Scanned the article for any significant insights and only notable reference to me was the distribution of neurons, though it was lacking any meaningful details.
Quick Google turned up: “Octopuses have an extensive nervous system, with over 500 million neurons, similar in number to that of a dog. But unlike dogs and other vertebrates, where the majority of neurons are in the brain, over two thirds of the octopuses' neurons are located within their arms and body.”
Appears neurons are evenly distributed among the each the eight arms. That said, unable to find any research that shows the arms show any signs independent intelligence.
I can imagine that arms equipped with significant neural processing would allow higher-level signals to be sent from the central brain. Rather than in most animals like ourselves, where the motor cortex basically tells which muscles to contract, the arms might be given a signal to feel around and lift a rock, or just explore and report back. Basically, leaving some or all of the planning to the brain, but offloading control. That, plus tightening the feedback loop between sending (touch) and actuation would make for some pretty useful advantages I would guess.
the necessity to see explains the position of the eyes but not necessarily of the brain. After all, you'd see just as well if your brain was in your stomach as long as it's connected to your eyes.
duration for the signal to travel is one explanation, but not a great one because nerves are fast. (a lot of nerves take weird detours, the visual cortex itself is located at the back of the head). More mundane explanation is heat dissipation. The brain consumes a lot of energy and emits a lot of heat. Would be pretty hard to get rid of it anywhere else.
they definitely are, about 300 mph (so that'd be about 10-20 ms). In fact one of your nerves, the laryngeal nerve enabling you to control your vocal cords wraps from your head around your heart back to your head, even in giraffes! Evolutionary it was apparently just more beneficial to make the thing longer.
That should be up to about 300 mph. There is wide variability depending on the structure of the nerve. Even so that maximum is tremendously slow compared to a wire.
You're forgetting to account for response latency as well. It's not enough to see danger quickly, you also have to be able to move your body in response.
If the brain was in the abdomen, the latency to move your head would be a roundtrip longer.
I'd bet on it being too hard for evolution to put eyes far away from the brain. The eyes are basically a part of the brain in every species, the optical nerve is so thick and does signal processing midway, I don't think our genes has a concept for dragging more optical nerves around the body or extending it etc.
Maybe with some more billion years of evolution our DNA would create more abstractions, like how to generalize optical nerve paths and not just touch/muscle etc, but as we are now it seems like the logic isn't there. Such a thick nerve bundle would need to get added to the spine with everything else, or risk getting pinched and disabled just by moving around, but for now it seems like the spine abstraction doesn't support optical data.
> After all, you'd see just as well if your brain was in your stomach as long as it's connected to your eyes.
Assuming the brain and stomach both need mutually exclusive physical space, let's swap the two and see what happens.
If the stomach is in the head rather than where it is now then it is right by the mouth, so food need to travel less. So it might be a first order solution to optimal spatial layout of the body components. Except, now there isn't enough time for the food to be digested so you have to move the stomach further away...
Latency. Nerves are not fast, but it's irrelevant if they were: if the man-eating giant cat has a the edge, speed becomes relative, yours vs cats, not absolute.
Actually, the visual cortex of mammals is in the back of the head, about as far from the eyes as possible (https://en.wikipedia.org/wiki/Visual_cortex), whereas the visual area of octopus are (I think) right next to their eyes. That said, the comparison is not quite that straightforward because our axons are myelinated, which speeds up signals tremendously.
Edit: Ok, "as far as possible" is an exaggeration, but they aren't close.
The visual cortex has a direct pathway to the retina via the lateral geniculate nuclei.
And there are lots of reflexes to fire, before we have time to understand what we're seeing. Like putting your hands in front of your eyes when an object approaches rapidly, long before you recognize what the object is. There needs to be something between the retina and visual cortex to respond to this quickly enough.
In many ways it is a bad thing compared to a faster centralized brain. Seems like an evolutionary local minimum where the organism supports an energy expensive peripheral nervous system that is just good enough to function.
Had already reviewed that article; from the article, “This study makes it clear that octopus's arms don't behave totally independently from the centralized brain.”
Limb independence, is not the same thing as limb independent intelligence. Independently intelligence would mean when separated from the would the arm still shows signs of intelligence. Limb independence is simply central brain with limbs that are concurrently able to independently and meaningfully able to function, either in unison or separately; for example how a human drummer can use four limbs to tap out four separate beats, each with different timings, but that are in sync.
There’s not a lot of meat in the article. Cephalopods it seems are a fantastic jumping off point for rewarding thought experiments about the mind and it’s relationship with the brain. But any claim to insight about “how cephalopods think” is pure bunk.
Or maybe people like reading Nautilus like I do and wanted to share what they thought was interesting. I think Nautilus has substance. It briefly introduces me to new ideas, people, and other miscellany that I otherwise would have never come across. It’s not an academic publisher or college-level lecture.
It's interesting when you think about the concept of how our own minds work. There are lots of studies about splitting the path between the two hemispheres of the brain called corpus callosum. In these situations people will sometimes exihbit "split" behaviour but over time that subsides and the brain is able to more or less subconsciously coordinate. Some of the patterns of behaviour are contrary; one hand does one thing, the other hand stops it. It's truly remarkable.
In any event it might be a loose association, but I do agree with the overall _title_ (and that Octopuses are pretty cool). I hope we're able to recognize any intelligence way may create if we do so successfully.
The book mentioned in the article, Children of Time by Adrien Tchaikovsky is fantastic. Adrien does an excellent job of trying to imagine how distributed intelligence would work for a species as it evolves.
Yes! Tchaikovsky's brilliant and under-read. Hyperintelligent spiders using ant colonies as computers... it's far-out amd yet strangely plausible given the suspension of disbelief...
The novel /referenced/ is Children of Ruin, the sequel, which is just as amazing. I assume the author is referring to the diptych by its first part's name.
Small discrepancy which might confuse exactly the readers who would enjoy the whole work.
I think this will be the way towards general intelligence, multiple decentralized 'brains' working together. For example, you may have something that processes vision, something that does task planning, something that does motion planning, something that processes audio, etc.
One day we will bring these parts together into single systems, but for now it's much easier to train and test these independent 'brains' until we figure out a proper architecture. There are quite a few different things we need to answer, for example, can motion planning and motions be de-coupled? Can vision detection be de-coupled from object classification? Can hearing be de-coupled from vocalization?
Octopuses and whales are two of the most interesting cases of high-level intelligence. Many species of whales actually have much more dense folding than our brains, and they are of course much larger. I know that larger brains are thought to be required for larger bodies, but I haven't seen reflection on that that considers aquatic animals have have different energy efficiencies than land based mammals. Given the social cohesion of some species, like orcas, I would not be surprised if their intelligence actually exceeds ours.
The other thing is that I think humans greatly and consistently underestimate the intelligence of almost all animals.
35 comments
[ 3.2 ms ] story [ 73.9 ms ] threadQuick Google turned up: “Octopuses have an extensive nervous system, with over 500 million neurons, similar in number to that of a dog. But unlike dogs and other vertebrates, where the majority of neurons are in the brain, over two thirds of the octopuses' neurons are located within their arms and body.”
Appears neurons are evenly distributed among the each the eight arms. That said, unable to find any research that shows the arms show any signs independent intelligence.
Octopi live underwater. Light diminishes with depth. More data gained from feeling around currents, so brain goes in tentacle.
duration for the signal to travel is one explanation, but not a great one because nerves are fast. (a lot of nerves take weird detours, the visual cortex itself is located at the back of the head). More mundane explanation is heat dissipation. The brain consumes a lot of energy and emits a lot of heat. Would be pretty hard to get rid of it anywhere else.
No they’re not. If your optic nerve had to travel from your stomach we’re talking a delay of 100s of milliseconds.
If the brain was in the abdomen, the latency to move your head would be a roundtrip longer.
Maybe with some more billion years of evolution our DNA would create more abstractions, like how to generalize optical nerve paths and not just touch/muscle etc, but as we are now it seems like the logic isn't there. Such a thick nerve bundle would need to get added to the spine with everything else, or risk getting pinched and disabled just by moving around, but for now it seems like the spine abstraction doesn't support optical data.
Assuming the brain and stomach both need mutually exclusive physical space, let's swap the two and see what happens.
If the stomach is in the head rather than where it is now then it is right by the mouth, so food need to travel less. So it might be a first order solution to optimal spatial layout of the body components. Except, now there isn't enough time for the food to be digested so you have to move the stomach further away...
Edit: Ok, "as far as possible" is an exaggeration, but they aren't close.
And there are lots of reflexes to fire, before we have time to understand what we're seeing. Like putting your hands in front of your eyes when an object approaches rapidly, long before you recognize what the object is. There needs to be something between the retina and visual cortex to respond to this quickly enough.
especially for animals which orient using scent.
In many ways it is a bad thing compared to a faster centralized brain. Seems like an evolutionary local minimum where the organism supports an energy expensive peripheral nervous system that is just good enough to function.
Limb independence, is not the same thing as limb independent intelligence. Independently intelligence would mean when separated from the would the arm still shows signs of intelligence. Limb independence is simply central brain with limbs that are concurrently able to independently and meaningfully able to function, either in unison or separately; for example how a human drummer can use four limbs to tap out four separate beats, each with different timings, but that are in sync.
That's reddit-level submission spam[1]. Altough single user account with multiple links to the same content[2] would possibly get filtered there.
[1] https://news.ycombinator.com/from?site=nautil.us
[2] https://hn.algolia.com/?q=We+Might+Already+Speak+the+Same+La...
In any event it might be a loose association, but I do agree with the overall _title_ (and that Octopuses are pretty cool). I hope we're able to recognize any intelligence way may create if we do so successfully.
My Octopus Teacher https://www.youtube.com/watch?v=3s0LTDhqe5A
Small discrepancy which might confuse exactly the readers who would enjoy the whole work.
One day we will bring these parts together into single systems, but for now it's much easier to train and test these independent 'brains' until we figure out a proper architecture. There are quite a few different things we need to answer, for example, can motion planning and motions be de-coupled? Can vision detection be de-coupled from object classification? Can hearing be de-coupled from vocalization?
Those don’t need to be singletons, you can have multiple of each.
And this is, to a large extent, how organized groups of humans work. Why don’t we recognize them as intelligent beings?
The other thing is that I think humans greatly and consistently underestimate the intelligence of almost all animals.