Interview with psychiatrist and writer Awais Aftab—clinical assistant professor at Case Western Reserve University, editor of Conversations in Critical Psychiatry, and author of the Psychiatry at the Margins newsletter—on how AI could help reform psychiatry. The host describes how his OCD took a decade to diagnose, and the discussion covers pluralistic thinking in psychiatry, a pragmatic approach to mental healthcare, why disorders like depression and OCD get flattened into diagnostic checklists, why psychiatry is stuck where machine learning was decades ago, shifting the field's focus from explanations to predictions, and how AI is already changing the psychiatric profession. # Channel Presentation Observations ## Explicit Self-Description Evidence - **Channel name and host**: The host is identified as **Dan**, and the outro names the show as **"AI and I"** hosted by **Dan Shipper**: > "smash that like button and subscribe to AI and I... a journey into the future with Dan Shipper as the captain of the spaceship." - **Stated editorial focus**: The host explicitly frames the show's scope as AI's impact on science, especially psychology and psychiatry: > "I've been really obsessed with how AI might change science... in particular how it might change psychology, psychiatry, neuroscience, more complex areas of science." - **Audience addressed**: The host addresses an existing listener/viewer base — "for anyone who's been a listener of the show for a while" — and preempts a likely audience question ("why are you on AI die [likely 'AI and I']"). - **Guest introductions**: The host's standard practice is credentialed-guest framing — the guest is introduced as a psychiatrist, clinical assistant professor at Case Western Reserve, and editor of *Conversations in Critical Psychiatry*. - **Cross-referencing behavior**: The host says, "we'll link to the Substack in the show notes," demonstrating show notes/links as part of the channel's presentation format. ## Observable Presentation Patterns - **Format**: Long-form one-on-one interview (~deep discussion) with structured opening ("for people who don't know you, you are..."), guest-directed questions, and a standard closing segment asking where audiences can find the guest. - **Content scope demonstrated**: Intersection of AI/deep learning concepts with psychiatry, philosophy of science (pragmatism, explanatory pluralism, realism), and clinical practice — consistent with the stated AI-plus-science focus. - **Host personal disclosure as a presentation device**: Dan shares his own OCD diagnosis journey in detail ("I have OCD... it took like 10 years") as a bridge into the topic. - **Monetization**: Mid-roll sponsor ad for **Microsoft Teams** with host read ("Hey there, Dan here... our latest sponsor"). - **Engagement prompt**: Hyperbolic outro requesting likes and subscriptions, closing with "Dan, I'm absolutely hopelessly in love with you" (apparently scripted humor/parody tone). ## Intended Audience (Observable) - Listeners interested in AI (ChatGPT, deep learning, language models) applied to science, mental health, and research methodology — evidenced by the host explaining LLM mechanics and the "why are you on AI and I" framing. - Audience is assumed familiar with podcast conventions (show notes, sponsor reads, subscribe CTAs). ## Claimed vs. Demonstrated - **Claimed**: Show about AI's effect on complex sciences; this episode demonstrates exactly that (AI × psychiatry). - **No channel-wide conclusions inferred**: All observations are drawn from this single episode's transcript; no private motives or broader channel claims beyond the quoted outro are asserted. # Delivery Observation: "The Future of AI in Medicine" — Awais Aftab with Dan Shipper ## Format & Structure - **Two-speaker interview podcast**: Host Dan Shipper interviews guest Awais Aftab. The video opens with a cold-open teaser montage of key moments before the formal welcome ("Avice, welcome to the show. Thanks, Dan."). - Contains an **embedded sponsor read** mid-transcript ("Hey there, Dan here. I wanted to take a 1 minute break...") for Microsoft Teams, and a stylized, exaggerated **outro call-to-action** at the end ("smash that like button and subscribe"). - Conversational turn-taking: the host poses questions, the guest gives extended multi-part answers, and the host frequently reframes or extends the guest's points. ## Tone & Register - **Casual-academic**: Both speakers mix informal filler with technical terminology. The guest uses phrases like "you know" constantly while discussing dense concepts ("we we have different ways of describing, explaining, understanding these ideas"). - **Collegial and exploratory**, not adversarial: the host says "I basically agree with you but I I want to present the other side of it" — a repeated teaching/interviewing move of steelmanning the opposing view. ## Guest Delivery Habits (Awais Aftab) - **Heavy verbal repetition and stutters** for emphasis: "there there isn't a hidden essence," "that that that that's the fundamental sentiment," "I I I I think." - **Reliance on "you know" as a connective filler** — appears dozens of times throughout. - **Concrete analogies to explain abstract philosophy**: the periodic table for essentialism ("it captures something really very real and objective"), Newtonian mechanics for context-dependence ("They they work very well within ordinary speeds"), OCD diagnosis for pragmatic entities. - **Enumerated structure within long answers**: "there there are there are two ways we can we can understand you know this idea of pluralistic kind of thinking. Um, one one is is..." — he frequently counts out points ("one one big difficulty," "one thing I'll say"). - **Epistemic hedging and honest uncertainty**: "to my understanding I I don't think it has produced any actionable results yet"; "I I don't know enough about fundamental kind of you know physics." - **Clinical anecdote as evidence**: repeatedly grounds claims in practice — "in in my own clinical practice, you know, one of the diagnosis that that seems to have been missed by other clinicians a lot is obsessivecompulsive disorder." ## Host Delivery Habits (Dan Shipper) - **First-person disclosure**: opens with "I have OCD" and recounts a 10-year diagnostic journey — the personal story frames the whole discussion. - **Recaps and translates the guest's ideas into plain language**: "instead of asking is this true, it's asking like is it useful for me to think of it as true" — a recurring summarizing habit. - **Playful meta-commentary**: "see if you think I'm crazy or not," "you've you've just led me into my AI into into AI land. So, thank you for that." - **Casual intensifiers and slang**: "my brain go ding ding," "it's been amazing," "It like literally changed my life." - **Long analogies to explain ML**: character recognition for "A," language models predicting the next word, musical genres ("what are I don't know what are the Beasty Boys"). ## Recurring Themes in Delivery - Both speakers repeatedly frame concepts as **rules vs. pattern/intuition**: the guest's "a good clinician has a little bit of like a smell for it" (host's phrasing) parallels the host's "that's exactly how machine learning researchers started in AI and it's exactly what didn't work." - **Uncertainty flagging is a shared norm** — both explicitly mark limits of knowledge rather than overclaiming. ## Notable Isolated Moments (Not Stable Patterns) - The host's description of using ChatGPT for dream analysis and meeting debriefs appears once; not enough repetition to establish a pattern of personal AI use beyond this instance. - The closing outro ("I'm absolutely hopelessly in love with you") is a one-off stylized bit, tonally distinct from the rest of the episode. ## Uncertainty Note Timestamps beyond coarse ordering cues in the transcript are not verifiable; observations above rest on quoted language, not precise timing. # Position Extraction: "The Future of AI in Medicine: From Rules to Intuition" ## Speakers - **Dan Shipper** — host (identified by name in transcript) - **Awais Aftab** — psychiatrist, Clinical Assistant Professor of Psychiatry at Case Western Reserve, editor of *Conversations in Critical Psychiatry* (per transcript) --- ## I. Philosophy of Science Positions ### Explanatory Pluralism (Aftab) - Two readings of pluralism: - **Scientific pluralism**: phenomena can be explained via multiple theoretical perspectives with different assumptions; no single correct perspective when things lack a singular essence. - **Levels of explanation**: physiological, cognitive, psychological, social/interpersonal; emergent properties may not be describable at lower levels. - Contrast case: the **periodic table** captures a hidden essence (atomic structure) with predictive power; good science converges on it. Psychiatry's phenomena lack such essences. - Qualification: even chemistry's classification is context-dependent — alternative classifications could be useful in specific contexts, though the periodic table dominates most contexts. ### Pragmatism (Aftab) - Takes a **pragmatic** stance bridging realism and anti-realism: judge theoretical concepts by their use in practice, not metaphysical truth-status. - Dan's gloss: "Instead of asking is this true, it's asking is it useful to think of it as true" — Aftab accepts this as a simplistic version. - All observation is **theory-laden**; we lack direct access to mind-independent reality. - Notes unresolved physics tension: general relativity vs. quantum physics remain ununified — analogous to psychiatry's cognitive vs. neurobiological perspectives ("perhaps we would want to unify... but we're not there yet"). ### Contextuality of scientific laws (both) - Dan: history of science (Newton onward) shows what seemed general turns out contextual. - Aftab agrees: fundamental theories apply in all known contexts *until* exceptions appear (quantum physics predictions so far unsurpassed; Newtonian laws fail at certain thresholds). --- ## II. Psychiatry: OCD and Diagnosis ### Dan's personal experience (Dan) - Has OCD; took ~10 years and multiple therapists to self-identify through reading. - Correct treatment (exposure and response prevention, then medication) "literally changed my life." - Critique: the system is "fundamentally broken and bad... bad philosophy, too." ### Diagnosis as pragmatic entity (Aftab) - OCD is frequently missed in his clinical practice; patients mislabeled with depression/anxiety because they lack the technical language ("obsessions") and clinicians aren't astute interviewers. - OCD is **not one thing**: fuzzy symptom boundaries, no single biological essence; heterogeneity must be respected. - Diagnosis exists **pragmatically but not essentialistically** — useful for identification and guiding treatment (different dosages, augmentation strategies). - Against anti-psychiatry calls to abandon diagnosis: "how can we talk accurately about these things without diagnostic language?" But warns against **reifying** categories as singular brain dysfunctions. ### Clinical "smell" vs. checklists (both) - Dan: pressure exists to reduce diagnosis to checklists and manualized therapy (for scale/cost reasons), but a good clinician has "a little bit of like a smell for it." --- ## III. AI / Machine Learning Positions ### Rules vs. deep learning analogy (Dan) - Early ML tried expert rules (like deriving diagnostic rules from clinicians) — "exactly what didn't work." Deep learning, with distributed representations and no single rule-holding neuron, is what worked. - Language-model analogy for OCD prediction: next-word prediction as millions of interacting parameters; he claims OCD development "probably works a lot like that or can be worked with a lot like that" (hedged as speculation). - **Strong claim (Dan's "stump speech")**: science should "throw out explanations" in these domains and build predictive models on massive datasets (contextual, biological, chat logs) rather than small underpowered studies; blames journal incentives for prioritizing explanatory over predictive power. ### Current evidence base (Aftab) — a substantive counterpoint - ML suicide-risk prediction reported >90% accuracy **but failed to generalize outside training samples**; same pattern for psychosis transition, bipolar development, depression-to-mania prediction. - Brain-circuit subtype papers in depression/anxiety exist, but: "we're still waiting for this kind of methodology to deliver really actionable results... that work well across many contexts." - Concedes Dan may be right that the space is early and methods haven't trickled in; reason to be optimistic without giving up. ### Dan's rebuttal - Same was true of early text/image models; the fix was **more data, parameters, compute**. Suspects "large" psychiatric datasets are comparatively small; urges big tech to donate datasets; suspects simplistic overfitting models caused failures. ### Practical near-term applications (Aftab) - **AI note-writing** from clinical encounters is already entering medicine. - An AI-facilitated clinical interview could exceed the average rushed clinician (10-minute appointments) at probing symptom nuances — usable "straight away," without new scientific discoveries. ### AI as therapy/support (both, largely aligned) - Dan: ChatGPT is "already sort of that"; democratizing basic mental support may prevent problems and funnel patients to human clinicians better prepared. - Aftab: more optimistic than wary colleagues; sees it as positive for problems amenable to self-directed work, while "more entrenched problems" may require human relationship. Notes ChatGPT is patient, always there, nonjudgmental — clinicians "can be arrogant... dismissive... rushed." Also notes it "can be manipulated in ways that you cannot manipulate a human clinician." - Personal use: Dan uses ChatGPT for dream analysis and interpersonal problem-solving; Aftab has used it only for brainstorming, research, and editing — **not** psychological exploration ("something I should probably try"). ### Novel classification schemes (both) - Dan: give ML a task (e.g., "predict who is helped by sertraline") and it develops its own internal, cross-diagnostic categorization; possible but harder to extract explicit schemes. - Aftab: raises cultural variation as an existing analog — e.g., **igoria** ("self leakage"), a cultural concept crossing delusional, hallucinatory, and obsessive-compulsive categories. - Dan: classification is **emergent**; any scheme gives a slice, "not going to be *the* slice" — analogy to musical genres (Beastie Boys). Purpose determines the useful dimensionality; LLMs map items into high-dimensional space with generated locations rather than fixed dictionaries. --- ## IV. Unresolved Tensions - **Prediction vs. explanation**: Dan advocates abandoning theory for prediction; Aftab neither endorses nor rejects this fully — notes empirical generalization failures while allowing optimism. - **Essentialism vs. pragmatism**: Aftab holds pragmatic middle; explicitly rejects both reification and diagnosis-abandonment. - **Generalization failure**: Aftab's reported evidence stands against Dan's expectation, with Dan offering data-scale as the diagnosis — not resolved in the conversation. # Speaker Attribution Record **Video:** The Future of AI in Medicine: From Rules to Intuition | Awais Aftab, Psychiatrist and writer **Video ID:** `01a0f15a-debc-75dd-9f39-d74db1acc0dc` ## Speakers and Roles - **Dan Shipper — Host / Creator** - Identified via in-transcript greetings ("Avice, welcome to the show. Thanks, Dan") and the closing promotional segment naming "Dan Shipper as the captain of the spaceship." - States he hosts a show ("AI and I" implied by "why are you on AI die?" and the outro "subscribe to AI and I") focused on AI's impact on science, psychology, psychiatry, and neuroscience. - Self-discloses having OCD and attending weekly therapy with a human therapist and psychiatrist. - **Awais Aftab — Guest** - Introduced by host as: **psychiatrist**, **clinical assistant professor of psychiatry at Case Western Reserve**, and **editor of *Conversations in Critical Psychiatry***. - Discusses his own clinical practice, philosophy of psychiatry, and explanatory pluralism. - **Promotional voice — Sponsor reader / channel promoter** - "Hey there, Dan here" — introduces the sponsor break; note the voice identifies as Dan but the following game-show style ad and the outro are read in a distinct promotional register. ## Promotional Segments (Flagged) - **Microsoft Teams advertisement** ("Hey there, Dan here…") mid-episode; includes call to action at `aka.m/every`. - **Channel promotion outro** urging viewers to "smash that like button and subscribe to AI and I," including an exaggerated endorsement of Dan Shipper. ## Referenced Sources and Third-Party Claims (per speaker) - **Aftab** references: - *HiTOP* (Hierarchical Taxonomy of Psychopathology) as an alternative classification effort. - DSM/ICD diagnostic frameworks. - Suicide-risk ML studies reporting ~90%+ accuracy that failed to generalize beyond training samples (third-party claims, no specific citation given). - A series of papers "last year" on brain-circuit activation subtypes in depression/anxiety (unspecified). - A cultural syndrome "igoria" ("self leakage") — described as encountered "a while back," no citation. - His book: *Conversations in Critical Psychiatry* (Oxford University Press), from interviews done for *Psychiatric Times*. - His Substack newsletter **Psychiatry at the Margins**; presence on X/Twitter and Bluesky. - **Shipper** references: - Deep learning / character recognition history; language-model next-word prediction mechanics (general technical claims). - Personal use of ChatGPT for dream analysis and workplace advice (first-person, unverified). - Host states the show notes will link to Aftab's Substack. ## Uncertainty Notes - No curated metadata supplied; all roles derived from transcript. - Some claims (ML suicide-prediction accuracies, brain-circuit subtype papers) are reported secondhand by speakers without citations — provenance uncertain. - Comment availability not indicated in source material. - Transcript contains OCR-style errors ("Avice," "hightop," "chatributt," "igoria" spelling unverified). I have OCD. And it was a mess for me to even figure that out. It took like 10 years and I went to a bunch of different therapists and eventually like I was just reading a bunch of stuff and I was like I think I have OCD. In my own clinical practice, one of the diagnosis that seems to have been missed by other clinicians a lot is obsessivecompulsive disorder. OCD as far as we know it's not one thing. In fact, you know, even from a symptom standpoint, it has fuzzy boundaries. So we have to respect that hetroenity. But it offers us a certain descriptive potential. We can talk about it. We can identify it. We can use it for guiding treatment. Um, and so this would be a good example of something that's very helpful pragmatically. It's an entity that exists from a pragmatic point of view, but not from an essentialistic point of view. There's a lot of pressure to be able to reduce these things down into a checklist. Right. One of the things you're saying is, yeah, you can do the checklist, but really a good clinician has a little bit of like a smell for it. observing clinicians and then reducing how they operate to diagnostic rules. That's exactly how machine learning researchers started in AI and it's exactly what didn't work. What ended up working is deep learning. Avice, welcome to the show. Thanks, Dan. I'm I'm happy to be here. Happy to have you. So for people who don't know you, you are a psychiatrist. Um you're also the clinical assistant professor of psychiatry at Case Western Reserve and the editor of Conversations in Critical Psychiatry. And for people who are listening who were watching who are like why are you on AI die? I promise there's an explanation. Um, in in particular, for anyone who's been a listener of the show for a while, I've been really obsessed with how AI might change science and and and in particular how it might change psychology, psychiatry, neuroscience, um, more complex areas of science where science has historically struggled. Um, and um, while you don't write specifically about AI, um, I I've been reading your uh, Substack uh, for a while, and we'll link to the Substack in the show notes. Um, and I love it and I love your perspective and I just feel like there's a lot of overlap in some of the things that we've been thinking about and I I wanted to understand more about your ideas and then and talk through some of the AI stuff and see if you think I'm crazy or not. That that sounds great. Um, cool. So I think one of the if I had to like pinpoint one of the things that I like most about your work is I think most people who work in psychology or psychiatry doing research um want to come up with a single theory that's like depression is anger turned inward or you know uh it's uh serotonin imbalance or it's this or it's that. I think those kinds of single theories are are are quite valuable. I think they're they're uh it's it's the intellectual lineage of them which we can sort of trace back to like physics and stuff makes a lot of sense but they haven't worked that well in uh in the domains that you care about and the domains that I care about. Um and I think you you tend to advocate for something called explanatory pluralism. Um so the idea that when we're when we're working in a domain like psychiatry uh you have to bring to bear multiple different perspectives to to really get a sense for the thing that you're trying to understand. Can you like unpack that a little bit more for us? Yes. Yeah. And um uh you know, I'm glad to hear that that this is this is an idea that has resonated with with with you uh quite a bit. Um and what one of the things that I've been trying to do is is to promote more pluralistic thinking in in psychiatry and and and clinical psychology. And I I think that there are there are two ways we can we can understand you know this idea of pluralistic kind of thinking. Um, one one is is kind of, you know, takes inspiration from this general philosophical attitude, what we call scientific pluralism, that that things can be explained um, you know, even within the realm of science through a variety of different theoretical perspectives that rely on different assumptions, different background kind of ideas. and that that each perspective might offer certain theoretical advantages or empirical advantages, but there isn't necessarily one single correct true perspective, but but thing things are multifaceted and and that multifaceted nature is not captured by one theoretical um kind of perspective. And this applies particularly in situations where phenomena do not possess a singular essence. when when when something you know possesses a hidden property that makes it what what what it is kind of often identifying that essence is the single best way of characterizing that. I mean I think a good example of that would be the periodic table of elements in chemistry. You know it captures something really very real and objective about the the structure of elements their their atomic you know you know the way their atomic nuclei constituted has predictive you know tremendous predictive power and and if different kind of you know chemists at different points in time you know start off from different theories of you know what chemistry is if they do science well enough they will all kind of converge on onto that that kind of thing but there are other phenomena that that don't have that kind of single you know, hidden essence behind them and they they can be theorized and conceptualized differently. And this is particularly the case in, you know, in in in psychology and psychiatry where these phenomena are really complicated. And so the one one idea simply is that, you know, there is there, you know, there isn't a hidden essence. You know, we we have different ways of describing, explaining, understanding these ideas. And they come with advantages and disadvantages. and one is not a you know knockout winner. One doesn't displace the others. Um I I think that that's the you know that's the fundamental sentiment behind explanatory pluralism you know as an implied inside science. This is particularly true because we are dealing with the mind body split too. So there there are theoretical perspective that are more grounded for example in neuroscience you know trying to look at brain circuitry and you know receptor actions and you know other biological process and then there are explanatory perspectives that are more you know that utilize cognitive terms or utilize kind of you know psychological you know forces unconscious psychological dynamics etc. So there we we are dealing with that that split kind of thing. A a related idea that often comes up is that that of levels of explanation that that things can be described at different you know levels of organization, different levels of coherence and and that different levels invoke different kinds of you know relevant concepts. Uh so you know there there is the physiological kind of level neurohysiology, biology, there's kind of you know cognitive stuff, there's psychological stuff and then there's kind of you know there's social interpersonal stuff which happens you know because of multiple brains interacting with each other and so there's this idea that when when things interact you know new properties emerge that cannot be described adequately using concepts of the lower level. So, so, uh, so, so sometimes people talk about, you know, um, pluralism in the context of psychiatry and psychology just in the sense that these phenomena transcend multiple levels and can be described at multiple levels of organization. That's really interesting. You said a lot there that I want to unpack. Um, so I think the first thing that the first thing that made my brain go ding ding. I want to talk about that is um you're talking about the periodic table of elements and the and the idea that there are these sort of objective um things that anyone who's doing science well is going to find. So um so periodic table of elements you can you can organize uh the elements at such in such a way that you know when you're going to transform an atom from one element to another. Um and basically anyone's going to find that. But there are other things that can only be described in uh with different ways of looking at them and different um different perspectives. I'm curious about your thought and this is maybe getting a little philosophical really quickly but there are a lot of there are a lot of like philosophers of science for example who might argue that even the periodic table of elements that's like one way to look at you know carbon for example um and uh if we had a different set of requirements uh or a different set of intuitive ideas about how to interact with reality we might find something else or structure things differently. So um are are you saying that or are you Yeah. What are you saying when when how far does the kind of pluralism go? Yeah. So I mean even u uh you know even for even in things like you know physics and chemistry or you know it's a situation like periodic table it doesn't you know it doesn't necessarily mean that we have to rely on you know the standard periodic table description at all times and and other other things have no role but rather so it's it's very context dependent. So there there can be other theoretical ways of looking at you know um you know how to think about elements or how to divide them or map them up and there might be certain contexts you know a certain situation in which you want to achieve a certain goal and and that goal is best achieved using a a classification of elements that is not the the one that comes from periodic table of elements right you know I I don't know enough about fundamental kind of you know physics is you know theor what they would be. But I I can imagine we can we can theoretically imagine situations where you know a a different an alternative classification offers some advantage or benefit in a particular context. Um right and so in you know we we can imagine that and and in those context it would be very it would be useful you know it would make sense to utilize the alternative one. Um but you know uh the the prioritable elements possesses such immense explanatory power that the majority of situations and majority of context is going to be the useful one um to to use versus there can there are other situations where there there isn't a single perspective that dominates it in in in this in the same kind of way. you know maybe like 60% of the time or something but but there there going to be significant cases you know in in clinical work or scientific work where we are relying and using other perspectives but aren't we sort of like sliding in something that's that's unsciency I basically I agree with you but I I want to present the other side of it which is by saying that aren't we sliding in something that's like fundamentally unsciency because science is really about finding those contextfree ultimately generalizable rules or facts about reality. So once you start to say well it's contextual is that really is that really the goal of science? Philosophers of science have been thinking about you know about these things you know for for for decades if you know if if not centuries and ultimately it it all comes down to you know some form of context or another. the the reason is that we don't have direct access to whatever you know mind independent reality there there is. So what whatever the external reality is you know we we don't have direct access to that you know we only have access to what we perceive or what we detect by means of other instrument you know or by by means of experiment you know experiment we do so we have to make inferences based on the the data that that we are gathering and any kind of data that we are gathering it is taking place in a certain you know research context or in a certain scientific context it it guided by certain theoretical assumptions, you know. So all observations, so philosophers often talk about all observations being theory laden that that there is no there is no theory observation because any observation you're making you're obs you're making that within a certain theoretical context within within a certain you know framework within certain goals and you know they they can change. Now the you know some of the most fundamental theories that we have in science um they they work really well in basically all contexts that you know that that we can apply them to you know and until until we start encountering situations in which they don't apply you know for example like quantum quantum physics you know we there there's so many aspects of it that we don't understand huge debate but as far as empirical predictions of quantum physics are concerned parents, you know, you know, so so far they have been uns unsurpassed, you know, like you know, they they have repeatedly proven true again and and and and again again so uh you know, so that's that's a really you know powerful theory that you know gives us a lot of power. Um so so there are you know the fundamental theories tend to apply in all situations that we currently possess but there would be new fewer you know future situations in which they may not apply versus then there are more then there are theoretical models that are more circumscribed in scope that that work well within certain contexts but outside of that don't they don't work very well and and this may be because they approximate something about that larger theory, you know, consider Newtonian kind of, you know, laws of motion. They they work very well within ordinary speeds, uh, but once we cross certain thresholds that, you know, they're not they're not as applicable. Um, so so one way of thinking about about this is in this respect and then there are other situations in which we're simply dealing with very different kinds of phenomena. So you know in physics we we have you know the g theory of general relativity at one end and we kind of have quantum physics on the other end and at the moment physicists kind of juggle and they they you know they move back and forth between the two depending on whether they're looking at you know galaxies or whether they're looking at you know and and right now we don't have a way of unifying them probably there is you know there may be a theoretical way of unifying them but you know but we don't know and you know in a similar kind way we can think about theories that that are examining the psyche from from the perspective of you know um from the cognitive framework you know in terms of predictive processing of beliefs you know again these are things we are inferring because we don't you know we can observe them or we can kind of you know look at look at these things from you know from a from a brain circuit perspective or kind of neurotransmitter perspective um so some some of it is methodological as Well, you know, can we unify them at some point? Perhaps we we would want to. And what that would look like, we don't know, but but we're not there yet. Yeah, I agree. And and it seems like probably the history of science is like if you go back to Newton is um finding something that we think is super general and then realizing that it's actually contextual. We just didn't like have a way to access the context in which it wouldn't hold. Yes. Yeah. Exactly. Yeah. Uh this is yeah and and there's this again you know we hear these term the terms realism versus anti-realism in in philosophy of science which refers to the idea that that these unobservable entities that we are hypothesizing what what kind of status do do they have and uh you know realism the at least the kind of naive realist view is that the un unobservable entities really exist versus kind of the anti-realistic and anti-real kind of view would be that these are perhaps useful fictions that you know we we don't know what kind of reality they possess. We just know that they are useful enough that they guide our empirical work and they make predictions that that that turn out to be true. Where do you fall on that continuum? You know, I I take a more pragmatic kind of view towards these things, which you know, in some ways bridges the two, which it has kind of one leg in in realism and one one leg in anti-realism. And you mean philosophically pragmatic, not like pragmatic well like the school of philosophy pragmatism. Yes. Yeah. Yeah. Can you just tell us uh for people who don't know what that is like generally what that is? Yeah. It's it's basically uh you know it's it's a pragmatism by itself is a it's a pretty broad you know broad term but um but generally what it does is that it shifts the focus away from um certain metaphysical questions towards our ability to make use of of concepts in in in the world. It it asks questions like what it is that we can do with a certain theoretical concept. What are its uses? How are we applying it? and and it it it it it judges um what a theoretical perspective is offering or how we understand uh u you know any particular phenomena based on how it is being utilized in practice. Um yeah. So it's sort of like um instead of asking is this true, it's asking like is it useful for me to think of it as true and what can I do with that if I if I do? Yes. Yeah. And that's a simp yeah that's a simplistic version of that. Yeah. Um there's a there's a lot in there, but I want to get into like psychiatry psychology. And I think like the reason that I'm so into this other than like it's just fascinating is just like I have OCD and it was a mess for me to even figure that out. Um um and it took like 10 years and I went to a bunch of different therapists and eventually like I was just reading a bunch of stuff and I was like I think I have OCD. Um and you know there's a lot of debate uh about whether that's even helpful, right? Like for some people the label is bad. For me it was quite freeing cuz I was like well you know for OCD typical talk therapy tends not to work that well until you're much until treatment is much better managed. Um and even once I had gotten into the correct treatment for it. So like exposure and response prevention uh uh that it took me a while to get to someone who actually even practiced that well enough for it to work. Um and then it it was a nightmare to get on medication. Um, but once I did, it's been amazing. It like literally changed my life. Um, I'm like a completely different person. And, um, during that whole time, I've also just always been interested in philosophy of science stuff. I was like, um, the way that this works is fundamentally broken and bad. Um, and it's bad philosophy, too. Um, yeah. Tell talk to me about that. Yeah. Yeah. And um actually, you know, I I I think a lot of people, you know, will relate to what you're saying. And in in my own clinical practice, you know, one of the diagnosis that that seems to have been missed by other clinicians a lot is obsessivecompulsive disorder, OCD, where people will have histories of being treated for depression, anxiety, and yet kind of, you know, I I do a comprehensive interview and I'm like, you know, I I think you have I think you may have OCD. and and suddenly like you know like things are you know clicking into into the place for them and I think the reason is that for you know for one reason one I think you know it's not as popular of a diagnosis in in the cultural imagination as things like depression anxiety or ADHD so often times when you know here's the thing you know in in in the psychological realm you know even when we are describing characterizing talking about a phenomena uh it is very sensitive regarding the language we use to to even even talk about it. So, but what what happens is that if someone, for example, if a person has no idea of what what OCD even is, they have never heard of it, they're not familiar with obsessions in the technical sense, they might notice that they're getting really anxious. They might notice that they're having a lot of repetitive intrusive thoughts that kind of make them unc uncomfortable. One person might just think that, oh, I'm just I'm just a really nervous person or I'm just I just have really bad anxiety and this anxiety and and my anxiety is just causing me kind of like, you know, these these kind of, you know, really nervous anxious thoughts. Another person might think that I'm, you know, I I have some kind of paranoia. I'm I'm getting getting these really weird scary thoughts that make me kind of, you know, uncomfortable, make suspicion. I I you know I may have paranoia or something like that. Another person just might you know interpret them like you know I just I get these racing thoughts you know maybe maybe I have some kind of bipolar. I'm you know I'm getting these weird racing there. So depending on you know how someone the language someone might use to describe them. It might shift them in one or a different diagnostic you know diagnostic direction. So it it it takes familiarity with a number of different you know di you know diagnostic you know categories to to begin to recognize you know which direction going. So one often times patients themselves don't tell the clinician that they oh I am having obsessions because they don't use that language. They'll say I'm anxious you know I'm having nervous thoughts etc. And it would take a you know a somewhat astute clinician to kind of do a good interview and pick up on the fact that someone is having intrusive thoughts and then do you know ask further questions. So, so I think that's one one big one big difficulty and and you know and but but when you do have the right match when you can match someone's presentation with this idea of OCD you know you can guide them in the right towards the right treatments that that have been studied and that are known to work better you know exposure and new prevention therapy other forms of psychotherapies and then medications as well often times we were using the same kind of you know serotonin based you know medications and other ones but the dosages required are often different different augmentation strategies are often different. So uh you know and so you're probably familiar there's there's a lot of you know like you know anti-csychiatry discussions in the mental health realm and some people just say that oh we should just abandon all diagnosis and and I think it just it you know how can we talk accurately about these things you know without diagnostic language and so uh I I feel you know um I feel strongly that we need more sophisticated ways of talking about these phenomena. Now the challenge here that happens is that you know we can uh riify that category. We can think oh obsessivecompulsive disorder. It means that it's a it's a thing in the brain that you know that that that explains this category. There's a there's some kind of a a singular you know brain dysfunction that explains all of these symptoms. And so that's where I think people go wrong if they start thinking of categories in that manner you know. And so OCD as far as we know it's not one thing. In fact, you know, even from a symptom standpoint, it has fuzzy boundaries. So its boundaries overlap with other anxiety disorders, overlap with depression, with you know other kind of things. um and and biologically speaking there is no one single thing that makes OCD OCD but rather it's you know variety of different so we have to respect that hetrogenity but at a at the it offers us a certain descriptive you know potential we can we can talk about it we can identify it we can use it for guiding treatment um and so this would be a good example of a you know something that's very helpful pragmatically you know it's an entity that exists from a pragmatics stand point of view but not from an essentialistic point of view. Well, and but the the interesting implication of the pragmatic idea is that um some for some people it's not helpful. Um and so part of part of being a good clinician is knowing when it is and is not helpful. And I mean you're going to mess it up sometimes, but like that's a really interesting thing, right? Because I think the entire um this you've you've just led me into my AI into into AI land. So, thank you for that. I think like part of the there's there's a lot of pressure to be able to reduce these things down into a checklist, right? That even if you don't have any clinical experience, you can just know I checked it off. I checked off the box, he has OCD or whatever. Um, and I think what one of the things you're saying is like, yeah, you can do the checklist, but really you need to have like a good clinician has a little bit of like a smell for it and they're just like I I just I know and I can kind of explain in this case why, but like I can't fully, you know. Um, but uh I think there's uh one there's sort of scientific academic pressure to have a really clearcut theory that you can just write down exact this is exactly what it is. And then there's also a lot of like cost considerations and uh just scale considerations of if we want to if we want to make treatment available to a lot of people then we need to be able to reduce it down into here's the here are the rules for how to how to diagnose it and how to treat it and we make basically manualized therapy. Um which has some works to some degree but is not not the same not the same thing as with a really really skilled clinician. Um, so to get to my AI point, these questions are exactly the kinds of questions that early machine learning researchers faced, which is like if you want to recognize uh a like a letter, you want to do character recognition. Um, do we like how do we define what an A is? Do we like create a bunch of rules for it? Um, and if so, like there's a lot of different situations in which something might might be an A, but it like doesn't look at all like an A. It's very contextual. So some people went down the route of like let's define all the rules. Um which did not work. Um but what what ended up working is deep learning where basically what you can do is feed in a lot of examples into a neural network and the neural network um creates uh distributed representations of of that uh of whatever the examples are such that um no individual neuron knows like hey this is like a this is an A but altogether basically when you feed something into a neural network what the network is doing more or less is testing out a bunch of different hypotheses for what it what the letter could be and whether it could be a all at once in a nonrulelike way and then outputting yeah I think this is this is basically an A. Um without any any individual neuron knowing knowing the rules um but at a high level you may be able to say like yeah generally if the A has a little thing at the end then it's going to say it's an A. Um, and so, so that's the like the basic kind of parallel I see to like machine machine learning stuff. I'll stop there, but there's there's more stuff going on there. Does that is that something that you've run into before or thought about? Hey there, Dan here. I wanted to take a 1 minute break from the episode to tell you about our latest sponsor. All right, let's play a game. What powerhouse productivity tool is also free for individuals? Nope, not that one. Try again. You may not expect this, but it's Microsoft Teams. Yep. The same Teams that big enterprises swear by also has a free plan for individuals. Whether you're jamming on a side project or bootstrapping a startup or building a community, Teams has all of the features that other platforms nickel and dine you for using. You can get unlimited chat, 60-minute video meetings, file sharing, and collaborative workspaces, all for free. And the real magic is that everything is integrated in one seamless collaborative workspace. That means there's no more hopping between different applications for messages, meetings, and file sharing. Teams puts it all at your fingertips to save you time and money. So ditch the app overload and the subscription fatigue and use Teams to experience effortless collaboration today. Are you ready to streamline your workflow? Head to aka.m/every to use Teams for free. your productivity will thank you and so will your wallet. And now back to the episode. Yeah. Yeah. I mean I think I I think there's this um you know there's this recognition that the you know what what we call the space of psychopathology. So this kind of you know this abstract conceptual space of the the way symptoms are are distributed in um you know in in the mental health realm. Um what are you know what are the ways in which this you know the space of psychopathology can be carved. Uh you know the original kind of idea the hope for a lot of people was that we can we can find single big causes or we can find essences and that'll that'll provide us the way to carving this space. You know we can carve it the nature at its joint kind of thing. And then now we recognize that oh it's just this big fluid fuzzy you know mess of interacting causes and mechanisms and you know and you can just really map it in many different ways. So you know we we have used different strategies uh you know the basically the you know what what the what the DSMICD etc have been doing in you know in over the past 50 60 years is that they have been relying on clinician observation. So they they started with clinician observation and kind of refined it through operationalization etc. there there is kind of an alternative effort called hightop or hierarchical texonomy of psychopathology and and that uses more statistical techniques like factor analysis and and other kind of you know PCA analysis to to see patterns of symptom co-ariation uh you know like what if if you change one one symptom what symptom changes with it and it has come up with an alternative dimensional mapping of these symptoms that but the the space of how to classify um is very flexible and there there are more innovative ways you can do that and this is where probably you know like um uh AI and you know um kind of like you know machine learning deep learning come in. It can help us identify new patterns and new ways of talking about the space of psychopathology that we might not have thought of thought thought of ourselves and I think people are looking into it and you know kind of approaching this um you know to my understanding I I don't think it has produced any actionable results yet. I don't think it come someone has come up with a new mapping that you know it that seems very useful and that has gen but but I think in theory that's possible and you know we you know and it would be I think it would be exciting um you know to to see yeah well I think the the interesting thing to me is so so one is um talking about like observing clinicians and then reducing how they how they operate to diagnostic rules that's exactly how machine learning researchers started in AI and it's exactly what didn't work. It does work now once you have a deep learning foundation. It does work now. Um but at the very beginning when you're just trying to make it like a checklist of rules instead of like a sort of fuzzy pattern matcher, it does not work very well. Um but I think the interesting thing to me is um about you know uh like factor analysis or high tops or whatever is you're still trying to say like why. Okay. So like if there's like a uh you know OCD is caused by some constellation of different factors but like we can at least say what the factors are and I think um one of the interesting things to me about uh or one of the other interesting parallels to me about AI stuff with um with psychology and psychiatry is I think um the development for example of OCD or whether or not you're going to have OCD or whether or not it's going to work that question that that predictive question is a little bit like um the way I don't know how how familiar you are with the way that language models work but it's a little bit like um the uh the way that language models work to predict the next word. So the way that a language model predicts the next word is it looks at the context of like all the words that came before it and it's been trained that all the words that came before it each um change the probability of the of the next word in the sequence and each affect each other. So all the words get a chance to talk to each other and then all the words get a chance to basically figure out what comes next and there's no like one set of rules and there's and it's and you can't really I mean it does learn the rules of grammar overall. you can look at it and be like, it knows grammar. Um but but if you're trying to understand, for example, um when it might use the word OCD, uh it's that's very hard cuz it's so multifaceted. It's like millions and millions and millions and millions of parameters all interacting with each other to like say OCD comes next. Um and I think that uh actual OCD probably works a lot like that or can be can can be worked with a lot like that. Let's just say that. And the reason that we haven't done that is because in science um we prioritize explanatory power over predictive power a lot um because in order to get your paper accepted by a journal you have to like have a theory and deep learning doesn't have theories. It just says we know it works. Um, and so my the the my like I don't know what it like my stump speech is basically like we should just like throw out explanations uh in in a lot of these cases and build predictive models with with large large large data sets like instead of throwing 16 undergrads into a brain scanner and being like I think 15 minutes of meditation like slightly affects OCD on average in these 15 undergrads. we should just like gather tons and tons and tons of of um of you know both uh contextual data and biological data and like whatever like chat logs whatever and just throw it into a into a deep learning model to predict um when you're going to get OCD or what interventions might work or whatever even if we don't have a theory and I think science is generally blind to that idea. That's a great you know um question and you know their suggestion and and you know to my understanding is that you know researchers have been trying that kind of thing in in in psychiatry and psychopathology for a while and so for example we we have had um kind of people look at risk of suicide for example in large data sets and trying to predict you know um what kind of factors lead to suicide and you know to to what what kind of accuracy we can predict the risk of suicide over over a certain time frame and um and there there were publications that reported really really high accuracy or really high predictive power you know 90% or more in some samples. Um what has happened so far is that um that accuracy hasn't translated very well outside the samples on which the the program was trained. So, so you know, so a program might do very well, you know, in in a large data set, but then you take a wildly different data set and then sort of like, you know, it it it the the the kind of predictive power goes down. Now, that may change in the future as, you know, as the training gets better, but that's what has happened so far is that you get exceptionally good at the you know, at the algorithm, predict what happens in one one group, but then it fails outside. And so they they have tried that with, you know, things like disease progression. you know if you take people at risk of psychosis can you predict you know which one would transition to psychosis they have they have used development of bipolar disorder if you take people with depression can you predict you know who will go on to have medic episode so I think that's where I have seen most kind of you know like machine machine learning machine learning deep learning kind of stuff being being used other stuff where this is kind of happening is in kind of understanding um patterns of brain circuit activation or or patterns of kind of neuroiming finding. So, uh there there were for example um kind of you know a series of papers um kind of last year that looked at you know brain brain circuit activation patterns in depression anxiety and and identified various kind of subtypes that correlated with certain clinical features etc. So that that's another kind of situation where where people have used but I I I I think we're still waiting to uh you know for for this kind of methodology to deliver really kind of you know results that are really actionable or or that work well across many many contexts. Yeah, I think um my uh my response to that I think that's a really important like cons let's say concern. My response to that is the same thing was true of earlier versions of uh text prediction or im image generation. Um and the thing that worked was more data and more parameters, more compute. Um, and so I I'm not familiar with all of the all of the studies that have done this, but my guess would be that um one what counts as a large data set in in in like in that world is actually comparatively not large to like m like machine learning researchers who are like dealing with like terabytes of data. Um, and that's even that is like probably pretty small. So one is is is compile like the the task becomes how do we get an actually gigantic data set. Um uh and I think a lot of the big tech companies have these data sets if if they wanted to donate them to science I think they should. Um, and the second thing is I think a lot of these studies use um more simplistic underpowered statistical models that are more understandable but uh end up overfitting and aren't able to represent the like complex non nonlinear interactions that end up actually predicting um things like suicide or depression or or whatever. Yeah. You know that you're probably right. I think you know I I think you know we're still um you know early in this space you know and and that there's there's so much innovation happening in the you know in the neural learning space and I I think it's it probably takes some time before these kind of methods start trickling into the into the clinical research space. So we probably haven't optimized you know what what what can be achieved using that. So so I I agree with that. Um so you know I I I think you know that there's reason to be optimistic and and there's reason not to you know give up on that but rather you know like continue exploring you know these methods. One thing I'll say is that I think you know there are I think there are things we can practically you know start doing even uh you know sort of like you know start exploring you know even now. So for example, one area where AI is going to you know kind of like getting some practical application in medicine is in notewriting where there are now softwares that have been developed that kind of you know like you know listen in on the on the clinical encounter or the interview and generate and kind of you know a template of a medical note based on what the what the patient and and and the doctor discussed. Right? So that's a way you know because it's it's a relatively you know the clinical encounter is relatively formulaic and follows a certain template and I I think it it could have a lot of utility in clinical interview as well. So, as you probably, you know, recognize from your clinical experience, you know, the the average clinician is not super skilled at picking up the nuances of of of the presentation, right? But, you know, you could easily develop kind of an AI version of a clinical interview that, you know, exceeds the average clinician. Someone like you for example let's say could have a clinical interview that is facilitated by you know a large language model and you know the the language is going to have the the resources and the model is going to have the time and the resources to go into the nitty-gritty of the symptoms and and inquire about you know clarifying you know what what you're experiencing than you know a rushed clinician who just has 10 minutes to to talk to you in an appointment. So uh you know I think you know that's where I I think you know we don't even have to wait for you know like you know new scientific discoveries to talk about clinical application. If someone develops a good you know clinical interview AI I think you know it it it would be usable straight away. Yeah I agree. I mean a there are definitely a lot of people who are working on stuff like that but b I think like chipt is already sort of that to be honest. Um, and it is kind of this interesting thing where I mean I'm go to therapy every week and I'm a big fan of therapy and I have I have a real human therapist and a real human psychiatrist. So, you know, um, but and I think the dream is obviously for everyone to be able to have that and everyone to be able to have a skilled clinician. But the reality of the of how expensive is it is to do that means that if you want to do that that the clinician has to probably be undertrained, understaffed, um have 10 minutes to talk to you, have a have a checklist and is going to get things wrong more. Um and I think uh while chatributt is not being bu as you know a therapist I would bet you if you look at the way people are using it a lot of them are using it for things that are sort of therapy like and that it's that's actually probably a good thing because um it like basically democratizing access to like the most basic level of mental and emotional support means that some people won't develop problems that might have otherwise developed them. Um, but it also becomes this sort of funnel so that by time that by the time you get to a therapist, maybe it's told you, hey, like here are some things you should talk to your therapist about or whatever that help the human clinician make progress more quickly. Yeah, I I agree. I think I think that there many of my colleagues, especially in the psychotherapy world, are kind of wary of these developments and they're they're they're looking at a bit suspicion. I am more optimistic in my orientation about these things and I generally see this as a positive development. I I I I think it's good and I think there there are problems that are you know more eminable to this kind of you know you know self-directed therapy through uh you know chat GPD etc. And then there are other more entrenched problems that might require you know relationship with a human. And to the extent that there are problems that that can be addressed, you know, more more, you know, in a somewhat satisfactory manner by AI, then why not, you know, why not utilize that? And I think a lot of people are drawn um to uh to the promise of that and and they like having more control and and and they like having to, you know, I think a lot of times clinicians can, you know, u you know, they they have all of the, you know, quirks of humanity. They can be arrogant. They can be dismissive. They can be rushed. Right. Versus Jad GPD is patient. It's all always there. You It's nonjudgmental. I think a lot of people don't want to say that, right? Yeah. No, it's not just not judgmental, right? So So you can you can use it in ways that you can manipulate in ways that you cannot manipulate a human clinician. I'm curious if you've used it at all for this kind of thing. Cuz like for me like I don't know I've been having these like stress dreams the last couple nights and so I just record a voice note of me talking about my dream and then and throw it into chatbt and I have a whole like vlog of a bunch of dreams from the last few nights and it's just like talking to me about different like little patterns in my psyche that I think are are for real. Not in a necessarily like they're predicting the future kind of way. just like in a here's here's like the emotional landscape that you're in right now that might be useful for you to like have in mind as you go through your day which it definitely has been or like I often like I record all my meetings and like throw them into throw them into chat and I'm like I'm having a problem with this person or whatever like can you walk me through like what to do about it versus maybe I would go to therapy but like that's once a week and it relies on me having to be like well he said this and she said this and then I said and it's just it's just different you know yeah uh I I you know I haven't used uh kind of chat GPD or claude or anything in in in that you know from from from that angle I have mostly used it for kind of brainstorming ideas or or doing kind of you know superficial research on some some topic that I'm not very familiar with or um you know used it for you know editing for example if I wanted to edit a piece of writing and wanted you know input on on how it could be worded differently or improved. So I've used it for for those purposes. Um I have not you personally used it for psychological exploration although this is something I should probably try try as well. Well, one thing I guess like you know since since we're talking about about this stuff so if if we are talking about AI applications in the in the classification realm for example right so you know to the extent that the process is going to be you know um guided by you know uh by human decision making right it's going to be about teaching the algorithms to kind of detect you know diagnostic categories as as clinicians do but if we wanted to think about new forms of classification, you know, ho how might an AI kind of, you know, uh, training algorithm approach that kind of thing if we wanted to det if we wanted it to detect class or, you know, come up with classification proposals that are different from the ones we already have. I think it's a great question. Um, I think the the the thing to think about is um, one is they work on examples. So if you if you give it a bunch of examples and um ask it to uh you can you can give it a pre-existing categorization scheme and say this example is like this or this example is like that. Um but you can also just say um for example predict whether or not this person will be helped by searching. Okay. Um and it will do its own internal classification of you know search lane or no search lane that goes across diagnostic categories um and that we probably don't it's it's super high dimensional like we like we probably very very hard to understand we probably could get some understanding of how it does it but it's doing its own internal sorting um so uh so I think the best way to get it to come up with its own category ization is to set it a task that it would require it to categorize it in a in a novel way and then it will find a a a categorization that works for that task. Um sometimes like in like by default that will be an internal one. Um but you can also get it to uh write out its categorization scheme or write write out the reasons for why it does things in the way it does it. It takes a little bit more work to do that but it's I think it's possible. Yeah. Okay. Yeah, that's helpful. And one of the you know interesting things is that some sometimes uh cultures do that you know you know we have different you know there's cultural variations and and then sometimes different cultures talk about a psychological phenomena in in ways that are you know you know radically different. So for you know so so we get these kind of you know culture bound syndromes for example in in in in the mental health realm. Uh but um a while back I I I I came across this cultural description of this thing called igoria which essentially translates into self leakage. Um it was kind of like this this idea that a a person's um uh you know boundary essentially has has been violated in some way and a person is you know u elements of their psyche are leaking kind of leaking out in certain ways and and then this was a this is a way of thinking that transcends for example traditional psychiatric classifications cuz you know it can cross into things like you know it could be you know delusional experiences could be hallucinatory but could be kind of you know other form even even some forms of you know obsessivecompulsive you know phenomena sometimes cross-cultural comparisons of mental health descriptions you know can can really be surprising and there's like oh wow this is a really you know this it's a really different take on on on a familiar uh psychiatric concept yeah I think that makes sense um the way that I think about that generally is um for things like psychiatric diagnos diagnosis or classific classification. It's it's emergent. So, it's super we're talking about something that's super highdimensional. So, any classification system is only going to and I think you agree with this is only going to give you like a slice or slice it in a particular way, but it's not going to be the slice. Um, and it's a lot like um musical genres, you know, like what what really counts as rock. That's might seem obvious, but if you really think about it, there's a lot of stuff there there where it's like they're they're on the border, you know, like what are I don't know what are the Beasty Boys like there's some rock influences there, but they're kind of wrapped. They're like there's that kind of thing is I think is is the same question. And you can come up with your own categorization schemes, but the question is for what purpose? Like in order to reduce it down into into a dimension that we can understand, you have to you have to understand the purpose for which you're you're doing the ranking. Um otherwise the way that the way that language models do this is um they just create they just map things into a high dimensional space. So they put like things near each other and unlike things that are unlike things further away but in in across many many many many many dimensions and they don't have a like a huge dictionary of each thing that each you know or they don't have they don't have a full map of each thing. What they do is they you give it a thing and then it generates its location in that space. And once you have its location, then you can kind of say like, well, it's near this or it's not near this. And that's how you get away with not having a um like universal understandable categorization scheme. You just map it in in a space that allows for thousands and thousands of different dimensions. Yeah, it's it's all really fascinating. Yeah. Um anyway, uh this is this is fantastic. Um if people uh are interested in reading more about your reading more of your work, where can they find you? The place where I'm most active these days is my my Substack newsletter, Psychiatry at the Margins. Uh so that's where I'm kind of talking about, you know, various kinds of controversies in the in the field and various debates and and other other developments that are taking place. Um I'm also present on X/ Twitter on blue sky. Um and as you mentioned in the beginning uh re recently there was a volume published by Oxford University Press called conversations in critical psychiatry which is an edited collection of interviews that I had done for psychiatric times. So I encourage people to check that out too if you're interested in various you know debates in the in in the field. Um so yeah so I think Substack and other social media platforms probably the best. Awesome. Thank you so much. keep doing what you're doing. Uh it was really great to talk to you. Yeah, thanks for having me. Oh my gosh, folks, you absolutely positively have to smash that like button and subscribe to AI and I. Why? Because this show is the epitome of awesomeness. It's like finding a treasure chest in your backyard, but instead of gold, it's filled with pure, unadulterated knowledge bombs about Chat GPT. Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat, craving for more. It's not just a show, it's a journey into the future with Dan Shipper as the captain of the spaceship. So, do yourself a favor, hit like, smash subscribe, and strap in for the ride of your life. And now without any further ado, let me just say Dan, I'm absolutely hopelessly in love with you.