The Taproot Podcast - Part 8: A Psycho-History of American Psychology- You must never listen to this, It should be destroyed!

Episode Date: May 19, 2026

Episode 8: The AI Therapist, the Generational Wound, and the Real Medicine The American mental health workforce is on track to be displaced by AI within ten years—and the psychiatric establishment isn't fighting it. They are welcoming it. Backed by venture capital and smoothed by insurance endorsements, AI therapy platforms are the ultimate fulfillment of what the "apparatus" has been building toward for 40 years: a delivery mechanism for psychotherapy that finally removes the unpredictable, unmeasurable human from the room. In Part 8 of this 9-part series, we expose what the AI replacement will actually do to the field of psychology, and why the variables that truly drive healing are the exact ones the industry pretends do not exist. In this episode, we explore: The AI Takeover: The meeting in San Francisco, what is actually being built, and why the psychiatric apparatus embraces the automation of therapy. The Generational Wound: How trauma shifts from the Greatest Generation to Gen Alpha, and the specific therapeutic interventions the "AI generation" is being shaped to need. The Convergent Rediscovery of Depth Psychology: How independent pioneers—including Richard Schwartz (IFS), Peter Levine (Somatic Experiencing), Bessel van der Kolk, Stephen Porges (Polyvagal Theory), and David Grand (Brainspotting)—all converged on the exact same picture of how trauma lives in the nervous system. The Dodo Bird Verdict & The Real Active Ingredient: Why 30 years of empirical research points to the therapist's regulated nervous system as the primary driver of successful outcomes—and why the industry ignores this. The Cost of Ignoring Culture: Groundbreaking insights from Tanya Luhrmann, Arthur Kleinman, and WHO data showing why non-Western cultures often see better long-term outcomes for schizophrenia. Beyond the DSM: Breaking down the 8 layers of human suffering, predictive processing, HiTOP, RDoC, and Karl Friston’s free energy principle. Why replacing the DSM with dimensional models will still fail if we strip away the human connection. The active variables of psychological work are inherently untrackable. The industry has spent 40 years pretending that only the measurable is real, paving the way for the cold efficiency of artificial intelligence. But the real healing continues anyway, transmitted hand-to-hand in the rooms where it has always lived. About the Host Joel Blackstock is a Licensed Independent Clinical Social Worker (LICSW), Clinical Supervisor, and the Clinical Director of Taproot Therapy Collective in Hoover, Alabama. He specializes in Brainspotting, Emotional Transformation Therapy, qEEG neurofeedback, Jungian psych, and somatic/depth approaches to trauma. 🌐 Learn more and connect with our Therapy Clinic in Hoover, Alabama. Listen to the full 9-part series to uncover the history, capture, and future of American psychiatry. Tags: #AITherapy #Psychotherapy #MentalHealthPodcast #TraumaInformedCare #SomaticTherapy #PolyvagalTheory #InternalFamilySystems #Brainspotting #BesselVanDerKolk #FutureOfTherapy #DepthPsychology #JoelBlackstock #TaprootTherapy #PsychiatryReform #DSM6

praying in the landfill on our knees, trading old songs for new IDs. Are we warm ghosts in the machine? Hey, guys, this is Joel, and welcome back to the Tapreet Therapy Collective podcast. Psychotherapy on the couch, a psychohistory of American psychology. What psychology can see, what it can't, and why it has trouble telling the difference. This is part eight. I'm on vacation in Alaska, so I don't have all the bells and whistles for editing. I hope this one turns out similar to the other ones, but we'll go ahead and jump into it,

and I did want to go ahead and get these out so you all don't have to wait another month. But there is a meeting happening right now in some office building in San Francisco or Austin or one of the other usual places where engineers and venture capitalists are working out exactly when they can replace your therapist with a large language model based on AI. And they're not figuring out if they're going to do this. They're trying to figure out when they're going to do this. The technology is already good enough by the metrics that they care about that we've already talked about to produce the text that looks like therapy, the technique that is evidence-based practice. The remaining problems are regulatory.

They're not technical. And they're going to solve those problems. And they're going to roll out that product. And when the evidence indicts that, we'll just have to figure out what went wrong inside of this system that we're already doing as we're doing it, because that would be evidence-based. The American mental health workforce is going to be displaced, you know, in 10 years as much as we let them. And the apparatus that we've spent seven episodes tracing, the apparatus that built the DSM and captured the credentialing pipelines and engineered the cold machine version of psychotherapy across 40 years and still can't see what works and what doesn't, is going to embrace the replacement enthusiastically. Because the replacement is what the apparatus has been working towards this whole time. And the AI is not a threat to the apparatus.

The AI is the fulfillment of the apparatus. The AI is not even really the villain here. It's just the latest iteration of a system. And the apparatus has spent 40 years training the field to deliver something a large language model could produce already. You said therapy is if this than that. If you go through and you replace these things with just a flow chart that says like somebody uses at a call center. You know, when someone says this, you say that. When they say this, you say that is a cognitive distortion.

The interventions for this cognitive distortion are these three things. If you think therapy works like that, then why not use a large language model? And that's what's happening. And the answer that I have kind of been trying to get at this whole series that isn't I know it answer. It's a you should know it answer. And we all really, on some level, know it answer. Is that the active ingredient in psychotherapy is not trackable.

Not metaphorically untrackable. really empirically untrackable either, but the variables that determine whether the work succeeds can't be measured en masse inside of systems that hold no humanity. The depth of a practitioner's work on themselves is an important variable, and it's one that the system doesn't want to acknowledge because, one, it's hard to track, and then two, it's expensive to produce. The sole accumulation of clinical maturity over a career that nobody measured because nobody knew how is important. And a lot of times the bad therapist, the people who are in school with them, know that they're going to be bad therapists. But the machine, the people who are supposed to be finding what makes therapy

good and bad, can't see that because it means, one, it's not teachable. Two, if you pay us a lot of money, we can't just teach anyone how to do it. And then three, it might require a process that is not a formula. And that is largely what academic training programs are scared of, but they're not really the biggest problem or they don't have the most power inside of this apparatus. You know, what produces outcomes when outcomes get produced in psychotherapy, you know, the meta-analysis going back to the 70s have been saying this, it is not really one technique versus another technique. You know, every honest piece of outcomes research that has ever been done

has converged on the same finding, which is that the technique is not really the active ingredient. The therapist is, something about the therapist. and something that may be different for multiple therapists, which is the competition that drives people to sell and create and market new models. Good ones. And the apparatus has spent 40 years receiving that evidence that the therapist is the thing, and filing it and continuing to operate on this assumption that the technique is really what matters

because the technique is what can be measured. And the therapist's interior cannot be measured. The apparatus has built itself around this comfortable fiction that the measure is the real, because the measurable is what people who cannot tolerate mystery in themselves and outside of themselves can defend and can see. And if they can't see it, it's not real. And that's the thesis. I'm going to defend that again. You know, the work of psychotherapy is irreducibly mysterious, maybe to the therapist as much as the patient. In the specific sense that the active variables resist quantification and the apparatus that has captured American

psychiatry is composed disproportionately of people who entered the field because they could not tolerate the mystery and the need to convert the work into something that could be procedurally administered. And that aligned with the interest of capital. But capital could not do that by itself. Capital needed the insecurity of the providers and the experts and the hierarchies of the profession. And the cold machine is the institutional form of the personal refusal to encounter what the work actually is, the work of therapy. The AI is what happens when the cold machine finally finds a delivery mechanism that does not require a human at all. And the field celebrates the arrival because the field has been preparing for the death of the warm ghost

the whole time without really admitting that this was the destination. Because they didn't know. Again, you could hold a gun to these people's heads. Just because you're motivated and building towards an end does not mean that you know what that end is. And this is the end that they will. wanted. But it's not the end that they knew they wanted or what it meant. And a lot of them will post about it on LinkedIn and complain while furthering the very apparatus that got us to this point. So I'm going to start this episode in a specific room because I need you to see what is actually there. Before I tell you what is going to replace it, because when it's replaced it, it probably will be too late. And these things won't die. They'll just go into fields other than American psychology,

which is sad. That's the field I work in, and it's the field that I want to see succeed and survive. So picture a room, two chairs, a lamp, tissues, a rug, a print on the wall of something that is, you know, meaningless and obtrusive is not going to distract, like a tree or a butterfly or something. And it'll target, you know, like a brown with a swirl in a bird somewhere.

Two people, one of them's a therapist who's been doing this for 15 or 20 years. The other is a patient who's tried everything. Six therapists, four psychiatrists, eight medications, two rounds of manualizable CBT, a week of intensive outpatient, a residential stay. You know, largely the kind of patients that come to me. They're chart lists six different diagnoses collected across the years, contradictory and redundant, and none of them describing what the patient feels and actually lived with, all prescribed, you know, or all given to them by providers in the diagnostic model, the biomedical model, that's supposed to result in consistent psychiatric diagnoses, but hasn't.

Something that is happening in that room will not appear in any randomized controlled trial. The therapist is doing something that has no CPT code, no manual, no protocol, and she's paying attention in a way that the apparatus cannot measure. Because the thing that that therapist, I said she, but it could be a guy, it doesn't matter, is paying attention to has no measurement instrument. And the reason that it has no measurement instrument is that the thing is not located inside, of either of them. It is located in the space between them, and the space between them is by definition what the apparatus cannot see. We can only be felt, and numbers do not feel. What is happening is the

nervous system that has been living in chronic threat for 40 years is for the first time in memory registering another nervous system as safe. Not the idea of safety, not the cognitive assessment of safety, the felt state of being accompanied by another animal who will not hurt you. Something that is very old now is happening ancient. Something that predates language, predates the diagnostic manual, predates American psychiatry, predates the apparatus and all its works. The therapist has been trained in something, but she is doing that, or he is doing that, right now, or Zay is doing that.

Pronome doesn't matter. The training got them to this room. But it isn't just that the therapist sees the patient and is not the rest of the because there's lots of formalized models of therapy that are not threatening, but they can't see you. And there's lots of systems, like the diagnostic systems, that can see you, but they cannot love you. The real encounter is to be seen and to be held and loved at the same time. The person is seeing things that the patient doesn't see about themselves. the person is understanding things that the patient may not understand about themselves.

And at the same time, the, and at the same time, the therapist's nervous system is not rebelling against what they see. There is no avoidance. There is clarity. And that is the closest that I can articulate to the ingredient of what works, is that you are witnessing and you are not flinching. When many people have had people who accept them avoidantly, but they don't really, feel witnessed or held. And there's a lot of people who have been seen and they don't feel loved. And there's a lot of people who have been loved without feeling seen. And the goal of therapy, the goal of the successful ingredient of change is when that witnessing can be accompanied by an

acceptance. And there is a limited, there are some formulas that are better than others, but there is a limited extent to which you can formulate that process. And that is the river that has run underground while the apparatus was building walls. It never stopped being there and it never could stop because the thing happening in the room is not a modality that the apparatus can credential or withhold. It's a capacity built in the human nervous system. And the question of whether it gets exercised depends on whether the clinician has been trained to recognize it and trust it and follow it. And this is what the language model can't do, the LLM or the AI. Not in 10 years and not in 100. The large language model doesn't have a nervous system. It has no body. What it has

What it has is prescriptions and intellectualized information and an approximation of what it thinks empathy sounds like, which a lot of therapists have been trying to do too. I'm not saying that it's worse when an LLM does it than when they do it. They shouldn't do it either. This approximation of understanding this, I totally understand what you're going through. It's so frustrating when this happens. You know, my guidance counselor in high school used to always, when you came into the room, would be like, I understand what you're feeling. Could you tell me what it is then? Oh, no, I can't.

What are you feeling? Why did you just say that to me? You know, the macro asks what the micro asks. And I've said this every episode. I'm going to say it now because this is the episode where the answer matters the most. Or we're creeping up on what the ultimate answer is. The way that the field defends a failed paradigm against its own data is the same way that a patient defends a failed self-image against their own felt experience.

The institutional dissociation is the personal dissociation. It's just scaled. It's just exponentially built. The apparatus is the patient writ large. The worst parts of the patient writ large. The parts that they're coming in to change. And so it is impossible for the apparatus to change those parts of the patient. And the question of whether the field can grow is the same question as whether the individual practitioner can grow,

which is the same question as whether the individual patient can grow, which is the question that has always been worth asking. It's the same question of whether or not you can teach psychotherapy from one person who is good at it to somebody who wants to do it but can't yet. And so the apparatus has one remaining move here, which is to hand the practice of psychotherapy over to an algorithm and declare the problem finally solved. And the infrastructure for this is already being built. You know, BetterHelp is the most obvious example, but it's not the only one. You know, Teledocque Mental Health, Talkspace, Cerebral, which briefly imploded when they hit a regulatory scrutiny for two seconds. It's just been followed by a dozen successors.

You can't kill these things. It's a system. And the model is consistent across them. A tech company builds an app that matches a patient with licensed clinicians, often with packages or some subscription model, like all you can eat therapy in a month except, oops, there's nothing available when you want to schedule. So how much can you really eat?

And it delivers the session through video or chat. And they're always there a few. It's like a best friend. The thing that's missing from the world, you can now get through this. And then, you know, It handles the billing and collects the data. And the patient pays a subscription fee.

You sign a waiver that lets AI process all of this stuff and that your sessions be recorded, of course. And then that's routed through this employer wellness program so that employers see this as cheaper than paying for insurance for you or having an EAP, you know, employee assistance program. And the clinician is contracted, you know, as this independent gig worker like Uber or Amazon delivery. They're paid per session. at the rates that the platform sets with minimal professional autonomy over caseload or scheduling. And what kind of providers are going to sign up for this?

What kind of providers do you think that you're going to get? And it's what they've always wanted, you know, to have all this data to make therapy more efficient based on figuring out all this technique that can be part of the apparatus but not part of the human to really get it what works. So it can be the most evidence-based kind of therapy ever. The session is no longer a private conversation

between two people in a room. It's a data event. Everything that happens is, The patient's chat logs are stored. The video is sometimes recorded. The clinician's response times are tracked. The outcomes against whatever metrics the platform has decided constitutes outcomes. Usually the PHQ 9 or the GAD 7 scores that the patient self-report at the beginning and the end of the engagement. And the data is the product. The person purchasing the data is the real customer, not the patient trying to get well. Therapy is just this loss leader that gets you to the real product. And the company is not a business to deliver. therapy, the company is in the business to capture the data that therapy produces and monetize it. Oh, but don't worry, it's not for bad reasons. We need all of this data so that we can finally break down the walls, see into the therapy room, and do better research studies and meta-analysis on the variables that actually produce change because we care so much about you.

You know, BetterHelp was fined by the FTC for selling user data to Facebook and Snapchat, including the fact that users had sought mental health treatment and what they had sought for despite explicit promises that the data would not be shared and the fact that this was a violation of HIPAA law. And the fine was a rounding era in the company's revenue. You know, they make more than they do selling the data. And this was done so that if you went to therapy and you said that your mom died, Facebook could target you with ads that said this crystal helps magically through the planet Venus for people who have a dead mom. That's how it was used. And the next move, the one the industry is openly planning for is the replacement of the human

clinician with a large language model. Why pay a contractor $60 a session when you can run an LLM that produces text that is statistically indistinguishable from what a therapist would say anyway if they were good? Because what the LLM says is the most evidence base. So the closer the therapist gets to that, the better of a therapist they are. Why pay anyone at all when the model can scale to millions of users simultaneously at near zero marginal costs? The technology is already being developed. There are multiple startups offering AI therapy right now, and the models pass surface-level tests of therapeutic response. They produce plausible reflections, reasonable refrains, appropriate sounding empathy. They do all this without requiring a licensure board or a training

program or a clinical supervisor or any of the slow human infrastructure that the actual therapy requires. And one of the things that's the most terrifying to me is I've had multiple people reach out and say, yeah, yeah, yeah, I don't like Silicon Valley either, all of these things, whatever. And these are providers, pretty high up business people, and then sometimes patients, or not patients of mine, but patients of other therapists that say, you know, I get all that and I'm afraid of these things too, but I've tried therapy multiple times. And I went to this AI group where I went to a chat bot. And the advice that it gave me was better than I got in therapy. That is evidence of my point. I'm not saying that AI is, that there's this wonderful therapeutic

apparatus that I've spent seven episodes talking about out there, and that AI that is getting worse every single year, and that AI is this villain that's going to take away the wonderful, warm, fuzzy relationship by the fire that you have with your therapist. I'm saying that this thing is dying, and AI is the coffin nail that it is dead. And the fact that AI is better than a lot of people working should fill you with horror, not with hope. You know, the AI therapists, they don't have to work in the deepest sense to succeed commercially. They only have to work well enough to produce a PHQ 9 score that looks acceptable in a 12-week follow-up, which is all the apparatus was measuring anyway. And once we measure it once and publish the study, we're just going to assume that it

works forever and probably never research it again. We'll just hold that up as evidence-based practice for the rest of time. A lot of these studies people fight about, no, no, no, it worked. I mean, what, 15 years ago, you have two studies that said it worked? And now we're marketing. this is solved? I'm saying it's not solved and that the human psyche doesn't work that way. Culture changes. Things change. We have to keep measuring these things. You know, the LLMs will be declared evidence base. The insurance companies will cover them when the human clinician will be priced out of the market. And the deepest work and the work that requires two nervous systems together in a room will become an unaffordable luxury available only to patients wealthy enough to pay out

a pocket for the kind of therapy that apparatus no longer remembers. But the bad thing about that is that at that price point, the patient has too much power. The patient has enough power to make sure that the therapist doesn't challenge them. You know, look at these gurus in Silicon Valley that occasionally have a couple, you know, credentials or PhDs in something. But there's scandals all the time because they charge, you know, $500,000 an hour and they're trying to live in, you know, Palo Alto. They're trying to afford this stuff. And that means that the relationship is not fair to the patient because the patient can essentially pay for a yes man.

And that happens a lot of times in these communities. Look at how the American generational wound has changed over time. Just to do how the average American patient has changed, you know, in pace or against pace. of American clinical psych. The greatest generation got the depression and, you know, the war, and they organized around scarcity and duty, all sleeping I'm dead. And the wound that the generation was over-identified with, you know, the role of the provider, the suppression of the vulnerable child, the demand that emotion be so subordinated to function.

You know, they love their kids through this material provision and sacrifice of their own presence as humans for performance, you know, in a market. And most of them never went to therapy. And they didn't have a language for it. And then they drank. But they still were relatively functional and they had this kind of wisdom and they had a life that was more or less fulfilling. That was available to them. You know, the boomers got the inheritance of all that suppression.

And the countercultural rebellion of the 60s and the 70s was the first generational attempt to name what had been suppressed. And we traced in episode four how that rebellion failed. and got absorbed into market individualism that promised self-actualization through consumption. You know, the boomer wound is the inflation of the self into a project of personal optimization. The boomer narcissism is this downstream of an actual psychological injury, the experience of being raised by parents who could not be present, met with a culture that told them the response to find themselves through buying things, having experiences, and this is what got passed forward.

And then Gen X got the latchkey years. Their parents were often boomers and they were in the workplace and were in a divorce. And the Gen X wound is abandonment. The Gen X coping strategy is ironic detachment. You wore baggy pants and did like, you know, grunge or something. And you didn't ask, you know, what you need because you have decided in advance that asking for that would be embarrassing. Cynicism became this armor. You know, your response to advertising when you had no control over the corporations was just a make

fun of the advertising, but that didn't really solve a problem. And you can recognize, you know, Gen X sometimes, like in culture, because they will narrate their own suffering with sarcasm before you have a chance to take it seriously. And millennials are my generation, and we saw this trap. We could see it, the student loans, the precarity of the economic future. We were hooked up to the firehose of information of the internet, which, you know, kind of took Gen X off guard. When Gen X is graduating college, they think that they're going to go on to New York to make like zines and rock magazines and publishing and like all this stuff that's just gone. They're struggling to open a PDF when like, you know, these babies are like, have iPhones at six. And, um, you know,

these political institutions had stopped functioning by the time millennials were coming out. They increasingly just had no hope for any sort of ecological or economic future. Um, and they saw the boomer trap of self optimization as meaning and they said, no, that's not really real. Um, that doesn't work. We're not going to do that. And then. And then I guess the refusal was correct, but there was no antidote. You know, hanging up a bunch of Edison classic millennial light bulbs on strings at your brewery and then being like, well, we have to have like a personal relationship with the person that makes the beer or something. Like, it's like that dream kind of imploded. Like it didn't really work.

It was this fake world, fake alternative, because eventually you have kids and you've got to go to work and you're trying to figure out how to survive. Same thing that happened, you know, to the hippies, basically, but less dramatic. So where did that pain go? I mean, the people that didn't get hit by the opiate crisis, which hit millennials harder than any other generation, I mean, it largely became this nostalgia. We have no vision for a future, but we can go back and play with toys. I mean, I saw a study yesterday that the amount of toys bought by adults because of millennials has now surpassed the amount of money spent on toys for children.

And a lot of that is because there's no vision of a future. They're just looking back to things that they were nostalgic for, that were on TV. Even when the culture was at its peak, you know, something like Tim and Eric, the Tim and Eric Awesome Show, great job, you know, Eric Wareham, Tim Heidegger, then they were preempting this. You know, the things that you see on a show like that were visual techniques from the 70s and 80s.

They weren't for them the 2000s when, or the aughts, you know, in the 2010s when that stuff was being made. It was inherently this sort of like hopeless looking back because there is no solution culture. And now Gen Z is inheriting this dissociation as they start, you know, thinking, basically. From the very beginning, they don't even have hope before they go to college. They never knew anything else. They came of age inside this algorithm swimming amongst this information. And the screen as this co-regulator was something that happened in elementary school. They spent more time looking at a Chromebook than a teacher than a person.

They spent more time looking at a little animated avatar that said, I don't know, ancient Egyptians did this in 600 BC or 12 BC, I'm not sure. Do you want a clue? While the teachers like sitting at their desk trying to manage the algorithmic functioning of the Chromebook cart, the diagnostic identity is the substitute for the selfhood that the apparatus left them because the apparatus stripped out the deeper categories of meaning over the previous 50 years. And the only language for selfhood that remained was the language of a disorder. That's why these kids think of basically the DSM like it's a astrology. To watch an 11 and a 12-year-old on Reddit talk about their identity in terms of pathology is pretty wild. They're not grown up yet. I am anxiety. I am Arfit or something. You know, not because the diagnosis are wrong, which a lot of times they are. If you can call me and say, I have dissociative identity disorder and one of my parts is Pikachu and one of my parts is Naruto and one of my parts of my parts of

whatever. I mean, I see a lot of DID. It's very rare and most of those referrals come from psychiatrists. But if you can do that, you don't usually have DID. It's not that your suffering's not real. It's not that you're lying. It's that you are looking for an appropriate container for the dissociated neurosis of your emotional self inside of a diagnostic manual. And you find this disorder and it makes sense. And so it becomes the language of your pain, which is an enormous problem. An enormous problem. And that Gen Z wound that is the dissociation that began as a millennial coping mechanism became for the next cohort the only state of consciousness that they've ever known. They didn't lose contact with their bodies. They never had it. And they're growing up unmerged

from the algorithm because they've never been merged with anything else. And this is setting you up for a cyberpunk dystopia where these kids are told that their bodies are the enemy and that they're They're supposed to break them and change them that they're bad. I don't know what this means for Gen Alpha. I don't even want to go that far. I mean, it's the first generation that's been algorithmically optimized at a developmental scale by social media systems that didn't even exist when their developmental theories were being written.

You know, their suffering looks different. Their language for it is going to be different. And their relationship to embodiment and identity and continuity is different. And the apparatus has nothing built for them. The apparatus has at most the same 12 sessions of CBT repurposed for a teen audience. We'll put another two letters after CVT to pretend it's a different thing. You know, with this treatment resistant by design population, that it's going to make the apparatus' numbers look even worse.

And then its reaction to its numbers looking worse will be worse. Because when this system feels insecure, it doesn't change. It doubles down on what doesn't work. And that's the future. That is, this is the population that the apparatus has built and that is speeding towards oblivion. This is the last phase of this cold machine project. The nervous system that was asked 130 years ago to become compatible with the factory clock is now being asked to become compatible with the language of a model. It has mistaken diagnostic codes that were supposed to describe pathological elements of human experience for identity itself.

and a lot of times political movements that were supposed to be reactions to negative material forces for identity. So it's not good. In the last 20 years, there's something that the field failed to notice. There's a series of clinical innovations that have been independently developed by different clinicians in different countries from different starting points that are converging on the same picture of what the mind is and how it heals. And the picture that they're converging on, and there were alternatives to this the whole time, you know, because you can't kill that river. You know, Tanya Lerman is this cultural anthropologist at Stanford, and she trained as an ethnographer, and she spent her career embedded in clinical and religious communities watching how minds work in context. In context, the thing that the machine can't do anymore.

We've driven anthropology and philosophy and a lot of co-related fields out of psychotherapy. So her 2000 book, you know, book, you know, of two minds is this ethnography of American psychiatric residency training. And it's one of the few honest ones. I mean, one of the only ones anyway, but Lerman recruited people diagnosed with schizophrenia in three different places. San Mateo, California, Akra, Ghana, and then Jahani, India. And all of these people met criteria for schizophrenia under the DSM, all of them. And the team interviewed each participant about the content, the quality, the relational character of the voices that they heard, and what the voices said.

Were they recognizable? Were they kind or cruel? And did they have a relationship with the person hearing them? any kind of relationship. And did the person have any control over them? Did the person consider the voices a problem? None of this data was terribly subtle in the story that it told. You know, the Americans, the patients in San Mateo,

almost uniformly described their voices as violent, intrusive, and hateful. The voices commanded them to hurt themselves or to hurt other people. And the voices were experienced as foreign, invading, and as evidence that something was deeply wrong with them. And the Americans hated their voices, and they wanted them gone. The voices were torment. You know, the Indians, the patients that they interviewed in Chennai, described something almost completely different.

Their voices were often family members, mothers, fathers, deceased relatives. The voices gave advice. The voices reminded them to do household chores. The voices were something sometimes irritating, but mostly they were experienced as part of a relational world that included both the living and the dead, a continuum, you know, something that was a part of meaning-making. And the Indian patients did not generally consider their voices evidence of disease. They considered them part of how the world worked.

And some of the voices were unpleasant, but the voices as a category were not. In their experience, they were just a sign that something was happening and something needed to be paid attention to. And the people from Ghana, you know, they described something different. Their voices were often experienced as spiritual presences. God spoke to them. Their voices were frequently benign, helpful, morally instructive, and the Ghana patients often felt special, chosen or marked for a spiritual purpose. The voice hearing was not a private disease. It was part of a religious world that included direct contact with the divine and ancestral beings that included relevant information. And some of the voices were difficult, but the experience of hearing the voices was not

pathological to those patients. Now, go back. to that the wound speaks episode of this series. Psychosis, sometimes autism, and definitely conspiracy theories in a society. They are canaries in the coal mine for what's going to happen and where things are going. Lerman doesn't have an answer here. She's an ethnographer. Her argument is that the cultural framing doesn't just overlay an identical underlying experience. the cultural framing constitutes the experience.

And when a culture is sick, psychosis will tell you that a culture is not working. We're a part of a culture, like a mental health apparatus. The Americans are not reporting the same voices that the Indians are reporting through a different filter. They don't have different associations. The Americans are having different voices, and the voices that the Americans here are violent and intrusive because the American framework, the framework of the medicalized cell, the framework of the bounded individual, the framework of pathology, as something foreign invading a clean self, of purity produces voices that have those properties.

And the voices in Ghana, what they hear are spiritual and morally instructive, because the Ghana framework, the framework of a poor itself that exists in relationship with ancestors and gods and spirits and a gentle, non-invasive, non-literal mythology produces voices that, have those properties and meaning that has that character. And the framework is not interpretation laid over a universal substrate. The framework is part of what generates the phenomenology. And this is a very strong claim. And Lerman backs it up with a lot of careful work.

And what I want to do is connect it to where we are going next in this episode, which is the predictive processing framework. Because what Lerman is saying, because what Lerman is saying, you know, ethnographically, is what Carl Fristin, who's a neuroscientist that works mainly with math, he's also a mathematician, has been saying computationally. And the two of them have been pointing at the same thing from completely different sides. I mean, even someone like Carl Jung, I think, is ending up through phenomenology

and just exploring his own kind of felt sense, or Eugene Glendon, who does the same thing. He's kind of a 70s somatic medicine guy. But they're entering into the same thing that Carl Fristin is saying with neurodine, neuroscience and numbers. You know, Carl Fristin is using numbers in a completely different way than Theodore Porter is critiquing or something like the evidence-based practice paradigm and its worst implementations I'm critiquing. You know, and Fristin basically says the brain is just this prediction machine. The problem with an organism is that it can stay alive if it can predict the future better, yet we can't see the future. So what do we do? So his answer is that,

you know, the brain maintains this internal model of the world. And the internal model is built from cultural priors, you know, from learned expectations, from the high-level interpretive frameworks that the culture supplies. But when the brain generates a voice, the voice is not a raw sensory event. The voice is constructed by the brain's interpretation machinery, using whatever priors that the machinery has been trained on. And if the priors are American medicalized priors, the voice gets constructed as a violent intrusion. And if the priors are, you know, a Ghanaian spiritual prior, then the voice gets constructed as a divine address. Same neurological capacity, different cultural priors, different phenomenology, different outcome.

But the apparatus is not able to absorb a critique or an understanding like that. The American outcomes are worse. And the American schizophrenia in Lerman Sample suffered more, hated their voices more, experienced more functional impairment than the Indians or the Ghanians. And the pathology framework, you know, the framework that the DSM produces and that American culture transmits makes the disease worse. You know, the cultural overlay is not a confounder of the universal disease. The cultural overlay is part of what determines how severe the disease is and how much the patient suffers. And the apparatus, by training the entire American population to interpret voice hearing as evidence of a foreign invader inside a clean self,

is a structurally implicated, you know, in producing this form of schizophrenia that is the hardest to live with. That even in America, you know, we have research studies like we talked about in the hearing the wound episode, the wound speaks episode, episode, episode four about conspiracy theories and psychotic delusions changing over time. What's if schizophrenia would become in the 1970s was not what it was before. And I don't think that it's a good thing or that people should live with it. And, you know, I've heard people make this critique that you could, you know,

just sort of accept schizophrenia until it goes away. Like Artie Lang, you know, had clinics like that in the 70s. And I don't know a ton about that. That's not what I'm saying. I do know people that their life is profoundly disrupted. And schizophrenia is something that does need to be medicated. I'm saying that we can learn from the way the rain works and what the data says. And that's really all that I'm saying there.

I'm not making a case for how we could treat it. That it is psychosis and conspiracy theories, again, are kind of a canary in the coal mine. And sometimes people that are, you know, very highly intuitive, sometimes like ASD type, highly intuitive people that understand systems really well. You know, they feel these things coming before they're here. And so I want to put this, you know, parallel finding, you know, which is the World Health Organization. You know, they did a pilot study of schizophrenia, which it's been replicated multiple times. the 70s, but, you know, the IPSS found repeatedly that patients with schizophrenia in low income and middle income countries had better long-term outcomes than patients with schizophrenia

in high-income Western countries. And the finding has been controversial because people have tried to attribute it to, you know, selection bias or a measurement artifact or, you know, misdiagnosis, something. The data is not really the same. So apples to oranges, we can't compare it. But you look at the iPS study and then replicated something like that across a couple decades. It looks like apples to apples. Or closer than apples to oranges at least. So the basic finding is held up across decades. And it's that people with schizophrenia, you know, they recover more often.

They function better and they live more integrated lives in places where the medical apparatus is less dominant. And there could be a lot of reasons for that. Like there are other technological apparatuses that are less dominant in those cultures. And so you may be able to function better if you don't have to wake up at nine o'clock and go to a job. Or if you're able to subsistence farm, it just creates less problems when the world is less structured. And you have something that makes your ability to conform to a structure less. So there's other elements to that. But the apparatus is interventions, the chronic antipsychotic prescriptions, the institutional framing.

of the patient as a chronic case, the social isolation produced by the diagnostic identity, they are part of what produces the chronic course that the apparatus then treats as this inherent part of the disease. And Lerman's Voices Research and the IPSS data together form one of the stronger pieces of empirical evidence available for the layered model that I would propose is probably a better one than the way we think about diagnosis through numbers and checklist right now. And again, I'm not saying that, you know, diagnosis is meaningless or anything. I'm just saying that we need to think about the context of these things. We need to think about the role of symptoms in what we call mental illness or even self. And that looking at that with a little bit more context and looking at the

space between the numbers is important. So if we take seriously the anthropology and the sociology, the numbers from other professions, I think one of the things that emerges is that the cultural layer is causal of a lot of the stuff. Not just the biology. You know, it shapes what symptoms become, how patients experience them, how families respond, what outcomes are possible. And any model of psychopathology that does not have culture as a semi-load-bearing layer with mechanisms by which it shapes the other layers is not modeling reality. Only the black keys will tell you the story if you can hear them in the noise. There is something more human, unique against the melody rewritten every year by the year by the story. young, we hear it better, older when our ears lose the soft sounds losing wonder, but if we are not too careful gaining a sense of all.

The dirt has been so many things that it cannot remember the earth, cannot separate who it is from wonder in itself, because it is always waking up into a new creation. Pay what you owe The babe says as it cries Who do you think you are Asked wild flowers in the graveyard Remember something else Says the misspelled word in the suicide note And this is what Arthur Kleinman has kind of been saying since 1980

Kleinman is a psychiatrist and a medical anthropologist at Harvard and his books, particularly recent thinking psychiatry and the illness narratives are the foundational text of cultural psychiatry. And Kleinman developed the concept of idioms of distress, which is the recognition that different cultures have different available vocabularies for expressing suffering, and that the vocabulary is not just a translation problem. The vocabulary shapes what the suffering becomes. So a Chinese patient in 1980 presenting with what an American psychiatrist would call depression, often described their suffering as physical exhaustion and bodily pain. And the semantic presentation was not a defense against the real psychological depression underneath.

The semantic presentation was the experience. It was the way that this person in 1980, based on their culture, experienced depression. Was, I'm hurting and I'm tired. And the American move to diagnose it as mask depression and prescribe an SSRI imposed a cultural framework that didn't fit and that produced outcomes that did not match the framework's predictions. And I'm not 100% of Kleinian. We're just looking at different people who have actually are serious scholars that publish about these ideas. And then I leave it up to you how these may fit together.

But, you know, Lawrence Kiermeyer at McGill, he's been doing parallel work for 30 years, building out the cultural psychiatry framework into this serious empirical research program. And Keirmeyer's work on cultural formulation, on the way symptoms get constructed at the interface between cultural narrative and embodied distress, on the way diagnostic categories travel internationally and what they do when they arrive is something that is some of the most important work happening in the field. Just when you put the whole world on the ICD10 or ICD11 and then you start to make them think that way, does that create any problems that weren't there before? That's an interesting question. That could be its own podcast. Even if you solve more problems than you create. Did you create any?

Ethan Waters wrote a book in 2010 called Crazy Like Us. That is this popular synthesis of all of this kind of work. And I think that's one of the better books written about American psychiatry that nobody who works in American psychiatry reads. But Waters traces these specific case studies of American diagnostic categories being exported to other cultures. and what happens when they arrive. And that, you know, exportation of American depression to Japan in the late 90s through this coordinated SSRI marketing campaign

that Glaxo Smith Klein ran with the cooperation of Japanese academic psychiatrists is one of the cleaner case studies there. And, you know, Japan in the early 90s did not have much of what we would call depression. Japan had a different framework for what we would consider a similar phenomenon. You know, the pharmaceutical industry, recognized a market opportunity there and they deliberately worked to change Japanese understanding of suffering so that the suffering would become

diagnosable as depression and therefore treatable with SSRIs because Japan was a lot more of a private culture you were sort of performative about certain unique things that in America might be private or less public and then you were very private about things that in America might be more public and so the way that the language of diagnosis doesn't fit that You have to change the way that people make private and public space in order to diagnose them because you're taking a checklist that is not understanding these things as a dynamic process, which I would argue, you have to understand them as a dynamic process that is changing and blocked energy and forces. If you're not thinking that way, if you don't have some kind of model that does that, you've killed them. When you have checklist, and this is evidence to my point, when you have checklist and you take them to a different culture and they don't work anymore, that's telling you that the truth's not really in the checklist.

right? So within a decade, you know, Japanese depression rates had risen dramatically and antidepressant prescriptions had become routine. And a country that had previously understood suffering through a substantially different cultural framework had been brought into the American framework. And the frameworks in position produced the disease that the framework was designed to treat. And so now it could be treated. So just hold that picture because, you know, we don't want the series on American psychology to leave America too much. But America did become an international model or part of one. So if we were going to go back and say,

what if there's a layered model that considers all of these variables? What would it look like? What could it look like? It wasn't a checklist. Because I don't think biology should be discarded. You know, I definitely think that a lot of mental illness, a lot of self has this biological component.

I just don't think it can hold all of it, you know, as I'm saying. So if you had a cultural layer that's shaped with the biological layer, expressed, then the biological layer shaped what the cultural layer could develop. And the relational layer shapes both. So the behavioral layer is built out of the interaction of all of these. And the apparatus has been trying for 40 years to identify a single layer that causes all of it. You know, usually looking at neurochemical. And then saying, well, we'll just go ahead and say dopamine disorder and then, you know, the field will catch up, the science will catch up. And then the science has failed to do that.

And I don't think the problem is in any of the directions they tried. I think the problem is in trying to make one single layer that can be screened for and quantified do everything that psychology does, that human psychology is capable of. You know, the disease does not live in any single layer. The disease emerges from the interaction of all of them. And the only instrument that can hold all of those layers at once is a human being who has been trained to do so. If you don't train them to do it and you don't think that those layers exist, then you

can't see them, no matter how much you spend on research. And the modern apparatus of American psychology fundamentally cannot do this. The protocols cannot do this. Even when, you know, these people at Harvard say that maybe we should. You know, the DSM can't do it. A.I. can't do it. Only a clinician can do this. And the apparatus has been systematically training clinicians to be unable to do this for a long time. Static in the... You say you're fine, but your eyes read... So what is a better apparatus than the one we have?

Right? Like, is there a better alternative to everything that I've talked about in the series? Because, you know, there's some good things in the system that we have. What direction do you go in? I think all of them. You know, what categories could replace the DSM's categories? Well, my answer would be rethink what category is. You know, the science has been moving and the science is pointing somewhere.

So let's look if we tear the silos down, the silos of insecurity. Where is all of the science? Other people's numbers that arguably are better than the ones we've been collecting in American psychology. Where do they point? What's converging across multiple independent lines of work, which I don't have time to talk about all these people. I've got a big blog.

I've got a podcast, but even half of what we write about, it's not all recorded. And most of what I write is it jumping off. point to you should go read this if you like this consider these perspectives but you know I can't mention all of them but you know there's a picture of psychopathology that I think could be a layered dimensional mechanistic and transdiagnostic model that is probably more useful to the to even the hospital and that's high top you know the hierarchical taxonomy of psychopathology this is a direction that Thomas Nsel even though he talks about the cultural um he's the nymh who dumps the DSM.

But, you know, he's, he believed in the cultural component of mental illness a lot and the environment. But what he believed in a way that I didn't is that until you understand these things as completely biological in their function in the brain, you can't treat them. So when he was asked about the kind of things that I brought up here, why did he not use high top? Why did he go with RDOK, which was this model that the NIH said, this thing's not going to be done for 10 years.

Like, it's a decade away from being done. it's already better than what the DSM has been doing for 60. We've lost all faith in this. And so we're going to jump to this unfinished model, RDOC, and use that. So what RDoc tracks is just these sort of objective numbers. And, you know, Insull said something about diabetes one time he said, you know, like mental illness, you know, we know that fast food causes diabetes. But when doctors are treating diabetes, they don't go and then pull like all of the fast food

density in your area. they diagnosed based on the biological mechanism of blood sugar and insulin. And I want those same things for psychiatry or else psychiatry is invalid as a science. Well, then I think, you know, again, I don't know, maybe a hundred more years of research in the direction that we've been going, you'll find the biological causes for all of these things. I don't think we ever will. I think that what the deep brain is doing is pretty context dependent, pretty complicated, and that you probably can never treat something like depression in the same

way that you can treat diabetes. Sorry. And again, maybe I'm wrong. But so why did he choose not to go in the direction of high top? I mean, that's what it is. He said basically if psychiatry, even though in culture and environment influence the science, if the science can't be boiled down to numbers in the biology, then it's not really a real science. And I don't think that perspective is right. So, you know, the high top people, the hierarchical taxonomy of psychopathology, they have built this really interesting model 20 years. It's empirical and it's from the bottom up. They take observed symptoms across enormous populations and they look at how the symptoms actually vary in the context of each other in real human beings. And they produce a structure. So you see these interesting relationships

between symptoms when you look at diagnosis that way that you wouldn't have ever really seen before. You know, not a structure imposed by a committee that just says this is the diagnosis, now let's go find the symptoms, but it's a structure that emerges from data. You get all of the data and then you let it tell you what the diagnoses are. You look at every single thing that the person says is going on that's a problem and then you look at how those problems are all showing up in relationship to other problems that may or may not be a diagnosis that you might want to treat as a category and pay attention to. So there's an internalizing spectrum and there's an externalizing spectrum. So there's a thought disorder spectrum and there's a somatoformed spectrum. There's a detachment

spectrum. And above all of those, there's something called the P factor. And the P factor in high top is the general dimension of vulnerability to psychopathology. How vulnerable are you to having symptoms at all? And let's look at risk factors and what kind of things those risk factors make more present. Some of that's going to be genes. Some of it's going to be environment. So you can sort of see through all of these layers by just looking at the data, which it's like, hey, you keep asking for a number, right? This is a lot of numbers. It looks like your allegiance is not to numbers. It's to the insecurity of the way that the current establishments need numbers to run the show,

which I would say is not a very scientific project. So, you know, the Ardoc framework, which Encel introduced in 2013, and then, you know, he alongside his transforming diagnosis, you know, speech, has been pointing in a parallel direction. You know, Ardoch says, let's stop pretending that we know what schizophrenia and depressive. and anxiety are. Where I think high top is a little bit better is that when you're looking at vulnerability itself, you know, there may be a parent with a major depressive disorder.

And it is not really increasing the odds that their child has major depressive disorder by very much. They are increasing the odds that their child has something, though. So what that something is depends on a thousand other factors, that the apparatus has been actively training itself not to, attend to for 40 years. You could look at maybe a propensity to, you know, pathological relationships or pathological relationship to a religious organization or how maybe some of these things aren't genetic, but the parent who grows up with them has a child that is more likely to have

a personality disorder based on their genetic disorder. I mean, there's all sorts of things that you could see and understand in order to see, you know, how we function. So, You know, Ardoch has not really delivered everything that didn't sell promised. And we've had it for a while. It's about done now. And I want to be honest, you know, about that, especially because I don't love it. But the criticism that Ardoch was too reductionist, to circuit focus, too dismissive of the psychological and the social, was a fair criticism when it was unveiled. But what Ardoch produced, even as it stumbled at the beginning, has been this generation of researchers who stopped believing.

All of that DSM stuff. and they've discovered some interesting things when they moved away from those categories. So one of them is the network theory of psychopathology, which Denny Boersboom and his collaborators have been developing. And they kind of went in this third direction. They stopped looking for the latent disease beneath the symptoms. And they're saying that the symptoms are the disorder. So the symptoms causally activate each other. And insomnia produces fatigue.

And that produces concentration. problems, and that produces failure at work, and that produces shame, and that produces low mood, and that produces insomnia. And then that produces a changing self-image that on a long enough timeline could become semi-permanent. And the disorder is not under the symptoms. The disorder is in the loop, the symptoms form. So if you view these things as living, interdependent processes interacting with

each other, I think you're going in a better direction, which this one is. And, you know, the leverage point and the treatment is not the imaginary disease underneath. It's the bridge, the symptoms, you know, connect and loop against each other. And the nodes, if you intervene in them, you can collapse the whole network back into a non-pathological state, which this would be harder. It would mean that a lot of people have been doing it the wrong way, which I think is the reason that we haven't moved to something better, is that we just keep clinging to the life raft that we're in because it's the one that we got. And even though we know there's better things out there, it'd be a risk to get to them. Some people may not make it or might not be able to. But hey, you're not dying.

You're just not able to continue to practice psychotherapy the way that you think it should be practiced in a way that I would say is wrong. So, you know, it's not as bad as swimming, you know, to another ship. It's just, you know, just that thing. But anyway, this explains, you know, comorbidity in a way that the categorical model can't, really. You know, the comorbidity is not two separate diseases. co-occurring. It's this natural propagation of symptoms activating across this bridge that connects different symptoms. And most people that have one symptom might have another, but other people may form a different bridge that is still related to the symptom, but more rare. And the way that that

could treat something like depression, something like addiction, you know, which addiction is comorbid with a whole lot of different, you're likely to have a lot of other disorders if you have issues with addiction, but you don't necessarily have all of them and some people have none of them. That is a more useful model. So if you look at, when I mentioned Carl Fristin before, you know, he has this idea, the free energy principle. And it is a mode of analyzing predictive processing. And, you know, the framework that grows up around that gives us this unifying language

across all of these converging lines. The free energy principle says that the brain is fundamentally a prediction machine. And the brain maintains this internal generative model of the world. And the model is hierarchical. High level priors come down from the top of the hierarchy. Shaping with the lower levels are allowed to perceive. And so allowed. It filters out information that is too against our ideals,

too against our ego, even evidence that you're giving the patient that the way that they're acting is not correct or consistent with their own values. That can be removed. You know, so sensory evidence comes up from the bottom, comparing against what the prior is expected, and then generating a prediction. And sometimes there's an error in that prediction when the sensory evidence doesn't match. So the whole architecture is organized around minimizing prediction error. It wants to make sure that it doesn't encounter something that it didn't think would be there.

And Fristin calls that minimizing free energy. So the brain wants its predictions to come true because predictions that come true, the model is accurate. And an accurate model means the organism can act in the world successfully. It becomes pathological when we start pretending that our predictions are accurate, even though they aren't. Because the reward system of, you know, getting that prediction correct becomes more pleasurable than actually being effective in surviving as an organism, right? You know, the brain creates something outside itself there. So there's this critical. variable in the Friston system, which is precision weighting. You know, the brain is constantly

deciding how much confidence to put in its priors versus how much confidence to put in incoming sensory evidence. And so if the priors are weighted heavily and the sensory evidence is weighted lightly, the brain ignores contradictory evidence and it stays committed to its existing model. If the priors are weighted lightly and the sensory evidence is weighted heavily, then the brain updates its model rapidly in response to whatever it encounters. And the balance between top-down prior weighting and bottom-up evidence weighting is set by neurotransmitter systems, by attention, by emotional states, by cultural training, by every layer of the system at once. And here is the move that makes the framework so powerful for psychiatry. Every major psychiatric

phenomenon can be reframed as a specific computational failure of the prediction error machine, and with a specific signature in terms of how the precision weighting has gone wrong. So that's pretty useful for something like a personality disorder. That's very useful for something like schizophrenia, a delusion or a hallucination. When you're looking at a model instead of just, is the symptom there or not? What was the gene that made it be there? What's the biological reason? Well, that's biology and its numbers, but it also isn't as simple as American psychology wants the numbers to be. So psychosis is the classical predictive processing account of, you know, Psychosis like schizophrenia involves underweighting high-level priors and

overweaving low-level prediction errors. So the brain stops trusting its model of the world. Every minor sensory anomaly generates a strong prediction error that determines explanation. So the brain scrambling to explain this cascade of unexpected events generates new high-level beliefs to account for them. And the new beliefs are not constrained by the previous model because the previous model has lost its precision. So the explanations become bizarre, but they feel true.

Delusions. They form not because the brain is broken in some random way, but because the prediction error machinery is doing exactly what it's supposed to do, except that the precision weighting is set in a way that produces a runaway cascade of error correction. So hallucinations in this framework are something slightly different. You know, there are cases where high-level priors break through and become perceptually vivid where the brain's internally generated predictions get treated as if they were external sensory events.

And the model predicts the voice. And because the precision weighting has gone wrong, the prediction is experienced as perception. So the brain literally loses the ability to tell the difference between what is outside of it and what it is making up. And then what is making up becomes more real. Then what is outside.

And I stop making sense to other people, but I still feel completely right. to myself in a way that you can't talk me out of in a way that it kind of isn't wrong because usually I'm making some sort of metaphor based on information that was true at some point but I've lost the sense of kind of time updating and context now you know hold this next to and this would explain why CBT for psychosis and schizophrenia which there is a manual for that I read and I thought it sounded crazy and tried to use it once in the hospital and said, yep, this does not work.

But it would explain why something like that does not work. You know, the American patients in, you know, San Mateo from Lerman's research, the, you know, foreign, violent, intrusive voices, the Ghanaian patients in Acra. And, you know, the Fristin framework, if you take that with the Lerman stuff, you know, it would predict everything the same. The brain generates voices based on the high-level priors it's been trained on. If the high-level priors are the American medicalized framework where voices indicate disease and a foreign agent has invaded the clean self, then the brain generates voices that have those properties. If the high-level priors are the Ghanian spiritual framework where the direct address from divine beings is part of how the world works, then the brain generates voices that have those properties.

And so the neurological capacity is the same. The precision weighting failure is the same. The cultural priors that shape what the failure produces are different, and that does. difference becomes enormous. You know, autism. The autism story in predictive processing is the computational inverse of psychosis. Autistic brains may have underweighted high-level priors and overweighted attention to low-level sensory detail. So sensory information that the neurotypical brain filters out as a relevant noise gets through to consciousness with the autistic brain. And the autistic brain cannot rely on its priors to predict away the buzzing of the fluorescent light,

the texture of the shirt of the collar, you know, the the slightly off pitch in the speaker's voice, every detail becomes salient and overwhelming sometimes. And so every change is a prediction error. And then the world is overwhelmingly informative. And the behavioral signatures of autism, the preference for sameness, routine, the restricted interest, they're the brain's compensatory strategies for managing a sensory environment that has not been pre-organized by sufficient prior waiting. Depression. You know, depression at a predictive process is something like locked in pessimistic priors with low precision on positive prediction errors.

You know, the brain is built up this model in which nothing will work. Hope is never justified what IFS would call a protective part that wants you to be hopeless because you've learned that all hope is just effort that is wasted. And the model is then defended against the contradicting evidence by reducing the precision waiting on any prediction error that would update the model in a positive. direction and so the depressed person walks past beauty and does not see it and it's not just an intellectual process right like there are somatic markers in these things too so it's I don't want to budget any energy at all I don't want to budget any hope I don't want to budget any feeling creativity it's these unconscious like an emotional somatic things it's not just the kind of intellectual things that I can read to you

it's the entire body deciding the world is hopeless you know anxiety trauma these explain quite a bit in a way that I think the the checkboxes of the DSM don't and the thoughts leave to beliefs which lead their beliefs lead to thoughts lead to behaviors thing of CBT you know kind of can't get and so the clinical implication of the predictive processing framework is is pretty big because every model of psychotherapy can be reframed as a specific intervention on the precision weighting of the prior structure of the brain so CBT and the manualized institutional version of it intervenes at this explicit level of cognitive priors,

asking the patient to challenge the high-level beliefs and then update them. So CBT works, but when it works, it works because explicit cognitive priors are real and can be updated. CBT fails and gaslights people when it tells them that their cognitive faculties have power over parts of the brain that cognitive faculties don't. You know, when it fails, it fails because much of the relevant prior structure is not explicit. It's not accessible to verbal challenge. It is encoded in autonomic and semantic relational layers that the cognitive intervention cannot reach. Behavioral activation works because it generates enough prediction error to shift priors even when verbal challenge can't. You know, exposure therapy works because it generates enough prediction error on the trauma

prior in a context that the trauma prior did not predict to allow the prior to update. You know, something like EMDR and brain spotting, which my industry is pretty hostile to, you know, they work by accessing the precision weighting machinery directly, skipping those cognitive layers that CBT is trying to mess around with, which is why the CBT people don't get it. They're like, how could an eye lead to a thought? Well, it's because you're talking to the body a different level of the hardware, at least as the, you know, the theory goes. Semantic experiencing works by accessing the autonomic priors,

the body's own prediction machinery and allowing it to discharge and re-regulate. You know, IFS works a little bit deeper. It's more of a depth therapy than cognitive and behavioral therapy because it works by activating the multiplicity of the brain's prior structure, being like, yes, this is one structure, but there are many others, and there is also a structure-making mechanism. You know, the parts and renegotiation of their relationship from a different vantage point, psychedelic therapy also would mess with this process. There's a lot of things that I think if clinical psychology paid more attention to Friston would be explained, but they don't.

So why not? Right? He's interesting. Why not? So why resist the temptation to just throw out the DSM now and then use high top completely, you know, if it's better and it accounts for all these different kind of layers of cognition? Well, let's give every patient a dimensional profile. Let's run R-Dog matrices.

Let's compute their personalized network of bridge symptoms. Let's measure their HRV and their autonomic flexibility. Let's get an inflammatory panel. Let's do QEG subtyping. Let's get a polygenic risk score. Let's compute their layered profile across all eight layers and produce at the end of it. You know, this is much more accurate diagnostic document than the DSM ever produced. And then let's bill insurance for that document.

and then let us get new code set in CPT so that we know how to call these new categories something insurance can understand, and then let's train clinicians in the new framework, and then let us build the new apparatus that the new science demands. That's a pretty big lift. You know, that's a pretty big ask. You know, I want to argue that this is the heart of the episode. You know, this is the exact wrong move. And the reason it's wrong is the same reason that the original DSM was the wrong move. Because the science isn't done.

The science is not going to be done. the science is right now in 20206 moving faster than at any point in the history of the field. There's some recent regulatory things and defunding that's making it move in really weird directions and then also losing some of the information that it was collecting. But let's just assume that if science was to keep moving, generally it moves faster in this area because technology gets better. But whatever apparatus you build in 2026 to capture the current state of science, that apparatus will be obsolete by 2031.

and the high top structure will get refined and the RDoc matrix will get revised and the network parameters will get updated and the layered model will get more layers. You know, some of the existing layers will turn out to be artifacts that we realize weren't actually what we thought they were. They're, you know, something related to something else. You know, so any apparatus you build will start aging the moment that it ships. That's just how bureaucracy works. And the apparatus knows that.

And that is the thing that we should be aware of because, you know, the APA has been publishing across 2024 and 2025 and in 2026 something called the DSM6 roadmap, a series of position papers and a working group reports and, you know, consultation documents laying out what the next edition of the manual is supposed to look like and how it's supposed to be developed. And if you read the roadmap, you find some really interesting stuff that you maybe would not I've expected from the same institution that published the DSM 4 and 5. The roadmap concedes in significant areas now that this basic critique is right. It acknowledges that the categorical model has validity problems.

It only took 45 years, right? And it acknowledges that the dimensional model like ITOP has produced better empirical structures. And it acknowledges that the biological universalism that the apparatus has been promising has never come. So it might look like the critics have finally been admitted to the conversation, and the apparatus has conceded this, you know, vastly at this point, empirical point to reason. But then the roadmap announces a multi-year process to incorporate those critiques into the next manual with working groups and consultations and field trials and a publication target somewhere in the early 2030s, but, you know, subject to change.

So I'm going to spend serious time on, what the DSM roadmap, what I think it's actually doing in the next episode. But I think this is a defensive move because they don't want this thing to change. They would rather sink on that ship than join another one or swim to another one. But I want it on the table now because it's proof of the structural argument that I'm making. The apparatus cannot be reformed by a better apparatus. Even when the apparatus itself agrees that the current apparatus is wrong, even when the apparatus invites the critics into the room to help write the next version. The apparatus has at this point evolved a defense mechanism that's very sophisticated,

that it can absorb its own critics by publishing thoughtful documents, acknowledging their critiques, in order to not change. And continuing to operate exactly as it has been operating while channeling the reform energy of an entire generation into a bureaucratic process that will conclude eventually, but just incrementally needs to eke along because any big, new change that might make something materially different on the ground is just not reasonable at any point ever. Does that sound like psychology or politics in America? You know, the trap is the acknowledgement without the change. The trap is the apparatus saying, like America said to the

hippies, like America said to the millennials, like America said to the Gen X, like, you know, here is a 10-year working group to address your rightness. While the billing code, do not change and the training programs do not change and the captured measurement infrastructure does not change. You should still know that we see you and we feel you and let that emotional connection between you and the bureaucracy dissipate your requirements for change by alchemizing it into an emotional energy that I'm reflecting back at you. Does that sound like a large language model? because if you complain about something, that's what a large language model will do.

And I don't know if anyone's seen the movie Avatar, but what I call the bureaucratic move, where bureaucracy says, hell no, we're not going to do that, but we can't tell you that because there's a whole lot of energy asking us to change, what you'll do, and they do this right and left. You will see a politician do the thing from Avatar,

where they say, I see you, I feel you, Jake Salih. and then they don't do anything that the group is actually wanting. They just want to make sure that, you know, you feel heard. You know, the reform within is being staged by the apparatus itself as a defense against actual reform. And the structural argument is that no apparatus can fix this, including the apparatus that is currently inviting reformers to help fix it, because the apparatus that is doing the fixing is the problem. And it is always participatory.

in its own reform by its own design. And deep down, it knows how that works. And so do the people who are doing the working. But they can't know that. They can't let themselves know that they know. Just like a patient in therapy that's participating in their own symptom. And that's the move I want you to track. The frontier is moving too fast for any apparatus to capture.

And that is what is actually different about this moment from any previous moment in the history of psychiatry. We don't need a better manual. We need to stop pretending the manual is the right kind of instrument. Because the instrument that can actually keep up with a fast-moving frontier is not a manual.

That instrument is an informed human being. And the instrument is not just a human being in the trivial sense of needing somebody to deliver the protocol that's not an AI. The instrument is a human being in the strong sense that the active variables of the work are inherently untrackable. And only a human being can navigate untrackable variables.

And the depth of the clinician's own work on themselves limits their ability to do this. And that includes how many books they've read that year. The quality of their attention. The capacity to stay regulated in the face of dysregulating material. The one who says, I don't work with this issue because it just, what do you mean? That's an entire, like, gender of people. You don't work with a whole gender of people. That seems like a lot of the population. Maybe you need to do some work. You'll hear clinicians say that. I just can't work with men. I just can't

work with women because they always, well, I mean, anyway, there's some problematic stuff there. Even when it's smaller issues, that's a strange statement from somebody who's supposed to be participating in this industry. You know, there's an intuition that comes from having sat with hundreds of patients and developed a felt sense for what each patient needs. And none of these can be measured in advance, none can be specified in a manual, none can be transmitted through a brief training, and all of them together are what produces outcomes when outcomes get produced. And the apparatus has spent 40 years pretending that these variables do not matter because they cannot be measured. And the pretense has produced a generation

of clinicians who have not been allowed to develop the variables that matter because the training programs were busy training them on the variables that could be measured. And the result is that the variables that matter have been getting weaker in the credentialed workforce year over year. And the apparatus has been congratulating itself for the improvement and fidelity scores, while the actual quality of care metrics, like the replication crisis in CBT shows us, have been getting worse. We made up a new number that we're getting better at. It's just not how many patients are getting better.

And the science is pointing at the limits of any standardized apparatus, And the question that follows is, if no apparatus can keep up, what do we do? And the only answer, I mean, honest answer, is a well-trained, curious, integratively capable, and supervised clinician with the time and the autonomy and the freedom from the bureaucracy and the freedom from the numbers to actually think about the patient in front of them. And the clinician is the instrument that catches the frontier. Not because the clinician is magic, but because the clinician is the only thing in the entire diagnostic system that can read across all of these layers and context

and space between numbers and wait them in real time against the specific patient's specific history and hold the dimensional profile in their head alongside the network of presenting symptoms notice when the autonomic states shift mid-session integrate the work the reactions and inflammations and attachments and all of the history with the cultural symbolic context with the the immediate behavioral pattern and produce working clinical formation, formulations that update session to session as new evidence comes in. And if you want to say, yeah, but who can do that? The kind of people who the incentive structures and the apparatus should be pulling in

to work in psychotherapy again. Well, that's not reasonable. Why? You better go out there and make it reasonable if you want anything to change. because there's not a system and there's not a metric and there's not an assessment. And there's not something that doesn't kick people out and attracts the wrong kind of person that looks like an educational system. They can tell you something that can make you do that.

The clinician is the instrument because the clinician is the only thing in the system that can hold all of those things at once and see through them. And this is what the apparatus has been actively working against for 40 years. The apparatus has been trying to make the clinician fungible. That word matters, fungible, interchangeable, replaceable. The apparatus wants any clinician to be substitutable for any other clinician. How do we just find out the important thing that they did so that we can make anyone do it? Because that is what makes the billing infrastructure work, and that is what makes the insurance panels work, and that is what makes the credentialing pathways work.

But more importantly than that, it is what makes your faith in those things work. contrary to the evidence that it doesn't. And that is what makes the eventual replacement of the clinician by the algorithm possible. And the apparatus has spent 40 years systematically deskilling the clinical role, narrowing what the clinician is allowed to know, restricting what the clinician is allowed to do, shaming and judging any clinicians that know anything that they're not supposed to, that it's the forbidden knowledge, the black magic.

Which that forbidden knowledge is just the things that Harvard professors are publishing in other fields that are related to psychology, I'll point out. To convert the clinical encounter into a procedure that can be replicated by anything. And the question that closes this whole discussion is, how was that allowed to happen? How does a field that produces a generation of clinicians who can see what is wrong, who can name what is wrong, who have the evidence base that says what is wrong, fail to change, what is wrong? And the answer is not that the field is stupid.

The answer is that the clinicians are complicit. The answer is structural, and it has to do with the way that American psychiatry has positioned itself in the culture, and the way that it has built its sociological armor, and the way that it uses complexity as a shield, and reverence to numbers as this wall. And accountability is something that only ever flows downhill, from the institution towards the clinician, from the highest paid people to the least paid people, and never the other way around. And this is the next episode. This is what episode 9 will be about.

In the first episode, I said that psychology is America's eye and is the apparatus through which we have learned to see ourselves. And we are teaching other countries at this point, or at least influencing the way that they see themselves through our eye. We've put it in their head. And the lens that we use to describe our inner life is this stuff that I'm talking about. I can't even get away from the words that I may or may not agree with in order to discuss these things because they're the words that we have.

And, you know, we cannot see the lens because we're always looking through it. Every time we try and examine psychology, we use the concepts that psychology gave us. That's what I said in the first episode. We critique the DSM using diagnostic thinking. And we question therapy using therapeutic language. The apparatus of examination is doing the examining. And this series has been an attempt to get outside of that trap, not by rejecting psychology, which, you know, I said in the beginning would be,

you know, crazy to do. But by tracing the history of how the apparatus came to be so that we can see it as a historical object rather than this eternal feature of reality that was handed down by God and the only way that things could ever be. But be honest about the limitations of that move. You know, stepping outside the apparatus does not give you a view from nowhere. It gives you a view from another apparatus, which is why the people who jump ship from this into, you know, kind of wild men's rights Instagram or new age stuff

or well everything means nothing man and diagnosis isn't real I mean that's like saying categories aren't real what categories where

sometimes diagnosis is applicable I mean this podcast is an apparatus the blog post is an apparatus the book that I'm writing is you know will hopefully not be I wrote it to resist

something like that but who knows you know every form of articulation is an apparatus because every form of articulation has rules and the rules shape what can be said

and what cannot be said gets left out and what gets left out is still part of the truth. And the eye that cannot see itself in isolation can be seen by another eye. And this is the secret of psychology that the apparatus, you know, could not capture and never will. Healing happens in relationship because seeing happens in relationship. And this stuff is in most fundamental religious texts. You know, the spec that I cannot see in my own eye, or the spec that I see in your eye is caused by the blog that I cannot see in my own. You know, you can't find your own blind spot. You need someone else's eyes for that.

And the other person cannot find their own blind spot either. In systems and bureaucracies, don't provide the kind of accountability that accounts for the subjectivity that is required to check these things in a psychological entity. Psychological bureaucracy. They need yours. We need each other. And the work at its deepest is the mutual offering of sight. It is the communal activity of seeing what could not be seen alone. You know, the church at its best did this, and Union Halls in America,

and the village council and the extended family, and all of the other meaning-making load-bearing structures of our world. The community in its moments of actually being a community does this. Psychology, when it has been good, has done this too, because of those times it was a community. and now it's more of a bureaucracy than a family. And the hour in the room when something works is the communal seeing in miniature between two people. I would like an apparatus, you know, a cold machine, that the ghost can warm up a little bit more.

You know, it's tried to eliminate all the communal ways of seeing. It's tried to replace it with procedures that could be delivered interchangeably and providers that could be interchangeable and measured on interchangeable. and measured on interchangeable metrics and build on interchangeable codes. And it did all of that because it wanted a machine that it could build, that would do accountability and do self-awareness and do a system of growth and discovery in place of the human. And that project failed. If this convinces you of nothing else,

it should be that those things cannot be done bureaucratically. You know, the communal seeing cannot be eliminated. It just goes underground when you make it illegal. But it doesn't die. It waits. And while it waits, it does what it's always done, which is to lie in private practices and supervision groups and accidental conversations between trainees and elders

and the small daily moments in which one person looks at another person and accidentally tells the truth. because they forgot to act and they forgot what they weren't allowed to know. And the cold machine does not know how to see itself. So it builds bigger structures of measurement that look like seeing, but they're not seeing. And the warm ghost knows that seeing requires another eye. So it keeps finding the other eye in whatever crack the apparatus has not yet closed. And that is what the field is underneath everything that I've described.

Two eyes in a room looking at each other. one able to see itself and both willing to see the other as old as humans that is not going away and it's not really winning either it's holding you know the questions remain what is consciousness what itself what is meaning what is healing what is the relationship between body and the mind and what does the mind call itself what does it mean whatever these aren't questions that are going to ever be answered definitively and they are not questions I think they can ever really be answered biologically. They are questions that we are going to live inside no matter where they come from. And that is what psychology in its best sense always has been.

Not a science that closes the questions. A practice that stays with them alongside the suffering person for as long as the staying is needed. And that is the right to be uncomputable that you have. The right not to be a gear in this thing or a bit of data to be sold. hold the right to withdraw from measurement, the right to defeat the apparatus, not the promise that the worm ghost will win, the right to be fundamentally more than the measurement can contain. Because you already are, whether or not you know it.

So don't forget, you know, to remain in the depth of what we are is an open question. And to be met by another open question in a room where neither of you knows the answer. And the not knowing is the practice and the practice is what we have. have and what we have is enough. And the apparatus cannot give you the right to that. The apparatus cannot revoke it either. It's already always, always will be always was yours. And the work at every scale, individual and collective is to live as if we already knew that. It requires the faith to do something as loud and as crazy as myself and a lot of other clinicians have done. Or we just say, hey, come on, wake up, this is it.

And isn't that kind of weird and isn't that kind of silly? And in the next episode, I'm going to turn from the question of what the science is pointing at to the question of why we are not allowed to act on it and what it has already told us. You know, why the field produced this evidence, and it's still structurally prevented from reorganizing around it, and why the apparatus persists when the people inside it can see exactly what is wrong with it. and why reverence is the social technology that protects captured institutions from reform in a completely secular and objective way, and what would have to change for the practice of sitting with another human being to survive the era that we are entering.