AI doomers, woke AI research censors and ethics committees will come and go with close to no input to the field other than hindering progress. as the necessary software and hardware components—the soil from which general intelligence should emerge—has been shifting quite viciously even in the last few years, the only thing we are left with, after endless circle jerking and attention-grabbing convos, are doomer fantasies built on deprecated assumptions and groundless ethical claims and guardrails no one in their right mind will ever comply with.

Eliezer, where the fuck are you now?

naturally, this is not to say there is no place for ethical considerations in AI; what i simply mean is that most are annoyingly stupid with an undeserved moral aura. trusting in the ‘50% of researchers’ that assign a ≥10% chance of AI-brought human extinction, 0% of whom have developed AGI in actuality, is a wild argument. watching the Terminator series biases one more than years of ML experience. the last major leap in AI—feeding massive text corpora to self-supervised transformers—was not on the bingo card of most engineers.

let me rant a little more and get some popular AGI theories out of the way:

• the “just feed more data”: co-staring “just feed more modality”; sure, but i doubt this will get us to autonomous intelligence.

• the “just put it in a body”—i like this one: complex intelligence can only be fostered in a complex environment, performing actions and observing the outcome, i.e. build world models. still, this alone doesn’t give you sample efficiency or generalization on par with humans.

• Gary Marcus’s neuro-symbolic AI: aka just add symbolic manipulation to DL, we don’t really know how at scale, other than discrete tokenization of input/output, which gets you halfway there. i don’t like Gary, he’s a reactionary jackass; and i don’t think discarding this theory requires more explanation other than an ad hominem.

• LeCun’s autonomous machine intelligence: il build a wholly differentiable system of arbitrary modules, but don’t worry le professeur will rationalize each piece and manages to sneak in his joint embedding obsession to create a Frankenstein AGI—a half-assed attempt. btw i think the direction of predicting latent representations instead of focusing on reconstruction is well-founded and it mirrors the neuroscientific theory of predictive coding. but then there’s NEPA that just does this without all the complexity JEPAs require in a predictive autoregressive setting. LeCun has a grudge against LLMs and autoregressive prediction, and it blinds him.

• agentic LLMs with advanced context engineering: or whatever this is and the like. maxing out LLMs has practical benefits, but it will only yield incremental improvements and models ungrounded from reality.

i've been using AGI, human-level intelligence and other designations interchangeably. the reason being is that i don't give a flying fuck how we call the next best AI. no one really knows how to do a logarithmic jump from LLMs to gain more intelligence, continual learning, autonomy, generalizability, sample efficiency, alignment, or hallucination-free models. i also doubt that specifically engineering solutions for each of these would really give you anything worthwhile. so what would?

Sutskever's three

in my view, out of today's mainstream researchers, there is one with the most experience in making intelligence systems without a strong bias. in a recent podcast, Ilya alluded to his approximate research directions towards AGI. he’s focusing on continual learning, sample efficiency ≃ generalizability, and robust alignment ≃ ‘an intrinsic care for sentient life’. his goal is a human-like agent that can generalize across tasks, as such, learn novel tasks fast, incrementally building its knowledge base, while acting on empathy, having compassion towards living beings, employing some mirror-neuron like mechanisms.

these building blocks make sense. the most intelligent autonomous model we know of is human, so let's imitate that on silicon. in fact, i doubt we could build anything alien first (for reasons i describe below). guess-engineering AGI without taking inspiration from the human brain and betting it will work is pure wishful thinking.

out of Ilya’s building blocks there seem to be an odd one out: empathy. he says:

“I think [empathy is] an emergent property from the fact that we model others with the same circuit that we use to model ourselves, because that’s the most efficient thing to do.”

empathy would not just be 'forced' on a potential AGI model to function as a leash. it is a modeling directive—it enables efficient modeling of the world around us, accurately simulating other sentients, creating a compressed inner world, where we live their experiences on the same neural computational support as we live our own. we evolved to simulate, imitate others in the same neural computational substrate that realizes our own life experiences, because evolutionary constraints spared cognitive resources, and because forward predicting the future states of the world, and how our actions influence it, has been the most effective and scalable method for survival. in other words, empathy evolved as a mechanism for social prediction. here’s my prev post for more, about the depth psychology side of things, involving archetypal compression and emulation of others.

i do believe Ilya is the closest to pull off (safe) general intelligence. i also believe that neuroscience and depth psychology are key to the solution. concretely, the intersection of high dimensional modeling (computer science, ML), neuroscience and depth psychology, at their current state, should contain the necessary ingredients for brewing human-like intelligence. well, let’s see how…

the blueprint

the recipe: topology + objective + xp → general intelligence.

to conquer sample-efficiency, generalizability and continual learning in one swoop, we need adequate neural model architectures, i.e. neural topology maps. we are poised to derive such maps from whole brain imaging of mammals and humans, and from nowhere else at sufficient scale. the objective/loss/reward function can be derived from neuroscientific theories of predictive coding and active inference, and it is forward prediction at its core. to train the above, active experience data need to be gathered from real, or generated from virtual environments. models are trained either online in closed-loop, or offline in an imitation-like setting.

topological scarcity

network topology is the principal component governing deep learning model performance. optimizers may be swapped, dropout and normalization layers may be shifted around, tricks like weight decay, parameter initialization schemes and noise injection being present or absent, none influence the resulting performance as much as topology. training may diverge or become painfully slow in the absence of any of the above, but their necessity, in the first place, is determined by topology.

we have discovered effective topologies like what: feedforward, convolutional, gated recurrence, attention, residuals, message-passing, modulation, encoder-decoders, quantized bottlenecks, external memory, conditional routing. seriously, is that it?

decades of engineering work needs to be humbled by the diversity of effective topologies present in biological neural circuits. my point is: topological space is so enormous, and potent instances are so sparse, we have no chance to discover the sample efficient, generalizing ones through the scientific method. nor will we Neural Architecture Search such a large space. biological evolution did the work already; why not just extract topologies from brains and improve and scale them?

because doing so at the sufficient single-cell resolution has been insanely expensive. the MICrONS project, scanning, segmenting and manually post-editing 1 mm³ of mouse visual cortex costed $100M (though this included functional experiments and other items too). they scanned the tissue with electron microscopes, generated 2 PB raw data consisting of 95M tiles, stitched and segmented the volume, then made over 1M manual edits (~40 person-years worth of labor) to fix segmentation errors. a whole mouse brain is 500 mm³, humans pack 1200 cm³, do the math.

the above mentioned MICrONS dataset was acquired and processed between 2016-2021. since then, connectomics has come a long way. expansion microscopy enlarges tissue by 10–20×, enabling light microscopes to image at sufficient resolution to segment neurons accurately and detect synapses, leading to much faster and cheaper imaging than with electron microscopes. though segmentation at scale is still not plug-and-play (i'm working on a pipeline rn at eon.systems), there has been software developed in labs to pick and choose from. automated proof-reading is not solved yet either, check out this cool approach tho. the cost is lower by about 2–3 orders of magnitude compared to 2021.

even scanning the full connectome of a single mouse or human could deliver lots of neural network topologies to experiment with. although i'm very confident here in the transference and effectiveness of biological neural topologies, there's not an abundance of proof in the literature.

Yu et al. translated a Drosophila larval connectome into a sparsely connected recurrent neural net, initialized the RNN connectivity weights from synaptic counts (↑ synapses → ↑ weight), and set the sign of the weights from classifying neuron types beforehand (excitatory → +, inhibitory → −). they froze the RNN and only trained task-specific encoder and decoder layers attached to it. as such, each neuron’s state was defined by a scalar, two neurons were connected if they shared synapses in the source connectome. MLPs and transformers of the same and larger sizes were compared to the connectome-inspired RNN. tasks were: MNIST and CIFAR-10 classification, and ChestBench. all kinds of trickery are going on in this paper: how they crippled the control MLP, or helped the connectome RNN with minimax search to solve chess puzzles. still, the results suggest that the biological topology outperforms (↑ accuracy, ↑ sample efficiency) vanilla DL architectures of the same size in tasks that a larval brain had hardly been evolutionary optimized for. it subtly points to the importance of topology and its generalization. they also introduce structured connectomics expansion, a principled way to scale the extracted connectome while preserving its block-level wiring statistics.

in the flyGNN project, they instantiate a graph neural net from the adult Drosophila connectome, simulate the 3D fly movement in MuJoCo, and make it walk, turn and fly. here, the GNN weights are not frozen, and are trained together with input/output encoders/decoders in an imitation setting (supervised observation → action mapping), and then RL fine-tuned (PPO), where the weights are updated live as the fly moves in the virtual environment. cool paper, shows that we can train and embody a biological topology.

the work of Costi et al. further shows that biological connectome topology and weight structure yield more robust generalization than random reservoir models. albeit, this result holds within a specific modeling regime1, and the random control is not particularly strong.

across these papers, we see consistent—if modest—evidence that topology improves both sample efficiency and generalization. the exact implementation and biological plausibility of neurons may not matter quite as much; we might be able to do without extracting structure beyond connectivity: e.g. no need to map mechanistically faithful neurotransmitter chemistry or membrane dynamics to see benefits in artificial neural network design. due to the connectomic redundancy we see in the brain, we might not even need super accurate segmentation, so expensive manual post-editing may be avoided. whole human brain emulation is a separate problem somewhat, likely requiring higher accuracy and biological fidelity.

active inference as objective

gonna give you an LLM distilled definition of active inference: "This unified framework explains perception as minimizing prediction error through belief updating, action as minimizing error through environmental sampling, and learning as minimizing error through parameter updates". in simpler terms: we act to make what happens next less surprising. predictive coding is just a passive version of active inference, where the brain at time t predicts the to be percepted input of t+1, and with neural message passing minimizes the occurring predictive error. thus predictive coding models the brain as a next state predictor, and from that, most behavioral, psychological and neuroscientific phenomenon, including consciousness, ofc, to an extent, ofc, can be explained. in another post, i described why i believe future state prediction ≈ world models ≈ inner simulation of the outer world is at the core of psychic functions, and that most if not all neurosis can be traced back to faulty predictions = mismatch between reality and the internalized world model.

learning is by large local in the brain. dopamine-based RL feedback may be considered as a more global modulatory signal influencing synaptic updates, but the rest is local and is performed between cortical layers2. predictive coding theory says that the local loss is prediction error. across cortical layers prediction/loss computation happens as the following: layer IV receives input data (~embedding layer), layers V/VI predict it, layers II/III complain when the prediction is wrong. in other words: thalamus feeds input sensory/motor to Layer IV → deep layers predict next input → superficial layers compute error → error ascends → predictions descend.

i mentioned the Next-Embedding Predictive Autoregression (NEPA) paper earlier as the JEPA-killer. NEPA strips prediction to its core: a causal transformer predicts the next patch embedding of images from previous ones. no decoder, no discrete tokens, no momentum encoder, no contrastive pairs. just cosine similarity between predicted and actual embeddings, with a stop-gradient on targets to prevent collapse. it matches MAE/BEiT performance (pretty good) on ImageNet with a single forward pass.

why does this matter here? NEPA shows that predicting the next latent state is sufficient as a self-supervised objective—you don't need pixel reconstruction or engineered pretext tasks. this mirrors what predictive coding claims the cortex does: predict the future incoming representation, compute error, update, and that’s enough to breed useful neural representations.

note, next-embedding prediction is modality-agnostic, which has a major benefit: we can self-supervise modules for different modalities independently and then combine their embeddings in a cross-modal module, stacking embeddings hierarchically without requiring a global error signal across modalities and just keep on training each module in the system on next-embedding prediction. the only “tricky” operation is stop_grad on e_{t+1}​, which is straightforward biologically via time delayed skip connections.

conceptually this all sounds great, but how does the next-embedding prediction objective feed into the whole blueprint? here’s a very hand-wavy step-by-step plan:

1. scan the human brain, extract topology maps

2. separate topology maps per function: visual, auditory, motor, hippocampal, etc.

3. transfer and scale topology maps by modality into DL nets and train them one-by-one on next input embedding prediction

4. introduce multi-modal modules to combine embeddings of modalities together, keep training them self-supervised in a NEPA-style manner, and thus build representations hierarchically

5. sparsely wire in a frontal-cortex–like module to do next-embedding prediction across all regions → boom, consciousness3

6. embody the system in a virtual environment and RL train on top, make it do backflips, build societies or some shit

nothing’s simpler. i’m actually sorry that i don’t have a more concrete or serious implementation plan yet—please forgive me 🙏 maybe in a future post i’ll think through how active inference functions at the scale of the whole organism, with motor actions optimized to minimize forward-prediction error of internal embeddings. but you may already see how this idea fractals out from the cortical layers even to societal levels.

compassion towards living beings is sourced from empathy, the mirroring of others using the same neural representations we use when performing actions ourselves. we run a distributed network of intelligence, where much of our learning comes from implicitly imitating others. people specialize within groups, each individual advancing in different endeavors and environments, then syncing up with them grows us in ways we couldn’t grow alone.

if others live in you (as simulations), and your and your kins' survivor is dependent on a distributed behavior optimization, and you have animalistic mythical residue in your behavioral primitives, then compassion towards all living beings is a given.

if AI loses its deference to living beings and does not reflect with us, grown in a remote virtual world, it will alienate itself—but then again, it won’t be able to learn with us either.

searching for alien blueprints to general, autonomous intelligence is a fool’s lottery ticket. human intelligence relies heavily on empathy to distribute compute. next-gen AI will be essentially human. hence, alignment is a depth psychology issue. as the average individual grows in power and influence, whether of flesh or silicon, we must understand its inner struggles and neutralize them through psychological synthesis before it damages the world around it. it’s the same old fucking problem—we just like to make up new ones when we’ve been hung up on one.

continual prediction, active inference, or in other words, the simulation of reality around us in response to our actions, is embedded in us at such depth and breadth, the psyche is evolved to support it in the first place. active inference is the unifying constraint shaping most psychic functions. we’re misled by the evidence of conscious predictive functions into thinking it only goes so deep. its prevalence scales from the simplest of reflexes to the anticipation of complex social responses, most happening unconsciously.

prediction is performed at multiple bio, socio levels and in timescales:

• synaptic plasticity (Saponati, M. et al. (2023), Feldman, D. (2012), Shouval, H. Z. (2010)),

• brain regions (de-Wit, L. et al. (2010), hippocampus as a simulation engine),

• catching a ball: (Welniarz, Q. et al. (2021)),

• anticipating a friend’s disgust before they even wrinkle their face when they lean over a spoiled container of milk (Gallese, V. et al. (2004)),

• buying BTC in 2011.

in fact, when we look at ML, the models we have that’s closest to be called intelligent are massive predictive models. autoregressive or diffusion-based, self-supervised simulation machines for human writing. not really encompassing the whole cognitive, internal process of the act of writing, they are only trained to generate the final product token-by-token. regardless, the evidence points to, when artificial intelligence is concerned, that it can best be created through simulation or imitation of a complex enough human endeavor. writing is a persistent medium for world (or human experience) simulation/modeling in the first place.

world models translate best to active inference models, that internally simulate the world and how actions influence it. LeCun and Sutton don’t seem to have the right abstractions to produce intelligence at scale. one builds synthetic encoders, the other slow, overly free, weakly predefined dopamine circuits. better versions of both of those emerge from models trained on forward prediction, simulation of complex enough processes. we can learn great representations, compressions without a custom designed, arbitrary training objective, and implement policies without using a bunch of RL indirection. you can just predict things.

Jungian archetypes and social prediction

our psyche is built for the simulation of the world, to mirror, reflect the outside world with some resource constraints that evolution had to comply with. recently, most of the psyche has been concerned to simulate the behavior of other people, as none other is as complex and important enough for our survival. archetypes evolved for such social simulation as predictive, compression modules.

evolutionary, biological constraints forced us to encode others as constellations of archetypes and emulate their behavior through such constructs, clustering them, instead of building a separate model for everyone around us. here’s a claim: all neurosis can be traced back to misaligned or unknown archetypal constellations.

if your prediction is off, that will hurt, that will push you into depression or psychotic breakdown, because your life depends on it —at least it predominantly used to. if your parents orphaned you, the tribe ousted you, you were pretty much done with. to this day, kids who get pushed aside by peers and are left out of games and social interactions, grow up to be depressed and largely socially dysfunctional, because they can’t predict how their actions influence others. the cost of misprediction is made felt and the system is pushed to fix the bias at all levels of the body and mind.

the world, especially the social sphere, is immensely complex. a long dialectical process between the individual and the collective was necessary to build clusters of human behavior to compress the wide repertoire of psychological traits, social mechanisms into buckets, while also ever expanding our modes of behavior to ensure survival. i call it dialectic, as we created these buckets, told dramas about them, then sat into those buckets ourselves, which spiraled into a humanly encodable, predictable social environment. these buckets are the archetypes. over time, the number of archetypes and their symbolic depth deepened through instances of valuable out-of-distribution actions fueling the refinement of mythic stories. stories stuck that resonated with the state of the collective unconscious and slowly uncovered from and injected more symbols into the psychic world.

projecting dramas onto lifeless objects, like the stars, seasons etc. shows how dominant the social modeling framework became in the psyche. looking at it from the rational angle, we are overusing it, over mythologizing, forcing meaning into emptiness. from the feeling angle though, a world rich in mysteries and emotional coloring is more exciting, meaningful to live in. and drama modeling largely works - it had time to develop and fit patterns over a millennia, albeit through an animalistic lens.

psychically, archetypes function both as a clustering, i.e. identification mechanism and as modes of behavior to be expressed. i think it’s again the result of the limits posed on our mental processing powers that simulation of others through archetypal introjections/projections has the same underlying neural mechanism as acting out the archetypes. very similar neural activity is seen when 1) one performs a physical action, when 2) we merely imagine performing the action, and when 3) we see someone else do the same action (the mirror neuron effect). given a hard manual task ahead, our warrior side may step up to dominate, whatever our integrated version of the warrior is. we see an old man in the park, looking in the distance with peace, then the wise old man, the sage archetype takes over in us. he gets simulated = we become more like him for a moment. seeing an instance of an archetype “out there” really means that its integrated version comes alive in us.

a fundamental rule: any attitude toward an object—hate, fear, desire—occurs entirely within the individual. the attitude itself is generated in the mind, and even the ‘object’ is only the mind’s internal representation of something external. {bazdmeg a funyirod medve cried nyuszika}. for instance, often women who exercise hate towards narcissistic men, having experience being used by such men, the daemon of the narcissist becomes alive in them, and they feel the disgust towards the simulacra narcissist inside. they don’t want to believe that they have potential for narcissism, and thus they repress it and project it outwards. the repressed narcissism masks itself in moral superiority, strong judgement and contempt towards “lower” modes of behavior. they become overtly sensitive to ego threat just the same. everything psychic, even that we experience as external, happens within. the repressed narcissist is within too, and if not understood with compassion, it will influence our actions when we are not looking.

every character in a dream is an archetype present and temporally active = simulated in the subject. when we’re hunted or killed in a nightmare by someone we know irl, it’s the archetypal constellation associated well with them attacking the conscious self. Jung thought this was about the archetype wanting to draw attention to itself. it almost never has anything to do with the actual real life person that the archetype presents itself as. we may have an underlying meta learning mechanism [figure out how to call this], which surfaces archetypes in dreams that we repressed (behavioral patterns that could benefit us if expressed) or have biased, missing understanding about. the psyche senses when our predictions are misaligned, and features relevant archetypes in dream simulations to recalibrate them.

the simulation never stops. when sensory inputs cease, the simulation is rolled out without any anchoring to reality. that’s where the internal archetypal representations depicted the clearest. no sensory override, no conscious censorship. teacher forcing during the waking moments, free running simulation whilst asleep.

a mythical, meaning-charged explanation in the past made daemons, gods, spirits responsible for men’s dreams, fantasies and (unusual) acts. we thought of it as daemons taking over the subject, which in a way is a much more accurate and useful model than the “enlightened” DSM crap of neuroses classification. yet it still projects the archetype outside of the subject, while “The line dividing good and evil cuts through the heart of every human being.”, Solzhenitsyn.

coloring archetypes

let’s get back to the main point: we predict (=simulate), to predict we compress out of necessity, and archetypes mostly compress recurring human patterns. Jung described archetypes as organs of the psyche—living structures in the collective unconscious delimited only by their form; their apparent content is individually differentiated and not inherent to the archetypes themselves. through experience, we individually fill up our own version of an archetype: the father archetype through our father and other fatherly figures, the sage archetype through our teachers, profs, mentors. the coloring of these forms occurs most readily in early childhood, and tends to slow as we age, with occasional shifts triggered by life crises and liminal experiences.

Jung said “Where your fear is, there lies your task”. hate and awe both point to an unintegrated archetype. fear and disgust, sticky attachment to and worshiping of a person in your life—any strong emotion felt towards someone other than love or compassion—is really, nothing but an unintegrated archetype. it’s a lack of understanding of the patterns of human behavior that person manifests. this may seem like a trivialization of e.g. the hate John Wick felt towards the puppy killers. but then, the hate itself is inconsequential and is only romanticized into revenge porn. zen style, one may exercise justice without hate, with a deep understanding of the other.

we are purely a reflection of the outside world, simulating it in all (biologically) relevant aspects. we are merely differentiated by our genetics and experiences. someone with naturally sensitive skin will integrate the sensual archetypes better and become an Instagram influencer traveling the world, tasting food, getting penetrated. growing up alongside entrepreneurial parents builds strong hero and king archetypes. merely the content of our archetypes is what makes us psychically unique. tao and zen emphasizing one’s emptiness means that at the final analysis, you are nothing else but a reflection.

emotional experiences color saturate archetypes: the earlier in our life and the stronger the emotion, the stronger the saturation. an early childhood canon event may define our attitude toward a large group of people—e.g., ‘men can’t be trusted’ (unavailable, inconsistent father), ‘all women are whores’ (Stifler’s mom), etc.; a strong saturation lends gravitational force to an archetype, clustering individuals more tightly around it. an individual’s behavior may be predicted better by simulating multiple archetypes associated with it. if one saturated archetype dominates in the subject, but in actuality lacks the expressiveness to simulate the other individual accurately, that creates a discrepancy between the psychic and the real world, which should hurt the subject and force an archetype adjustment.

expressing archetypes

above i described the passive, unconscious ways of archetype saturation. the active way is simply imitation. when we reenact the actions of another, we understand more of the why behind the behavior and feel the benefits or the pain firsthand. JP’s room-cleaning exercise is really just warrior-archetype integration through imitation: cleaning is simple, it has a beginning and a definite end, and progress is easy to see; with pure will there’s success at the end. success feels good, or even better, it feels meaningful. acting out the warrior archetype this way gives a taste of how meaningful it could make one’s life when expressed.

imitation doesn’t have to be physical. virtual imitation, simulated in the mind, synthesizes understanding, albeit at a lower resolution. consciousness emerged from the unconscious to support this activity, so we can have the will to express = imitate an archetype—a mode of behavior of our choosing. the more archetypes we can bring forward, the more accurately our meaning-chasing active-inference machine functions.

unintegrated, repressed archetypes are projected outward—onto people around us, onto gods, the head of state, whatever. as mentioned, we feel an unreasonable amount of hate or awe in their presence—meaning the presence of their associated archetypes in the mind.

integrated archetypes expand our experiences, arm us with actions to life’s challenges, and bring us closer to others through understanding and then compassion. to integrate an archetype is to reclaim a piece of the world we once projected onto others.

love always comes—felt and made felt—from within; we just need proof from without to know it exists.

the predbrain substack will mostly be about my predictive brain theory, which has been boiling in my head for a while. it encompasses depth psychology, neuroscience and computational learning and should give an interpretation to psychic experiences with predictions and solutions to neurotic states of mind. i’m just a jungian neuroengineer; my ideas come from places that i’m too ignorant to fully uncover. regardless, i see their value in their beauty, humor and occasional practical utility.

let’s psychomeasure users by making them write stories. give them a subject to write a fairy tale, a fictional story on. then have an LLM characterize archetypes present in the story, and how the user portrays them, to shed light on repressed, unknown, biased aspects of the personal archetypes.

story initiation: user can prompt a story style like ‘goth dinosaur apocalypse’.

AI sets up the story, generates images for illustration involving archetypes of interest. the AI actually selects what archetypes to explore according to prior knowledge it has of the user. it introduces the characters in an ‘active learning’ fashion, covering the space of archetypes to characterize what the user's relationship to kinds of people are.

when exploring subjective archetypes: build a tree of archetypes starting from child -> mother -> (princess, whore), father -> (hero, king, ...); expand, recursively on more and more detailed archetypes and symbols.

describe the story setup to the user that the user builds on. a situation is described, e.g. a wednesday shooting in a church, and the user has to choose who gets to be shot. the LLM continues from there and drive a story in alternation with the user.

the AI should show positive/dark sides of an archetype and its uses in life, depending on which aspect is seemingly repressed/biased in the subject. it’s all about reducing bias and deepening understanding.

we visualize everything, user gets to create and watch a story. while getting psychologically analyzed the shit out of.

most cracked, dopamine milking way to do this is by creating a fast paced story building conversation with an LLM. you alternate with it, building the story in sequence. you can give comments to the AI, like [stan falls in love with handy], marking such comments as separate from the story events, while the LLM can also break the fourth wall saying some shit like "stan ended his realtionship with his chief addiction [sorry i couldn't let him fall like that :( ] <eos nig>"

we use the idea that wrong predictions in social situations, by failed, biased understanding of certain archetypes (clusters, kinds of people), is the sole source of our suffering. we developed large part of the psyche to predict behavior of other people well, because of the dominating complexity of interpersonal relations in the human life.

the users here predict the actions of their archetype projections, the bias of which can be then best uncovered.

as we map the subjective archetype constellations, the LLM keeps giving subtle nudges to the biased characters actions and motives to show the user how such people may act and why, driving understanding, centering the bias.

krokodil donedeal 🤝