Transhumanism

[Tory]

Why do our brains need to be turing machines in order for the singularity to be possible? Our brains have nothing to do with the singularity directly, our only involvement with the singularity is to initiate it by creating an intelligent machine which can self-improve and create more intelligent machines, and the rest will be history.


The singularity isn't a silly idea, it's practically inevitable under the assumption that humans don't destroy themselves in the near future before it is achieved. The median year predicted for it to occur among AI researchers is 2050, a mere 30 years away. If a bunch of action potentials traveling through a mushy blob of neurons can generate human level intelligence, there's no reason to believe that electrical signals traveling through silicon can't also. A silicon brain, however, would operate much faster, more reliably, have a much greater memory and computational capacity, wont need to eat, sleep or get distracted in the ways that humans do, and can be mass produced in the trillions. From there, technological and industrial progress will clearly be insanely rapid.

OldM8? Is that really you? I missed you so much, welcome back. First Ashaebi came back, now you. This is a dream come true.

My favorite intelligent technology for transhumanism is love robots. You won't be forced to pay child support for children that aren't yours. No false allegations of domestic violence, sexual assault, and child abuse. Not accused of being a premature ejaculator. No alimony, no unwanted pregnancies, and no sexual transmitted diseases.

Also, does artificial wombs count as a transhumanism concept? If so, this would be my second favorite.

 

[OldM8]

If you are trying to build something that is in all ways superior to a human brain out of a turing machine, that will be impossible if the human brain is not a turing machine.

The point isn't whether the brain is a turing machine or not, it's whether it can be simulated by one. For example, push-down automata and finite state machines are examples of machines which aren't turing machines, but they can be simulated by one. Since the anatomy of the brain and its functionality can be represented as binary data and functions which perform operations on that data, it can be simulated by a turing machine.

I guess you might be able to build something that for all intents and purposes is better than a human brain, but, that being the case, it does kinda raise the question of how humans managed to evolve a super-turing machine when a turing machine is more than capable of doing the job. I think its reasonable to assume that if human brains arent turing machines, then there is some reason they need to be more than a turing machine.

Here you are pre-supposing that the brain is super-turing, when in fact it is not. It's important to keep in mind that from a purely physical point of view, there's nothing really special about the human brain. Ultimately, it's a jumble of atoms obeying the laws of physics, just like everything else is. So theoretically, if we had a turing machine with a huge amount of memory to store the positions and velocities of all the atoms that comprised the model of our brain, we could simulate it using the basic laws of particle physics. In other words, the brain is just a lump of matter with emergent properties like all other machines are, including computers.

And the singularity is one of those things its kinda impossible to argue against, cause you never really know what it means when someone is talking about it. If you are talking about the idea of a point in time where technological growth improves exponentially, then we already passed that.

The singularity almost universally refers to the point where humans create an artificial intelligence that surpasses the intelligence of humans, which is when the slope of the exponential curve increases dramatically to a much steeper level than now.

But basically, I will back mushy blobs of neurons over silicon pretty much every time. Silicon is just not very good at anything other than doing really dumb stuff really quickly.

Silicon is merely a substrate upon which information is stored and manipulated, just as neurons are. You seem to be conflating silicon and the computing framework that is implemented upon it. Current computer chips only do "dumb stuff" because that's all they have been designed and programmed to do. There are plenty of mushy blobs of neurons in nature which are also quite stupid (like snails, fish, etc) but that doesn't mean that neurons can't create human level intelligence. The development of strong AI requires the intersection of advanced hardware and software. Once we learn what algorithms/processes are responsible for producing general intelligence, we can implement them in dedicated hardware and we will witness the genisis of the singularity.

Do you think it's possible that people's tendency to get distracted is actually helpful for technological progress?

Not really. Humans get distracted and procrastinate primarily because of the way our biological reward system works. We seek experiences which release dopamine, which is why we frequently switch to watching youtube videos or playing games instead of focusing on the task at hand.

 

[termi]

OldM8? Is that really you? I missed you so much, welcome back. First Ashaebi came back, now you. This is a dream come true.

My favorite intelligent technology for transhumanism is love robots. You won't be forced to pay child support for children that aren't yours. No false allegations of domestic violence, sexual assault, and child abuse. Not accused of being a premature ejaculator. No alimony, no unwanted pregnancies, and no sexual transmitted diseases.

Also, does artificial wombs count as a transhumanism concept? If so, this would be my second favorite.

i take it you're very popular with the ladies

 

[Hipmonlee]

The point isn't whether the brain is a turing machine or not, it's whether it can be simulated by one. For example, push-down automata and finite state machines are examples of machines which aren't turing machines, but they can be simulated by one. Since the anatomy of the brain and its functionality can be represented as binary data and functions which perform operations on that data, it can be simulated by a turing machine.
Here you are pre-supposing that the brain is super-turing, when in fact it is not.

Just want to point out that it is mildly amusing to me that where I have clearly stated that my initial post was based on an assumption that the human brain is a turing machine, and haven't actually made any real speculation one way or another, I am somehow being challenged by both sides of that argument in this thread.

But just to clear things up, the original thing you quoted with emphasis added.

And if our brains aren't turing machines then at least I guess the singularity is an impossibility.

But anyway, I should thank you for proving my earlier point that basically all transhumanists think that the human brain is a turing machine.

It's important to keep in mind that from a purely physical point of view, there's nothing really special about the human brain. Ultimately, it's a jumble of atoms obeying the laws of physics, just like everything else is. So theoretically, if we had a turing machine with a huge amount of memory to store the positions and velocities of all the atoms that comprised the model of our brain, we could simulate it using the basic laws of particle physics. In other words, the brain is just a lump of matter with emergent properties like all other machines are, including computers.

I am not entirely convinced it is true. I am not suggesting that a human brain might have metaphysical properties, just that a turing machine has a bunch of limitations that theoretically a brain might not be constrained by if it operates in some kind of fundamentally different model. As far as I am aware, no one knows that for certain whether that is the case or not.

The singularity almost universally refers to the point where humans create an artificial intelligence that surpasses the intelligence of humans, which is when the slope of the exponential curve increases dramatically to a much steeper level than now.

Eh, the steepness of an exponential growth curve is kinda irrelevant. But a good rule of thumb when dealing with exponential curves in real life: they never last. Can't say when this one will end, but it might be sooner than you think.

Silicon is merely a substrate upon which information is stored and manipulated, just as neurons are. You seem to be conflating silicon and the computing framework that is implemented upon it. Current computer chips only do "dumb stuff" because that's all they have been designed and programmed to do. There are plenty of mushy blobs of neurons in nature which are also quite stupid (like snails, fish, etc) but that doesn't mean that neurons can't create human level intelligence. The development of strong AI requires the intersection of advanced hardware and software. Once we learn what algorithms/processes are responsible for producing general intelligence, we can implement them in dedicated hardware and we will witness the genisis of the singularity.

My point was really just that computers are really not very close to competing with human brains at this point in time.

Not really. Humans get distracted and procrastinate primarily because of the way our biological reward system works. We seek experiences which release dopamine, which is why we frequently switch to watching youtube videos or playing games instead of focusing on the task at hand.

Have you ever watched a youtube video or played a videogame where you didn't learn anything? What even is the task at hand when it comes to advancing technology? I suspect this stuff is actually pretty useful when it comes to any form of creativity.

But to be a little more direct about my point here. I think that a general intelligence machine is soooo different from our current computers that assuming that they are going to operate in a similar fashion, and retain all the benefits of current computers seems a little unlikely to me. It makes more sense to me to expect them to act similar to the actual general intelligence machines we have all around us.

 

[OldM8]

Just want to point out that it is mildly amusing to me that where I have clearly stated that my initial post was based on an assumption that the human brain is a turing machine, and haven't actually made any real speculation one way or another, I am somehow being challenged by both sides of that argument in this thread.

A turing machine is a theoretical construct. The human brain from a computational standpoint is actually less powerful than a turing machine because human brains only have a finite memory, whereas a turing machine has an infinite one. The human brain implements algorithms which we currently do not entirely understand, but that is completely beside the point. Once we understand how the human brain works, those algorithms could be run on a turing machine. Indeed, we have already simulated small parts of the human brain to incredible detail and have verified that the simulation behaves in accordance to experiments in neuroscience.

But just to clear things up, the original thing you quoted with emphasis added.

And if our brains aren't turing machines then at least I guess the singularity is an impossibility.

Why do our brains have to be turing machines in order for the singularity to be a possibility?

But anyway, I should thank you for proving my earlier point that basically all transhumanists think that the human brain is a turing machine.

You still don't get it. They don't think the brain is a turing machine, because it isn't. They think that the brain can be simulated by a turing machine and that there is nothing mystical about it.

My point was really just that computers are really not very close to competing with human brains at this point in time.

They're a lot closer than you might think. The big tech companies are investing heavily in specialized computer architectures for AI and research in the field is booming. If AGI is even 30 years away, that's very close considering that it will bring about the most profound change of all time.

 

[Hipmonlee]

Ok, to clarify, in my post when I refer to the brain as a turing machine, I mean "a thing that can be simulated on a turing machine".

Earlier in the thread Myzozoa and I discussed whether it is possible that the brain is something that could not be simulated by a turing machine. It is plausible that a human brain is in some way a more sophisticated form of computation engine, eg. a device that takes continuous input and generates continuous output. This would be impossible to perfectly simulate on a turing machine, because a turing machine can only operate on discrete data. Any simulation would be necessarily lossy.

I do not really have a strong view either way here, except to say that if the human brain can be simulated on a turing machine, then with a couple of other assumptions which I think aren't completely outrageous, then you can draw some pretty wild conclusions! As seen in my first post.

 

[OldM8]

Earlier in the thread Myzozoa and I discussed whether it is possible that the brain is something that could not be simulated by a turing machine. It is plausible that a human brain is in some way a more sophisticated form of computation engine, eg. a device that takes continuous input and generates continuous output.

That's not plausible because we've known for a long time that's not how the brain works. The brain does not compute on continuous inputs or data, it computes on discrete events called action potentials (spikes). All inputs to the brain from light, sound, etc are converted into spike trains by our sensory organs, which are basically biological analog-to-digital converters, much like how cameras and other electronic sensors work. All computations are then performed using these discrete units of information.

This would be impossible to perfectly simulate on a turing machine, because a turing machine can only operate on discrete data. Any simulation would be necessarily lossy.

It doesn't have to be simulated perfectly, it can be simulated to an arbitrary degree of precision, much like calculating pi to trillions of decimal places. All that is needed is for it to accurately replicate the observed functionality and behavior of the brain.

 

[cookie]

When you think about it, the fact that no extraterrestrial being has ever contacted us (as far as we know), considering the vastness of the universe related to its old age, is just stunning, and can only mean 2 things.

not necessarily: there's an obvious bias towards our existence (because we wouldn't be here discussing this topic if we didn't), but to calculate the odds of life existing elsewhere you have to multiply a very large number (the size of the universe) by a very small one (the odds of it arising in any one place). The first number being large isn't enough - the second number needs to be sufficiently big too. All that existence of life here on Earth proves is that the latter number is non-zero, but it could be so laughably small (laughinggirls.jpg small) that the a priori average number of civilisation- or even life-bearing planets is going to be very, very close to zero such that the odds are massively in favour of there being zero life rather than life on even one planet.

1) Either we are alone in the universe, which, when you think about it, means that we were certainly created by a God. Indeed the probabilities that we are alone in the universe by pure chance are absolutely microscopic.

thought about it, didn't reach the same conclusion