There is no magic in the brain. There is no magic in LLMs. There is just new experience we gain by interacting with the environment and society. And there is the trove of past experience encoded in our books. We got smart by collecting experience, in other words, from outside. The magic in the brain was not in the brain, but everywhere else.
What is experience? We are in state S, and take action A, and observe feedback R. The environment is the teacher, giving us reward signals. We can only increase our knowledge incrementally, by trying our many bad ideas, and sometimes paying with our lives. But we still leave morsels of newly acquired experience for future generations.
We are experience machines, both individually and socially. And intelligence is the distilled experience of the past, encoded in concepts, methods and knowledge. Intelligence is a collective process. None of us could reach our current level without language and society.
To say there's no magic in the brain drastically *minimizes the complexity of the brain.
Your brain is several orders of magnitude more complex than even the largest LLM.
GPT4 has 1 trillion parameters? Big deal. Your brain has 1 quadrillion synapses, constantly shifting. Beyond that the synapses are analog messages, not binary. Each synapse is approximately like 1000 transistors based on the granularity of messaging it can send and receive.
It is temporally complex as well as structurally complex, well beyond anything we've ever made.
I'm strongly in favor of AGI, for what it's worth, but LLMs aren't even scratching the surface. They're nowhere close to a human. They're a mediocre pastiche and it's equally possible that they're a dead end as it is that they'll ever be AGI.
That kind of explains why humans need to absorb less language to train. It still takes 25 years of focused study to become capable of pushing the frontier of knowledge a tiny bit.
Re: synapses being analog messages, isn’t this sort of true of neural networks? In my understanding, the weights, biases and values flowing through the network are floating point numbers so I’d argue closer to analog than binary.
You know that's a good point that I didn't consider in my comment. LLMs use 16bit floats, which is approximately the same number of values.
But then you get into the nuance of spiking vs non spiking neurons etc, which to my knowledge isn't emulated. The brain is extremely complex. I don't believe LLMs do anything similar to inhibitor neuronal function for example.
Neuroscientists would agree with GP, otherwise they would be neuromystics instead of neuroscientists. There is no magic. It's all physical processes that we can eventually understand.
The entire point was that we do not understand it? That much of how the brain work is "magic" atm.
It's our field that are the alchemist mystics, rambling about AGI/Philosopher's Stone in ever increasingly unhinged ways, while stirring our ML-pots that we have never even tried to prove have a chance to be anymore successful than the alchemists.
Visarga used the word literally and stated in broad strokes what the brain does (process input, update state, emit outputs) without the details of how it does so. You're the one who misinterpreted that as meaning we already know exactly how the brain works.
No, not at all. The only thing I did was to react on how ridiculous that oversimplification is and how such a thing can only come about due to an embarrassing amount of hubris currently going around our field with relation to "AI". It's a hand-wavy "Eh, how hard can it be?" comment to rationalize ML being a pathway to AGI.
There is nothing oversimplified in that description at all. Here is an equivalent description from a neuroscience textbook:
"Neuroscience is the study of the nervous system, the collection of nerve cells that interpret all sorts of information which allows the body to coordinate activity in response to the environment."
This is even simpler than the RL agent description of the brain that visagra provided earlier, ignoring that the brain must have some internal state instead of being a pure function from each input to each response. Would you say that neuroscientists have even more hubris?
Yet again you just pick out a word, "oversimplified" in this case, ignores the rest of the reply - and more importantly, the entire context of all these replies - and goes on to make a nonsensical comparison with the ever-present one-liner that all textbooks of all academic disciplines have.
Have you not understood by now that my critic of our field's AI-hubris is a general one - and thus not hinged on exact wordings? I could've written similar hubris related replies on multiple other comments in this thread. It's my reaction. It's my exasperation.
You keep ignoring that your criticism is nonsensical. There is nothing in Visarga's comment that indicates hubris. If there is, you would be able to plainly point it out instead of complaining that people are misinterpreting your complaint.
> There is nothing in Visarga's comment that indicates hubris
I think it does, and I've explained exactly how but you just pivots to snarky comments on specific words instead - first scientists, then magic, then oversimplified. Furthermore you seemingly can't accept no matter how many times that it's not specific to this particular comment but meant as a response to the general "AI-hype. So judging by this you're either wilfully obtuse or completely unable to read between the lines no matter how many times I explain it to you. Did you miss that all sibling comments to mine seemed to - shockingly - be able to react the same way I did - albeit less exasperated? But somehow you're completely unable to accept that reaction.
Seriously though, to make my general non-comment specific point for the last time; just do a Occam's razor on the "AI"-hype timeline between if it's a bog standard self-interested tech-hype, using the same methods but with more hardware, or an extraordinary scientific breakthrough justifiably triggering all the extraordinary claims about AGI being close and even an existential threat (!). I think it's the former. If you think that the latter is likelier, fine, but I do think that that requires an embarrassing combination of hubris & confirmation bias because it strokes our egos.
I assume that you're invested in this hype and has taken offense by me calling it embarrassing hubris. I hope you'll be okay when this bubble inevitably bursts. I'll leave it to it now, since I think I've made my general point even though it's clear that it will never be acknowledged.
> No, not at all. The only thing I did was to react on how ridiculous that oversimplification is and how such a thing can only come about due to an embarrassing amount of hubris currently going around our field with relation to "AI". It's a hand-wavy "Eh, how hard can it be?" comment to rationalize ML being a pathway to AGI.
Added emphasis, and others reacted the same way. You can disagree with the interpretation but to like disallow it like you seem to be attempting is really weird.
Visarga's comment is equivalent to comments that neuroscientists make, as I showed earlier, so inferring that it means, "How hard can it be?" is your mistake alone, not visarga's. The voters agree.
You picked an intro one-liner from visarga's comment. It's amazing that you really can't see that this is the same thing and yet are capable of using the Internet. Neither one claims to know how the brain does what it does, but they both say that the brain does the same thing.
I picked the line within a context, explained the context multiple times to you to make sure. You picked the line, removed all context of the discussion and the goes on to compare it a randomly picked textbook intro line. Simply mindbogglingly stupid.
Not randomly picked. The introduction. Visarga's comment is not a treatise on how exactly to reproduce the brain. The fact that you still don't understand this is seriously impressive.
What is experience? We are in state S, and take action A, and observe feedback R. The environment is the teacher, giving us reward signals. We can only increase our knowledge incrementally, by trying our many bad ideas, and sometimes paying with our lives. But we still leave morsels of newly acquired experience for future generations.
We are experience machines, both individually and socially. And intelligence is the distilled experience of the past, encoded in concepts, methods and knowledge. Intelligence is a collective process. None of us could reach our current level without language and society.
Human language is in a way smarter than humans.