Max LeibmanBrains are not computers. Computers are not brains.
Ben Ramsey@maxleibman @blogdiva Here’s a really good article related to this: https://aeon.co/essays/your-brain-does-not-process-information-and-it-is-not-a-computer
The Sleight Doctor 🃏🍉@maxleibman @blogdiva @ramsey I wonder if the folks who write these pieces have aphantasia (the inability to mentally visualise) and so can't conceive that some of us can do this vividly and in detail.
There's been significant progress in reconstructing both seen and imagined images from brain scans. And the IP metaphor really only says that the organisational processes of a mind can be modelled. The only way out of this conclusion is dualism (the ontological irreducibility of mental states). 🤷🏻♂️
@ApostateEnglishman @blogdiva @ramsey The weather can be modeled, but we wouldn't say a computer has a climate.
@blogdiva @ramsey @maxleibman Everything we are is cells following rigid rules fixed by DNA, so any argument you could use to "prove" that a philosophical zombie wouldn't be conscious, sapient, sentient or self-aware, applies equally to humans and other animals...unless you think there's some mysteriously undetectable secret sauce or "elan vital". We don't know yet whether what makes us human could be run in alternative physical substrates. It could turn out that only our mushy brains can do it!
@blogdiva @ramsey @maxleibman (But to be clear, if it does turn out that only our mushy brains can do it, it won't be because of the secret sauce, it'll been for *functional* reasons that only our unique biology can physically achieve. I'm currently an agnostic about it. What I do object to, however, is the mistaking of bad analogies and philosophical "thought experiments" for science. I'm Team Dennett on this one!)
feliz@ApostateEnglishman @blogdiva @ramsey @maxleibman
Why do most of these theories ignore that we are learning?
Our nervous systems, which includes our brains, are more than “cells following rigid rules fixed by DNA”. We humans are becoming human through LEARNING PROCESSES. They widely differ in individual, social, linguistic, cultural, geographical,climatic,nutritional and many other factors. All these factors influence how our nervous systems operate in detail.
That's where the modeling fails.
Why do most of these theories ignore that we are learning?
Our nervous systems, which includes our brains, are more than “cells following rigid rules fixed by DNA”. We humans are becoming human through LEARNING PROCESSES. They widely differ in individual, social, linguistic, cultural, geographical,climatic,nutritional and many other factors. All these factors influence how our nervous systems operate in detail.
That's where the modeling fails.
Antin@ApostateEnglishman @blogdiva @ramsey @maxleibman @feliz
THIS. The brain is a process, not a processor. It changes all the time; different parts at different rates at different times in different situations.
The function is ecological and evolutionarily compounding. Your brain is fundamentally different from your parents', because yours is strongly influenced by your parents' experiences. All the while, yours is trying to predict what your children's brains will need in their environments, and so on.
A climate system is more analogous to a brain than a computer is.
Not to mention all the other organisms (e.g. gut bacteria) that actively influence your brain and behavior. "You" are not just your brain! "You" are not just "your" DNA!
I agree with both of you. I even explicitly called the mind an organisational process. Indeed that was the basis for me arguing that there was no "in principle" reason it couldn't run on an alternative physical substrate (we don't know yet whether or not that might be possible).
My guess is that a fully conscious non-human machine would have to go through childhood, learn and grow just like a human. It would need to be sensorily and culturally immersed, just like a human.
@ApostateEnglishman @antinomian There is no indication of the possibility of a robot with sensors that are similar to the level of sensitivity & complexity & integration of the human system. And then find a person willing to have that robot implanted so it can have all prenatal experiences necessary for humanlike development? Science has shown that even that phase is crucial for psychological foundations. Let alone the 15-20 years of adolescence. Sorry, it's not feasible to recreate all that.
@antinomian @feliz You're talking about the limits of current tech, and ethical constraints, and I don't disagree with any of that. I do *not* think we'll have conscious robots, or even that it would be desirable. But we know physical stuff organised in a certain way can be conscious, sapient and self-aware, because you can look in a mirror and see for yourself. That's what you are: a vast assemblage of tiny cellular machines. Of course you're far more than that, but there's no magic involved. 🤷🏻♂️
@ApostateEnglishman
There's fundamental differences between machines and human cells. That's a scientific fact. Our cells' functions are not replicable with machinery, nor can their functions be sufficiently explained with metaphors of machinery. We have biology as a science for that. We have biological knowledge which represents a different level of integration than technology. We don't need to assume magic to happen on any of these levels of integration, we have scientifiic proof. (1/2)
There's fundamental differences between machines and human cells. That's a scientific fact. Our cells' functions are not replicable with machinery, nor can their functions be sufficiently explained with metaphors of machinery. We have biology as a science for that. We have biological knowledge which represents a different level of integration than technology. We don't need to assume magic to happen on any of these levels of integration, we have scientifiic proof. (1/2)
On the biological level,we observe processes that cant be described or replicated using technology. Then we have psychological and sociological knowledge. Our nervous systems can be understood biologically and bio-graphically. Humans are social beings,interacting through bodies. Neuroplasticity is a scientific fact. All this is not applicable to machines. We humans and our thinking function on the basis of symbols which are transferable in the form of cultural knowledge. No magic involved. (2/2)
@feliz Also, it seems you allude to "scientific fact" here as a rhetorical device, not because it's actually "scientific fact". In the cognitive and neurosciences, the consensus view *is* that we're complex biological machines, "designed" (note the scare quotes!) by evolution, possessing intricate, self-repairing systems like cells, neurons, and ATP-synthase motors. The current science *does* assume we're organic "computers" with advanced adaptability, rather than rigid, pre-programmed devices.
@ApostateEnglishman
The idea that the brain is an “organic computer” or that cognition is similar to “computation” is an outdated, mechanistic, reductionist approach. It is not the consensus in modern neuroscience and cognitive science. Because those are misleading simplifications which miss key aspects of biological cognition.
Maybe some researchers still use computational models for specific purposes, especially in books of popular science. And also AI marketeers do that.
@feliz Current scientific consensus is that the brain is an adaptive, biological and embodied information-processing system capable of constant self-reorganisation, and so nothing like a silicon-based digital computer. But cognition is still seen as a form of processing. Or how do you explain us already being able to interpret mental images from brain scans? It's not sci-fi today.
I also hate "AI" evangelists so you may be talking to a caricature you read about in a Searle book, rather than me!
@ApostateEnglishman
I read about that study where a latent #DiffusionModel is reconstructing visual experiences from human brain activity – you mean that? It's very interesting. The study correlates measured #brain activity with AI layers. The authors use a technical "system-to-system" language because there's a significant gap in understanding both the human brain and the inner workings of the AI. That's not a consistent theory, it shows that we don't have a clue yet how either works. (1/2)
I read about that study where a latent #DiffusionModel is reconstructing visual experiences from human brain activity – you mean that? It's very interesting. The study correlates measured #brain activity with AI layers. The authors use a technical "system-to-system" language because there's a significant gap in understanding both the human brain and the inner workings of the AI. That's not a consistent theory, it shows that we don't have a clue yet how either works. (1/2)
@ApostateEnglishman Correlating AI activity with brain activity doesn't mean that the brain is "decoding" or "calculating". It only shows that our mathematical models of AI are the best statistical tools we currently have for organizing and interpreting complex data.
I still think: As long as we speak in metaphors and say the brain "calculates","stores" or "processes data", we admit that we have no language yet for what the brain actually does. And we obscure what we want to understand. (2/2)