More than four months ago I published Sorry, But Your Ex Is Probably Not a Narcissist: Essays on Depth, Complexity, and Human Transformation, gave away dozens of copies, and asked that recipients write an honest review in return. So far, I’ve gotten one on Amazon.
Oh, I did get another on Goodreads, but that was written by someone trying to butter me up to hire them as a publicist, I doubt they even read anything more than the description.
I know, you might have gotten the book, but haven’t read it yet. Or you did and haven’t gotten around to writing a review. Or you started to, but didn’t know what to say. I get it. I don’t blame you. I’m just curious about what the reviews would say. I’m especially curious about what a book review from the New York Times would say. Not that I expect them to review it. My book is not from a major publisher with a strong publicity campaign, so the Gray Lady will never even hear of it.
It was then that I saw something another writer did on his Substack and understood something I hadn’t understood before. Bear with me, because it appears to be irrelevant. I’ve wondered why folks use AI when they could ask real people for help. Most don’t want to use AI. They’re nervous about it taking over the world, or at least their jobs. They dislike the way everything is turning into data centers. But people use AI anyway for one simple, good reason. Because real people are not there for them when they need it.
That’s what’s happening in my field, psychotherapy. Folks are turning their favorite Large Language Model into a therapist. We therapists say don’t do that, it would be better if you talked to us. But folks know that already. They know AI can never be as good as a real therapist, except for one thing. It’s there when they need it. Being there trumps everything else. You don’t need to be great or even competent if you can be there. Often, just being there is enough. It’s impossible for a therapist to always be there when you need them, so it’s a good thing we have AI to fill the gap.
So, anyway, this is where I’m going with this. When Dan Williams posted an AI generated review of his own book on his Substack, Conspicuous Cognition, I thought I could do the same. Since I’m so curious about what a review of my book would look like if I saw it in the New York Times, I decided to ask my favorite AI, Claude, to review it, as if it came from the New York Times.
I chose their reviews because I’m a regular reader and find them opinionated, contextual, and willing to be harsh. There’s often an attempt to explain why a general, educated reader should care about the book. I opened an incognito chat with Claude, so it wouldn’t know the request came from me, selected Opus 5, designed for complex tasks, uploaded the manuscript, and used this prompt: “You are a well-read and well-regarded social commentator. Please write an insightful review of this recent book. The review should be suitable for publication in the book review section of the New York Times and conform to their style. It should be no longer than 2,000 words.”
This is what it said:
SORRY, BUT YOUR EX IS PROBABLY NOT A NARCISSIST: Essays on Depth, Complexity, and Human Transformation, by Keith R Wilson.
The word narcissist has had a strange career. Coined for a Greek boy who drowned admiring himself, absorbed into psychoanalysis, codified in a diagnostic manual, it now functions mainly as a term of art in breakup narration. Somewhere between the clinic and the comment section, it stopped describing a rare and serious personality disorder and started describing a man who didn’t text back.
Keith R Wilson, a psychotherapist in Rochester, N.Y., with roughly four decades in practice, has written a book whose title promises to take that word away from you. It does, in about four pages. Then it goes somewhere considerably more interesting, which is both the pleasure of this collection and its central marketing problem.
The title essay is among the shortest in the book, and its argument is quick: the self-help content that spreads is not the content that asks you to examine your own conduct. It’s the content that hands you a verdict about someone else. Wilson attributes this to what he calls, borrowing from Gadamer, the hermeneutic horizon — the boundary beyond which a person cannot absorb what they’re being told. Left alone with a feed, we share only what falls comfortably inside that boundary. A therapist, by contrast, can aim just past its edge, because he can watch your face while he does it. The ex is not a narcissist, Wilson concludes. He has narcissistic tendencies, as you do, as Wilson does. The question was never whether but how much.
That’s a sound observation, briskly made, and if the book were only that it would be a pamphlet. What Wilson has actually assembled — 35 essays drawn from his Substack, The Reflective Eclectic, and arranged into six thematic parts — is something closer to a working therapist’s apologia for his own vanishing profession, delivered in the voice of a man who used to raise pigs.
He did, in fact, used to raise pigs. The strongest essay in the collection compares the family farm to private practice. Wilson farmed in Bath, N.Y., butchering his own livestock and fixing his own tractor, and watched agriculture consolidate into something that optimizes a single measurable variable — yield per acre — while ignoring the taste of the food, the pollution, and the people displaced. He then spent twenty years in a community clinic absorbed by a corporation, where his caseload went from 25 to 225 and the mandate changed from letting people talk to reducing symptoms and returning them to work. When the corporation instructed him to complete his paperwork during sessions, he left. He now watches his own physician type into a screen while the software recommends the prescription. It is the freshest argument in the book, and Wilson has the rare standing to make it, having been ruined by both industries.
This is his method throughout: a homely, concrete image doing the work that jargon would otherwise do. Schemas are outdated maps. The unconscious is a jazz band, and the ego is the front man who thinks the show is about him. When the ego gets pelted with tomatoes and flees the stage, the band plays on, which is what a complex is. Wilson calls the resulting integration “getting in the groove,” and cheerfully admits he has renamed individuation to smuggle Jung past readers who would never pick up Jung. His name for improvised, jury-rigged practice — borrowed from Lévi-Strauss — is bricolage, and his emblem for it is the set of jumper cables he once wrapped around his car to hold up a dragging muffler, to his daughter’s lasting humiliation.
The best essays are the ones in which the instrument is turned on himself. Early in his career, running a group for people recovering from serious mental illness, Wilson listened to a woman describe how she had once believed a picture falling off her wall was a message from God. As she spoke, a picture fell off the wall behind her. Everyone turned to the therapist. He reached for whatever authority he could find and pronounced it a coincidence, adding that reading personal messages into coincidence is a delusion of reference. Nobody in the room wanted to be delusional, so the subject closed. Wilson’s verdict on his younger self is that he was the only one in the room who went crazy that day, because he adopted a tone of certainty he did not possess. From this he derives the thesis that organizes the collection: the Establishment is not merely governments and corporations but “any voice that hands out answers,” including the quiz that assigns you an attachment style and the video that diagnoses your ex.
Elsewhere he recounts typing back and forth with a deaf client while proudly signing what he believed meant I understand. It meant I’m horny. He had also missed her schizophrenia, having assumed that a voice in her head was the ordinary figure of speech, not knowing that deaf people hallucinate voices too. The essay becomes a lucid short course in Gadamer: prejudice is not a defect to be scrubbed away but the scaffolding that makes understanding possible, and the honest move is to declare yours and invite correction. In a third, he confesses to a dinner-table standoff with his four-year-old daughter over peas, which ended with the child asleep in her chair and the dog eating the evidence. Taking the episode apart, he finds no nutrition in it at all — only his mother’s authority, a general faith in discipline, and a fear of raising a child who would embarrass him at dinner parties.
Wilson is also, when he wants to be, a genuinely useful clinical writer. His essay on venting assembles the unwelcome research — that catharsis tends to feed anger rather than discharge it, that debriefing after trauma predicts worse outcomes — and then makes the sharper psychological point, which is that compulsive venting is itself a form of repression, since the story you tell a hundred times consumes all the airtime the untellable things needed. His treatment of attachment theory is a model of proportion: the Strange Situation findings are real and important, the extension to adult romance is an analogy rather than a finding, one longitudinal study puts the infancy-to-adulthood carryover at about 17 percent, and the labels most often function as a cage, letting a woman pathologize her own accurate perception that her partner is unreliable. And a small, superb essay on forest bathing works its way to the conclusion that the active ingredient may be the exotic name — followed immediately by the observation that a placebo is expensive, because it is paid for out of the public’s trust in science.
The book has real faults, and they cluster in a revealing place.
Wilson’s argument against typologies is relentless — the DSM is dispatched in a fable as “one of the greatest works of science fiction,” diagnosis is an illusion of control, labels close inquiry rather than open it. Then, in a chapter called “Nine Kinds of Madness,” he presents nine strategies people grab to keep from falling into the abyss: perfectionism, helpfulness, achievement, individuality, learning, loyalty, enthusiasm, productivity, peace. That is the Enneagram, in order, and Wilson names neither it nor any source. His chapter notes, otherwise scrupulous, skip the chapter entirely. A writer this alert to borrowed authority ought to have caught himself.
Worse is CRUDD — Certainty-Based Unshakeable Reality Distortion Disorder — a diagnosis Wilson proposes for people who hold beliefs with a rigidity that survives all contrary evidence. The impulse is understandable and the underlying observation is fair. But this is a man who has just spent five chapters explaining that inventing categories for ordinary human trouble is how psychology sold itself to the pharmaceutical industry, and he responds by drafting an entry for the manual. He compounds it by specifying that the point of the diagnosis is not to tell the patient but to warn the people around him, which is pathologizing your opponent while he isn’t in the room. Three chapters earlier, that was the thing being condemned.
There are smaller lapses. A taxonomy of “reflective functioning” — otherwise the book’s most original piece of synthesis, with its explorers and mapmakers, its label droppers and firehoses — files autistic people at the bottom of the scale in a paragraph that flattens them exactly as Wilson has warned against. An extended set piece about a couple’s household conflict rests on masculine and feminine “functions” in a way the hedging at the end does not quite rescue. And the Substack provenance shows: the hermeneutic horizon is introduced three separate times, Hillman recurs so often he becomes a co-author, and the preface’s insistence that a coherent system would only produce another Establishment reads, by the last third, as a license for repetition.
Which points to what this book really is. Wilson presents himself as a revolutionary against therapy culture, and the title sells that. But he is not against therapy; he is against cheap therapy — manualized, insurance-shaped, fifty-minute-symptom-reduction therapy — and against the free version people now assemble from short videos. What he wants is more expensive, slower, and older: depth psychology, insight, the long relationship, the why rather than the how. His closing essays make this plain, from the portrait of Philippe Pinel unchaining the inmates at Bicêtre in 1793 to the demonstration that therapy plundered religion for its techniques and is now being plundered by chatbots in precisely the same way, with precisely the same objections. That last essay ends by handing the verdict to the reader, which is the most consistent thing in the book.
Wilson’s finest line is that uncertainty is a sign of health, not weakness. The collection is most alive wherever he honors it: at the fallen picture, the peas, the wrong sign made with his hand. It falters exactly where he forgets — where the man who distrusts every settled answer turns out to have a few of his own in the drawer. He would be the first to say that this is what happens to revolutions. He might not have expected it to happen quite so soon.
My reaction?
That Claude guy is really, really smart and spotted things my copy editor and I missed, like having no references for one chapter, “Nine Kinds of Madness”. Not that Claude needed them, it already knew I was writing about the Enneagram. Good editing would have also caught the repetitions I kept when I converted a collection of essays into a book. Next time I publish a book, I’ll run it by Claude.
It’s right with its criticism, too. In the book, I am willing to create taxonomies that I warn you against, although he must have missed my saying that people can’t help it. In my own defense, my creation of diagnostic categories demonstrates how easy it is to do it and how little rigor is involved.
All the usual AI tics are evident in the review. Em dashes abound; no semi-colons in sight. “Not X, but Y” appears in nearly every paragraph. And every paragraph ends in a short, hard unsupported statement, so frequent that you expect it. A human writer would switch it up, only because the monotony would bore him. Everything is put in threes. I mean everything three times. It can’t say anything once, it has to say it thrice.
Always eager to please, in a side note, Claude offered to write a more positive review, if I preferred. I actually don’t prefer. I like things real.
That is, I like it real when I can get it. Otherwise, I settle for artificial.
