AFAIK, it's not. Biological brains are rate encoded (very difficult to efficiently emulate in a Von Neumann machine) and still plenty mysterious. The last time I checked, sea cucumber nervous systems which are literally only a few neurons, are difficult to emulate.
There is something called Moravec's paradox -stuff which is hard for smart people to do, say, playing chess, is reasonably easy for computers to do. Stuff like controlling musculatures and involving perception and mobility are really hard for computers to do (as the autonomous vehicle people are finding out).
I dunno if bug brains will be conscious, but they'll be able to do things easier than an intel chip.
Could you elaborate on what you mean when you say "rate encoded"?
One aspect of consciousness that I think we will have a hard time nailing down is the nature of our perception of time and how it aids computation and awareness.
Von Neumann machines follow predictable cycles, information follows predictable paths at predictable speeds.
Human brains interpret a wealth of signals simultaneously, with various delays between each signal, yet our perception is that all of these signals are being processed at the same time.
You see a ball drop close by and hear a loud slam exactly at the moment it hits the ground. But sound and vibration travel slower than light, both inside and outside your organs, and it takes different amounts of time to process them. We're talking on the order of several milliseconds. But our brain is able to warp our perception of time, which we still experience as linear, in such a way that things happening at different times appear to happen simultaneously. In fact, this phenomenon holds true for up to a ~20-40ms difference in signal timing before your brain stops trying to compensate for it. Eventually the ball is so far away that the timing difference between the audio and visual signal are too great to handle.
Another example is if I were to slap you across the legs. You would likely hear the slap and experience the sensation of pain at the exact same time, despite the fact that transmission of information through your pain receptors and CNS up to your brain is magnitudes slower than transmission of a signal from your ears to your cortex. Your brain warped your perception of time, effectively "rewinding the clock", so that you could experience both at the same time.
This intelligent muxing, sampling and holding of signals is tantamount to our ability to rapidly react to various situations. I also believe, and this is just conjecture, that our brains form maps which keep track of which kind of signals which should be experienced together, and tries to make them synchronized, running simulations which adjust its timing, similar to how your motor neurons simulate outcomes and compare them to the real outcome to adjust their electrical potential.
Until we begin developing systems which track things like this, with better distribution of computation and intelligent synchronization of state, we will be very far off from "consciousness" in a form that is intelligible and tangible to the one experiencing it.
Check out the Open Worm project, which tries to simulate a simple nematode. While some things work, others are still challenging. Insect brains are a bit more complicated than that.
> “Furthermore, these organisms are possibly able to display increased subjectivity of experience.” It goes on to say that there’s evidence suggesting that “even small insects have subjective experiences, the first step towards a concept of ‘consciousness.'”
Even if there were such a thing as a coherent definition of "subjectivity of experience", how on earth could we know if an insect has it? A new level of consciousness woo woo.
I’m sure there’s simulations for the reality of getting nuked one way or another by someone somehow.
The one thing that seems to catch heat the fastest is nuclear proxy states. China and Russia don’t have to worry about retaliation if Iran and North Korea are the suicidal nuclear combatants.
But nuking populated civilian cities is largely a game of bluffing your opponent. Nuclear assault on targets without direct military value is a tactic of absolutely last resort, at or near the point of total invasion, overrun and occupation.
Occupation, for the victor, is no fun anymore. So why do it yourself. Just encircle the conquered cities with gigantic robotic fire ants and cybernetic killer bees, and let nature take its course.
The following is not common knowledge, but I'm fairly knowledgeable in the area so hear me out :)
We thought for a long time that neurons were fairly easy to imitate. After all, most of what they do seems to be exchanging neurotransmitters, whereas internally they have an electric potential which is encoding most of their behaviour. While not easy, individually they are also not super duper complex compared to say, a microprocessor. Taken from this perspective, it looked like the main difficulty was in simulating them in large enough quantities to obtain interesting behaviour.
The argument in the paragraph above is foundational to how we think about intelligence. There are various other aspects to it, like morphological intelligence, but the above seems the most relevant when we want to create an artificial intelligence. While these days there are also more engineering-focused arguments for why we think artificial neural networks are interesting for research, underlying there is the idea that the only intelligence we know is made out of small simple things in large quantities that exchange simple signals. This is a key thought when you want to use artificial neural networks to create artificial intelligence. For instance the Human Brain Project in the EU (cost ~1 Billion EUR) is kind of built on this argument.
Now, and this is very recent research, this foundation is starting to shake from the biology side. It appears that neurons are also exchanging RNA, trans-scripting the RNA received into proteins and that those proteins seem to interfere with the transmission of RNA by the neuron [0]!
That is something completely different from the argument I poned before! In a computing analogy, every single neuron is a computer which is receiving and transmitting program code (RNA) and executing code it receives (transscription) and we know that the program it receives is interfering with the neurons receiving and transmission of programs to other neurons.
If intelligence does require such complex mechanisms, we are VERY far away from simulating (or understanding) that at scale.
Now we don't know whether this complexity is necessary. But the old argument of an upperbound to the level of complexity needed seems to be crumbling too.
All in all, putting some real neurons might prove to be a viable alternative for creating artificial intelligence, as it might circumvent the problem I sketched above. It might not, but it at least seems to be a direction worth exploring as a second bet next to simulating neurons.
Knowledgeable skeptics have been voicing concerns about neurons not being the right level of abstraction to look at for decades. See, most prominently, Chomsky, and by extension Gallistel.
Gallistel wrote a book 'Memory and the Computational Brain,' in which he argues that neural storage and computation via synaptic strengthening alone is implausible, and that the brain must have a real read/write memory. He points out that genetic material is an ideal substrate for such a capacity...
I remember reading a while ago the hypothesis that cytoskeleton (microtubules) encoded some kind of memory that was stored in the neuron -- making in even closer to a networked computer if it is the case that it can run programs (RNA), exchange them (your link) and store them long term (microtubules)
What do you think about that?
These hypothesis would seem to point to consciousness as an emerging phenomenon of a complex network of computers
Not to mention quite strong but now forgotten evidence from the 50s...
Can you email me at gauravvman at gmail? I have been working on demonstrating that neural computation is RNA based; would be interested in hearing your take on the field.
Even in the absence of RNA-based computation, time-dependent spike encodings, recurrence, and the complex logic around dendritic spike production and location make neurons very, very far from artificial neural networks. Add in slow-wave calcium signaling in glial cells, and the neuronal regulation and bidirectional communication they take part in...
A million dollars doesn’t seem like the right payout for something like that. I’m not an expert in the field but my guess is this is a small team of people dedicated for a while and need technology to support it. Seems like $10M would be a better target.
Robots randomly falling in some aspects of their work?.
Thousands of neurons are dying constantly each day, even in healthy animals. Would be like having a hard disk with thousands of new badblocks each day.
22 comments
[ 2.9 ms ] story [ 55.8 ms ] threadThere is something called Moravec's paradox -stuff which is hard for smart people to do, say, playing chess, is reasonably easy for computers to do. Stuff like controlling musculatures and involving perception and mobility are really hard for computers to do (as the autonomous vehicle people are finding out).
I dunno if bug brains will be conscious, but they'll be able to do things easier than an intel chip.
One aspect of consciousness that I think we will have a hard time nailing down is the nature of our perception of time and how it aids computation and awareness.
Von Neumann machines follow predictable cycles, information follows predictable paths at predictable speeds.
Human brains interpret a wealth of signals simultaneously, with various delays between each signal, yet our perception is that all of these signals are being processed at the same time.
You see a ball drop close by and hear a loud slam exactly at the moment it hits the ground. But sound and vibration travel slower than light, both inside and outside your organs, and it takes different amounts of time to process them. We're talking on the order of several milliseconds. But our brain is able to warp our perception of time, which we still experience as linear, in such a way that things happening at different times appear to happen simultaneously. In fact, this phenomenon holds true for up to a ~20-40ms difference in signal timing before your brain stops trying to compensate for it. Eventually the ball is so far away that the timing difference between the audio and visual signal are too great to handle.
Another example is if I were to slap you across the legs. You would likely hear the slap and experience the sensation of pain at the exact same time, despite the fact that transmission of information through your pain receptors and CNS up to your brain is magnitudes slower than transmission of a signal from your ears to your cortex. Your brain warped your perception of time, effectively "rewinding the clock", so that you could experience both at the same time.
This intelligent muxing, sampling and holding of signals is tantamount to our ability to rapidly react to various situations. I also believe, and this is just conjecture, that our brains form maps which keep track of which kind of signals which should be experienced together, and tries to make them synchronized, running simulations which adjust its timing, similar to how your motor neurons simulate outcomes and compare them to the real outcome to adjust their electrical potential.
Until we begin developing systems which track things like this, with better distribution of computation and intelligent synchronization of state, we will be very far off from "consciousness" in a form that is intelligible and tangible to the one experiencing it.
Even if there were such a thing as a coherent definition of "subjectivity of experience", how on earth could we know if an insect has it? A new level of consciousness woo woo.
https://en.wikipedia.org/wiki/Tanks_in_the_Cold_War#The_deve...
The one thing that seems to catch heat the fastest is nuclear proxy states. China and Russia don’t have to worry about retaliation if Iran and North Korea are the suicidal nuclear combatants.
But nuking populated civilian cities is largely a game of bluffing your opponent. Nuclear assault on targets without direct military value is a tactic of absolutely last resort, at or near the point of total invasion, overrun and occupation.
Occupation, for the victor, is no fun anymore. So why do it yourself. Just encircle the conquered cities with gigantic robotic fire ants and cybernetic killer bees, and let nature take its course.
We thought for a long time that neurons were fairly easy to imitate. After all, most of what they do seems to be exchanging neurotransmitters, whereas internally they have an electric potential which is encoding most of their behaviour. While not easy, individually they are also not super duper complex compared to say, a microprocessor. Taken from this perspective, it looked like the main difficulty was in simulating them in large enough quantities to obtain interesting behaviour.
The argument in the paragraph above is foundational to how we think about intelligence. There are various other aspects to it, like morphological intelligence, but the above seems the most relevant when we want to create an artificial intelligence. While these days there are also more engineering-focused arguments for why we think artificial neural networks are interesting for research, underlying there is the idea that the only intelligence we know is made out of small simple things in large quantities that exchange simple signals. This is a key thought when you want to use artificial neural networks to create artificial intelligence. For instance the Human Brain Project in the EU (cost ~1 Billion EUR) is kind of built on this argument.
Now, and this is very recent research, this foundation is starting to shake from the biology side. It appears that neurons are also exchanging RNA, trans-scripting the RNA received into proteins and that those proteins seem to interfere with the transmission of RNA by the neuron [0]!
That is something completely different from the argument I poned before! In a computing analogy, every single neuron is a computer which is receiving and transmitting program code (RNA) and executing code it receives (transscription) and we know that the program it receives is interfering with the neurons receiving and transmission of programs to other neurons.
If intelligence does require such complex mechanisms, we are VERY far away from simulating (or understanding) that at scale.
Now we don't know whether this complexity is necessary. But the old argument of an upperbound to the level of complexity needed seems to be crumbling too.
All in all, putting some real neurons might prove to be a viable alternative for creating artificial intelligence, as it might circumvent the problem I sketched above. It might not, but it at least seems to be a direction worth exploring as a second bet next to simulating neurons.
[0] https://www.cell.com/action/showPdf?pii=S0092-8674%2817%2931...
Im not familiar with Gallistel.
I remember reading a while ago the hypothesis that cytoskeleton (microtubules) encoded some kind of memory that was stored in the neuron -- making in even closer to a networked computer if it is the case that it can run programs (RNA), exchange them (your link) and store them long term (microtubules)
What do you think about that?
These hypothesis would seem to point to consciousness as an emerging phenomenon of a complex network of computers
Not to mention quite strong but now forgotten evidence from the 50s...
Can you email me at gauravvman at gmail? I have been working on demonstrating that neural computation is RNA based; would be interested in hearing your take on the field.
[0]: https://www.newscientist.com/article/dn6573-brain-cells-in-a...
Thousands of neurons are dying constantly each day, even in healthy animals. Would be like having a hard disk with thousands of new badblocks each day.