take the money and vanish somewhere around delivery, and now AI that hands back a draft in a voice that isn’t yours, but is rather an amalgam of all the stolen intellectual property from pulped and scanned books the machines were trained on. Rebecca Ackermann described that last one in The Atlantic: a client who swapped their ghostwriter for AI got a draft with a cruel edge where they’d meant to be funny. “I’ve never heard you talk about people this way,” the writer told them.

The good version of this is old. In 1965, Malcolm X told his life to Alex Haley, and the book became one of the most important American memoirs of the century. Andre Agassi’s Open, Phil Knight’s Shoe Dog, and Prince Harry’s Spare were all written by J.R. Moehringer, a Pulitzer winner who insists his name stay off the cover; “the midwife doesn’t go home with the baby,” as he puts it. That’s the craft, and it runs on a strange discipline: the writer disappears so the author can stand there alone. The mills are the problem: they cash your deposit and mail back a manuscript nobody who knows you would believe for a single page.

What’s changed is the volume. When the investor Stanley Druckenmiller published a Wall Street Journal op-ed this year, readers caught the machine behind it within minutes: the giveaway “not this, but that” constructions, the cadence and tropes you’ve started to notice everywhere, and the paper shrugged and stood by it. None of the underlying impulse is new; I’ve been writing about it for years: the humblebrag economy and the phones that were always listening, the long work of performing a self. The counterfeit voice has been around forever. AI made a convincing one; it cost nothing. That’s exactly why the real one is worth more. A book written entirely by a machine can’t even be copyrighted; The Copyright Office won’t register work with no human behind it, so the one that carries your voice also happens to be the only one the law lets you own.

Doing it right is unglamorous. It starts with a recorded conversation, week after week, and a writer working from the transcripts to get the person’s story and voice onto the page, then going back over it until it reads the way they actually think and talk. I use every tool there is, AI included, and none of them do the part that matters. A transcript won’t write your book, and neither will a model; it takes a writer who’s spent years learning the difference between how a person talks and how they think they talk––one who isn’t afraid of throwing in an em dash or two when it fits.