5 ms·
The problem is, neurobiology proves there are no world models. Silicon Valley bet on the wrong cognition model, a psychological version trapped in 20th C bunk,
by Mallowram 1y ago
The problem is, neurobiology proves there are no world models.
Silicon Valley bet on the wrong cognition model, a psychological version trapped in 20th C bunk, and everyone pays the price listening to cult leaders like Scott Alexander worm their way out of consciousness.
https://pmc.ncbi.nlm.nih.gov/articles/PMC7415918/ https://pmc.ncbi.nlm.nih.gov/articles/PMC7415918/
- lukev 1y agoHow can you say there are no world models, when I can literally draw out a simple one for you on demand? You can argue that's they're not the governing principle of cognition, but it seems farcical to say they don't even exist, when we are trying to explain them to eachother all the time.
- mallowdram 1y agoNo what you're describing is arbitrary and idiosyncratic. The brain doesn't use that to survive, it doesn't need them. Anything external to that is completely separate from thought. What you're describing is an arbitrary game for entertainment to fill up your time and confuse yourself and others. It has no relationship to the choices you would make to survive, and can only interfere with it. The "world model" you're describing is arbitrary fiction. “We refute (based on empirical evidence) claims that humans use linguistic representations to think.” Ev Fedorenko Language Lab MIT 2024
- Izkata 1y agoWorld models aren't linguistic. You seem to be conflating (at least) two different things, then claiming because one doesn't apply the other doesn't exist. Edit: Also come to think of it, that quote is odd, like it's rather late to the party. The NPC meme is several years old and came from a study that most people don't have an inner voice - that they don't think with words.
- mallowdram 1y agoOf course world models are linguistic. What working memory or neural syntax bypasses linguistic externals when the term is in of itself linguistic. The entire concept of model is linguistic in origin. Biology doesn't have models.
- boxed 1y agoI think you might be confused about the expression "world model". In this context it clearly means that a person has an understanding of reality based on "math->physics->chemistry->biology->psychology" instead of "peer pressure->group identity" or whatever you see in QAnon or cults or whatever. If a person primarily evaluates the truth content of a statement based on identity or something instead of math/physics/etc then that person has no "world model", and vice versa.
- mallowdram 1y agoNo it's neurobiological, psychology and particularly cog-sci is in error, as both place language ahead of concepts. Our understanding of reality per language is post hoc, it's an illusion. Life is always ad hoc, any violation of a narrative model is easily evaded for survival. This is simple stuff folks!
- boxed 1y agoYou like to use big words. But you're not making any sense. Notice this.
- Mallowram 1y agoIt only makes sense to scientific thinkers who think in correlation, and know how degraded words and sentences are. Anyone who accepts word statements as valid is at loss. The only role of language is to refute itself.
- morpheos137 1y agoThe irony is intense. Classic psychotic dysphasia.
- _Algernon_ 1y agoThat's not what world model means, neither in psychology or artificial intelligence — two fields Scott writes a lot about so he should know how the term is used or define how he uses it if he uses a non-typical definition.
- dwaltrip 1y agoI don’t necessarily disagree with you, but I wonder if you are not being overly reductive / pedantic. What is there then? What words (heh) do you use to distinguish between someone who makes more accurate predictions about the world than someone else?
- Mallowram 1y agoFirst off, we don't use predictions, that's another model, it's false (Spontaneous Brain Northoff or read Mofakham's papers). In terms of words, they barely represent and never reference. Any statement like that serves primarily status gain, not know knowledge transmission (I proved this from the first statement above as well). The reality is CS built a math model from totally false premises as it relates to communication and knowledge. It works for efficient value trading using symbols in place of actions. Does it have a future, no. The problem is how do we shift to a real neurodynamic system of sharing? https://docs.google.com/document/d/1cXtU97SCjxaHCrf8UVeQGYajatFQ_MU69MBDmIFe58I/edit?usp=sharing https://docs.google.com/document/d/1cXtU97SCjxaHCrf8UVeQGYaj...
- dwaltrip 1y agoI think we are talking past each other. The ideas you mention sound interesting, but I’m not sure what the point is.
- mallowdram 1y agoWords do not function as communication. You asked a pointless question. It has no function except to extract values from actions: it was subjective, arbitrary. Until CS grasps this, it is irrelevant. All the symptoms from LLM failure rates stem from their reliance in arbitrary forms to extract value, and they are no different than the errors we experience in reality in climate, politics etc. CS didn't solve the initial conditions, it's maxing them out as errors.
- morpheos137 1y agoYou seem unable to communicate clearly what it is you are trying to say. Honestly i would be interested but I can't follow you. Your use of words is non standard. You say words are about social interelations, e.g. domninance, exchange, etc but then you also say that words don't communicate information. Of course thought is fundamentally non-verbal but since telepathy doesn't exist we use symbols, i.e. words to communicate with oth ers. Nobody thinks words ard thoughts. The represent thoughts. Your part about words being deception is one of many use cases. Honestly, and with all due respect, your comment fits the style of a person in an altered state of mind be it drugs, psychosis or something else. Why write words in a public forum with your weird little idiosyncratic meanings nobody else understands.
- suddenlybananas 1y agoNeurology has not proven any such thing. Our knowledge of neuroscience on the cognitive level is super limited and we don't have a good understanding about how any higher-order cognition works.
- mallowdram 1y agoNeurobiology has proved this, just read Buzsaki or Northoff. The brain doesn't need models, it needs differences.
- suddenlybananas 1y agoI have a PhD in cognitive science and my supervisor was a neuroscientist.
- Mallowram 1y agoThat's irrelevant. Cog-sci is largely folk psychology, and the problems in automating inference in AI demonstrate the model would eventually collapse. Question is how do we toss this model aside for an irreducible form of post-symbolic relationship between brains and machines?
- suddenlybananas 1y agoI appreciate your gumption but I really think that you don't understand things as well as you think you do. Maybe read someone other than Paul Churchland.
- mallowdram 1y agoBoth Churchland's are out of date. Note the references above, this is a neurobiological, dynamic approach they're not party to. If you don't know what those are or optic flow, neural reuse are, then study them. Trad Neuroscience and cog-sci is no longer applicable.
- vintermann 1y agoIt takes a pretty damn complicated model of the world to start explaining things with neurobiology.
- Mallowram 1y agoDoesn't require models in dynamics, coordination or otherwise.