[daxfohl]

There's a _lot_ of smoke and mirrors. Paste a sudoku into chatgpt and ask it to solve. Amazing, it does it perfectly! Of course that's because it ran a sudoku-solving program that it pulled off github.

Now ask it to solve step by step by pure reasoning. You'll get a really intelligent sounding response that sounds correct, but on closer inspection makes absolutely no sense, every step has ridiculous errors like "we start with options {1, 7} but eliminate 2, leaving only option 3", and then at the end it just throws all that out and says "and therefore ..." and gives you the original answer.

That tells me there's essentially zero reasoning ability in these things, and anything that looks like reasoning has been largely hand-baked into it. All they do on their own is complete sentences with statistically-likely words. So yeah, as much as people talk about it, I don't see us as being remotely close to AGI at this point. Just don't tell the investors.

[conception]

On the other side of the coin, I think people also underestimate the amount of human thinking and intelligence is just completing statistically likely words. Most actions and certainly reactions people do everyday involve very little reasoning. Instead just following the most used neuron.

[seadan83]

Human vision works this way. To fix the latency problem (actual event hff happening vs signal transmitted to your brain) human vision is constantly predicting what you should see, your brain tells you that is what you saw (the prediction), and then the brain does reconciliation after the fact. Your brain will scramble for coherency when prediction and reality do not match. This trickery is why it seems like you see events in real time, when there is actually a significant delay between event and perception.

Though, there are error correction mechanisms, systems for validation, and a coherent underlying model of the world that is used by tthee brain.

FWIW, it is likely the most used set of neuron connections, sets of millions in play and their interconnections being the important part. That subset being one of billions of others with thousands of connections between each neuron - keep in mind it is not the set of neurons firing that matters, but the set of connections firing. The set of connections is a vastly large number.

Like, if you have three neurons, your brain can encode 10 data points. Let's call these A, B,C. A firing and terminating is one (so three for each), each edge, eg A to B is another three, each set of two edges, eg A to B to C (three more), and all three edges for one more. Then keep in mind you have billions of neurons and they are each interconnected by the thousands.

[daxfohl]

True, and this even happens in a brain the size of a flea's.

Which makes one wonder, what is it that makes processing and reconciling millions of visual signals per second "easy", but reasoning through a simple sudoku near impossible?

[stevenwoo]

Are you sure about fleas? I thought we know this does not happen with frogs as far as we can tell https://courses.csail.mit.edu/6.803/pdf/lettvin.pdf

[daxfohl]

Oh. No, not sure.

[stevenwoo]

I do not know how many times that frog type experiment has been repeated on other species or branches of animal life but that one study up ended my preconceived notion of how vision could work with a brain, and most insects have a tiny brain, speculatively may be possible only of certain automaton type tasks, though wasps IIRC have very small brains relative to other insects but exhibit social behavior so who knows.

[ilbeeper]

Citation needed. The word reasoning isn't describing everything that the brain does, and "just following the most used neuron" is not even wrong.