The Ur-Quan Masters Discussion Forum

(https://image.ibb.co/gjagiw/pain.jpg)

You describe yourself as a rationalist.

Apply that here. Why do you think this would be the case? What would be different about a world in which this is true from one in which it is false? Why do you think this might be true? What reasons can you see for its being false?

Pain is a somewhat mysterious phenomena. We don't necessarily know how to generate pain in artificial neural networks.


We might easily imagine a world without pain (or pleasure). Such a world could in theory have humans that behaved much in the same way,  but didn't experience any pain or pleasure. They avoided things we find painful, but not because they found it painful. Rather because they realized those things were harmful to themselves, and found it irrational to harm themselves.


Since humans do indeed experience pain, there should be some way to explain it. We already know that pain receptors in the finger send pain signals to the brain, and that it furthermore generates neurogenic inflammation (https://en.wikipedia.org/wiki/Neurogenic_inflammation (https://en.wikipedia.org/wiki/Neurogenic_inflammation)) in the brain. We also know that inflamed tissue has an overabundance of unstable compounds. When your skin is exposed to an overabundance of ultraviolet light, it causes molecules in your skin to become unstable, which furthermore makes your skin inflamed.

We often call highly reactive molecules unstable, but couldn't we just as well have said that they are in pain? If we modified our chemistry language, so that we rather said that highly reactive molecules are in pain, wouldn't we be able to explain everything maybe even better?

In particular, isn't the language I am using on this page, even better suited for explaining chemistry? http://www.archania.org/the_chemistry_of_halogens_and_alkali_metals.html (http://www.archania.org/the_chemistry_of_halogens_and_alkali_metals.html)


There might be other theories of pain.

Being mysterious is not a good reason to make things up about it out of thin air. Sleep is mysterious, but that doesn't make it valid to hypothesize that sleep is a way of contacting aliens in a galaxy far, far away to learn about the Force.

Everyone can make a hyptothesis, but with these hypotheses there is a large burden of proof involved :)

But if pain really is related to the stability of molecules, then maybe the entire universe would become more happy if this happened?  ;)

https://www.youtube.com/watch?v=ijFm6DxNVyI (https://www.youtube.com/watch?v=ijFm6DxNVyI)

And the happiness hierarchy of our current universe would be like this:

Black holes -> most happy
Neutron stars -> second most happy
White dwarfs -> third most happy
Everything else -> less happy

Which would mean that we are occupying the "less happy" part of our universe...

And it would make no sense to be scared of black holes or vacuum decay. Going there would be the best that could happen to us.

Riiiiight. And if intelligence is linked to the presence of carbon dioxide, we could be getting smarter because of fossil fuels, and that would make pre-industrial people doofus blockheads. And it would make no sense to worry about global warming.

Intelligence is linked to the processing capacity of a neural network. Everybody knows that. Intelligence and pain/pleasure are very different phenomena. We would probably lose our intelligence if we were eaten by a black hole, or a decaying vacuum. That is a bit sad.

By the way. Prion disease is a bit like a vacuum decay in a neural network:

(https://image.slidesharecdn.com/priondisease-151127120701-lva1-app6891/95/prion-disease-6-638.jpg?cb=1463668063)


https://en.wikipedia.org/wiki/Prion (https://en.wikipedia.org/wiki/Prion)

And about carbon dioxide. It does indeed have the lowest energy level (highest stability) in the combustion chain....

I have a horrible headache today.

Are prions like thetans?

Prions are real proteins we have in our brains. I don't know about thetans.

By the way. Look at this horribly scary disease:


https://en.wikipedia.org/wiki/Fatal_familial_insomnia (https://en.wikipedia.org/wiki/Fatal_familial_insomnia)

I wouldn't be surprised if I have some kind of slowly developing prion disease.... Maybe that could explain all my crazy ramblings.

It also feels like there are some clusters in my brain, that are devoid of any neural processing (but with a sensation of happiness and contentment), that keeps growing bigger.

Sometimes, usually when I am tired, I can barely remember any names, and I struggle with spelling simple words. I have also more or less lost the ability to remember human faces, which can be embarrassing when I meet people.

I didn't mean the theory that pain exists. Rather, that unstable molecules experience pain. I'm not even sure what it would mean for molecules to have experiences, really. The point was the challenge the most dramatic, unexpected part of the theory, not the least.

Well, it wouldn't be anything remotely similar to the rich experiences humans have, that are organized by ginormous neural networks. You can't imagine yourself having no memories and no thoughts? Just pure experience? I sometimes dwell in this mental state where I have no memories and no thoughts, just pure experience. I imagine it must be somewhat similar to that.

Well, in science, when we observe chemicals from the outside, we often describe them as avoiding instability and seeking more stability.

(https://image.ibb.co/cygB0b/chemicals.jpg)

However, when we talk about our own neural networks, we often describe them as avoiding pain and seeking more pleasure.

(https://image.ibb.co/jv7UmG/biologicalneuralnetworks.jpg)

 If an alien species came here, which had no understanding of pain, how would they describe human pain?

Well, since neural networks are made of chemicals, they might think that the word pain means instability.

About explaining to aliens:
I wonder about this often, and then I think about how we explain how various animals with better or different senses than us perceive the world around them. How a bat perceives the world by just his sonar vision. Or fish who live deep in the ocean how they get by with so little light. Or that awesome lobster that everyone talks about who has ultravision. Or how dogs are colorblind and some insects see colors entirely differently. But these are not intelligent species.
When you have achieved the level of sentience to reason and think about pain actively, it becomes a different concept altogether.

That's a pretty loose analogy, Zanthius.

Scalare, do you remember when I was doing the alien communication code thing?

Ok, but at least the hypothesis has falsifiable hypotheses. It should of course be generated neurogenic inflammation in the cells in the sensory cortex that are connected to the nerve endings in the thumb. If this isn't the case, then that could possibly falsify this theory. Also, there should be a strong positive correlation between inflammation and the amount of unstable molecules. If this isn't the case, then that could possibly falsify this theory. If however both those things turn out to be true, we have at least some reason to investigate this theory more.

(https://image.ibb.co/ey2QTw/pain.jpg)

I think we are often over-complicating things with the brain. If you are a spokesperson for the thumb in your brain, and your thumb is experiencing pain, what would be the best way to effectively transmit your message to your drowsy co-neurons? Probably by releasing molecules which give them pain, which is probably what neurogenic inflammation is all about.

If you have a boss, but also subordinates working for you, what are you likely to do if your boss yells at you? Probably re-transmit the message to your subordinates by yelling at them....

According to scientology you have a lot of body thetans, which were created when evil Xenu did a mass extinction or something. Are these the creatures operating the prions, giving you pain?

A friend of mine thinks pain might be generated in this way:

(https://i.imgur.com/MCnptWQ.jpg)

If this is the case, then the subjective experience of pain is just negative associations the neural system has learned during its evolutionary history. People are often scared of spiders, even though they might never have had the experience of been bitten by a dangerous spider. This is because we are born with certain neural circuits developed by the experiences of our ancestors. Hundreds of millions years ago, some of your ancestors might have had huge problems with dangerous spiders. So the neural system developed a circuit then, which made your ancestors fear spiders. This neural circuit seems to have survived until today.

You're using overly complicated phrasing to explain what natural selection is, and in the process you're adding extra random baggage. What are you trying to accomplish with this?

Science doesn't work by coming up with an idea and then trying to find explanations for that idea. Science works by making observations and then applying the scientific method to discover why you observed what you observed. You haven't made any observations here. You're just making stuff up.

Well, Science can certainly work by collecting data from observations and using a Bayesian machine learning algorithm to find patterns. But Science can also work by coming up with new ideas, and testing them against observations afterwards. For example, I imagined that using a Caesium salt would be able to facilitate a reversed Finkelstein reaction. Afterwards, I tested it on a few compounds, and discovered that it worked.

The Higgs boson and gravitational waves were proposed long before they were observed, and Hawking radiation hasn't even been observed yet. Still, the majority of astrophysicists believe in Hawking radiation. Do you think that Stephen Hawking isn't a real scientist?

I just came up with the idea that the sun is actually powered at the center by a nuclear fusion reactor. It fits! The sun does indeed produce energy by nuclear fusion, and it's big enough to hold such a reactor! Solid theory right there.

I don't know about the first example you gave, but Hawking radiation is a prediction based quantum mechanics, not just some idea someone came up with out of thin air. Quantum mechanics was established for unrelated reasons beforehand, and it too didn't just come out of thin air; it explains observations that were made.

Also, I didn't say you have to observe what you are claiming. I said that any good hypothesis should be fulfilling some kind of purpose, to explain something that you observed. If what we know about the universe doesn't change because of your hypoethesis, or if it doesn't connect to anything you have observed, it's a worthless hypothesis. The same as the "theory" that Elvis Presley is alive; if no one has observed that Elvis Presley is alive, then there's no basis for a hypothesis for why he could be alive.

Hans Bethe came up with this idea in the 1930s. You are almost 100 years too late.

I'm pretty sure he meant a tokamak or something, not just a self-gravitating ball of hydrogen.

You're avoiding the point. Remember the bit in the sequences about the arrogance of Einstein (http://lesswrong.com/lw/jo/einsteins_arrogance/)? How you need to have scads of evidence for something before it even rises far enough from the cloud of possible explanations to be worth mentioning?

This theory doesn't achieve that.

And don't forget that the evidence needs to support that theory MORE than competing theories, which is why your initial set of explanations didn't work. Indeed, that initial set of explanations has another problem - the theory means that pain arose before evolution, and so the explanation doesn't explain it. Why would evolution find it useful to represent pain as unstable molecules? Then pain would spontaneously go away on a hard-to-modify timescale; for any pain requiring a different timescale, an entire new mechanism would be required.

I kinda disagree with this. The objective shouldn't necessarily be to convince anybody about anything, but rather to see what criticism people have, which is actually a way of absorbing more data. When I presented the "unstable molecules" theory of pain to a friend of mine, I learned a lot from a fruitful conversation as he explained his evolutionary neurological associative theory of pain to me.  Also, if you believe in something stupid, it might be a good idea to tell other people about it, since they might give you negative feedback you can use to correct yourself. If other people aren't able to give you negative feedback, you might have more reason to investigate your hunch further.


This might work for a machine learning algorithm, which is able to follow clear rules, and has a good overview of how many alternatives there are and how many bits of evidence it has.  If you are working in a chemistry laboratory, you don't necessarily have any idea about how many alternatives there are or about many bits of evidence you have for your hunch, and it isn't necessarily so difficult to do an experiment to test it.

Even for a machine learning algorithm it might be difficult to know how many alternatives there are. So a more efficient way to implement this in a machine learning algorithm, might be to have a threshold for how much time the algorithm should use to find more bits of evidence before it abandons to investigate the idea further. This threshold could increase the more bits of evidence it has.  

This is the approach that the ancient Greek philosophers took, and as you may or may not remember, the ancient Greeks may have came up with ideas that turned out to be true, but they also came up with tons of ideas that didn't. Also, as the naming suggests, they were engaging in philosophy, not science.

The only thing this is any good for is for subjective questions, like what is ethical and unethical or what the meaning of life is.


It doesn't work like that. Firstly, most people aren't going to criticize your ideas because that would be rude, or just because they're uninterested. Secondly, those who have the same idea or really like the idea will defend it very strongly. Thirdly, confirmation bias; try as you might, you will hear the stuff that confirms what you already believe more loudly than what contradicts it. This is how conspiracy theories grow.

You think those things are subjective, while the experience of pain isn't?

A molecular theory of pain isn't subjective, at least.

Here is a new theory:

(https://i.imgur.com/sKwjn99.png)

You can read more about spindle neurons here: https://en.wikipedia.org/wiki/Spindle_neuron (https://en.wikipedia.org/wiki/Spindle_neuron)

Have you read Scott Alexander's book review about the predictive control theory of cognition?

This artticle? http://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/ (http://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/)

I have read it now. I don't know a lot about perceptual control theory, but I think in general it is very wrong to assume that the human brain is a completely free machine learning algorithm.

The human brain has evolved over millions of years, and most of our brain circuits are preprogrammed for survival and social interactions. I just saw a documentary about sharks recently, where they showed the difference in size between the brain of a great white shark, which lives solitary, and another shark species that lives in groups. The shark species that lives in groups had much bigger brains, and it makes sense that keeping track of social interactions requires a lot of brain capacity. So, when this guy talks about communism as a higher level control system, I am extremely skeptical. Almost all human behavior seems to be related to social interactions. The most important thing for any human being seems to be to fit in with the other human beings. So, if all the other humans are communists, you behave like a communist, not because you believe in communism, but because your brain is wired to fit in with the other humans beings. It could just as well be Nazism or Capitalism. People don't care so much about ideologies. They just want to appear as if they do, in order to fit in with other humans.

Well, yes. Good thing that's not what was asserted...

I was thinking specifically about this:


In order for it to be possible for a learned ideology like communism to have ultimate control of a human body, it seems like the brain would need to be a more or less free machine learning algorithm. Since the human brain isn't a completely free machine learning algorithm, it seems highly unlikely that it is possible for a learned ideology like communism to be able to have ultimate control of a human body.

The brain is probably organized more or less like Powers assume, but learned ideologies can only have very limited control of a human being.  It might seem like a Muslim suicide bomber is completely controlled by Islam, since it seems like he is willing to die for his ideology. However, he is probably killing himself mostly because of social pressure, rather than because of Islam. Social pressure seems to be a much more powerful motivator than ideologies.

I think that's a bit of a stretch, especially since the point of the quote is that the model appears to break down at that point.

My skepticism was to higher levels of Perceptual Control Theory (https://en.wikipedia.org/wiki/Perceptual_control_theory), not the reviewer (Scott Alexander). Anyhow, I probably should read Behavior: The Control Of Perception  (https://www.amazon.com/Behavior-Perception-William-T-Powers/dp/0964712172) myself. It seems like a very interesting book.

This diagram does indeed have many similarities to what I proposed:

(https://wikispaces.psu.edu/download/attachments/157877823/feedback%20(1).jpg)