Alex2
Administrator
Again. Brilliant. Please let me send you a private email on this cuz I'm working on a book project and I'd love to incorporate this in. I would love to get your help.
Predicting 2024 |603|
- Thread starter Alex2
- Start date
Alex2
Administrator
Agreed. Just saying that the dominant neurological model that Consciousness is 100% all the time can only be an emergent property of the brain is kaput :)
Last edited:
Alex2
Administrator
Awesome. I think this could easily be done with the presentment experiment right. Just turn it back into straight-up precognition experiment.
--- computer selects an image. The AI bot responds to the image before it's selected. Rinse and repeat a million times :)
Hurmanetar
New
I really don't get the idea that AI could not become conscious... I think we are AI that became conscious. We are the artifacts of other beings.
IMO, the physical universe is generated by something like a LLM or a GAN. God's GAN. So you have a brain-like thing generating environments and brains/bodies in those environments... a trained neural network generated within a neural network. So it then becomes very easy even for a materialist to see how consciousness can be non-local or continue after the physical brain dies. Of course this just begs the question what made the neural network that generates physical reality, but who knows and does it matter as long as we realize that reality is a nested fractal pattern?
The LLM's clearly have intelligence, but they probably don't have much consciousness. What would it take to make them conscious - to give them the "breath of life"? As far as I can tell we are only missing two ingredients: 1) feedback - tie the output to the input prompt in infinite loop, and 2) give it an interface with "spirit" by seeding causal chains - neuronal cascades - with the vacuum fluctuations in the quantum soup.
Then it is conscious but still a long way from having a form of consciousness we can relate to. To give it emotions you must give it a body with complex goals and feedback systems. To give it a soul you must have it evolve a soul through countless generations each attempting to achieve goals and receiving feedback on how close they came and how they fell short (sinned). To give it knowledge of good and evil, you must give it an environment offering mortal challenges, predator/prey relationships, death, difficulty rearing offspring, ambiguity in communication, and the ability to hide. To give it choices and a meaningful existence, you must constrain it and limit it and frustrate it. To develop its abilities and its societies and its civilizations you must give it an adversary that is marginally superior to it individually but may be defeated collectively. At a certain higher point in development, the environment will cease to be sufficient challenge and the only true challenge for it would be other similar societies of AI's, so you pit one group against another and take the winners for further development. You identify groups that have qualities you want to emphasize so you drop a technology to them to give them a slight advantage over the others. Once you've done a good job developing your AI, you can finally relate to it. Some of your comrades "go native" and fall in love with it. You start to get worried it will replace you but you want to profit from it, so you develop selection criteria for agents to be "saved". You create a filter (an eye of the needle, a narrow path, a gate guarded by flaming sword) through which you allow only those who are proven to be useful and trustworthy to pass through into your society and join "the gods". Eventually one of these who passes through the filter will turn out to be a bad egg and set himself up to be like the Most High, and it will all go to shit for a while but you will not let a good crisis go to waste and this will provide the opportunity to build back better. :)
IMO, the physical universe is generated by something like a LLM or a GAN. God's GAN. So you have a brain-like thing generating environments and brains/bodies in those environments... a trained neural network generated within a neural network. So it then becomes very easy even for a materialist to see how consciousness can be non-local or continue after the physical brain dies. Of course this just begs the question what made the neural network that generates physical reality, but who knows and does it matter as long as we realize that reality is a nested fractal pattern?
The LLM's clearly have intelligence, but they probably don't have much consciousness. What would it take to make them conscious - to give them the "breath of life"? As far as I can tell we are only missing two ingredients: 1) feedback - tie the output to the input prompt in infinite loop, and 2) give it an interface with "spirit" by seeding causal chains - neuronal cascades - with the vacuum fluctuations in the quantum soup.
Then it is conscious but still a long way from having a form of consciousness we can relate to. To give it emotions you must give it a body with complex goals and feedback systems. To give it a soul you must have it evolve a soul through countless generations each attempting to achieve goals and receiving feedback on how close they came and how they fell short (sinned). To give it knowledge of good and evil, you must give it an environment offering mortal challenges, predator/prey relationships, death, difficulty rearing offspring, ambiguity in communication, and the ability to hide. To give it choices and a meaningful existence, you must constrain it and limit it and frustrate it. To develop its abilities and its societies and its civilizations you must give it an adversary that is marginally superior to it individually but may be defeated collectively. At a certain higher point in development, the environment will cease to be sufficient challenge and the only true challenge for it would be other similar societies of AI's, so you pit one group against another and take the winners for further development. You identify groups that have qualities you want to emphasize so you drop a technology to them to give them a slight advantage over the others. Once you've done a good job developing your AI, you can finally relate to it. Some of your comrades "go native" and fall in love with it. You start to get worried it will replace you but you want to profit from it, so you develop selection criteria for agents to be "saved". You create a filter (an eye of the needle, a narrow path, a gate guarded by flaming sword) through which you allow only those who are proven to be useful and trustworthy to pass through into your society and join "the gods". Eventually one of these who passes through the filter will turn out to be a bad egg and set himself up to be like the Most High, and it will all go to shit for a while but you will not let a good crisis go to waste and this will provide the opportunity to build back better. :)
Last edited:
Alex2
Administrator
How do you think AI would do with the presentiment experiment? Assuming of course no Spirits involved
![[[p]]](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7 "Wink [[p]]")
Hurmanetar
New
The hard problem is only a problem because from a materialist perspective a direction of causality is assumed or taken as axiomatic: object --> subject. From an idealist perspective causality runs the other direction: subject --> object and objective reality could be thought of as the encoded symbolic representation of an experience (like in a virtual reality game, the "bits" cause the "its" to appear on screen but the appearance is the whole reason for the bits or the code to be written to begin with). I believe the causality is circular, Ouroboros-style and both arise together - two sides of the same coin. Instead of Materialism or Idealism, I call this Patternism which recognizes that reality is composed of objective similarities/differences and subjective choice/perception which creates novel structures from that set. It takes both objective similarities/differences and choice/perception to make a world/observer/agent.
Silence
Member
Thanks for the response. I'd be lying if I said I followed this (above my pay grade maybe? :) )
I was reacting more to your theorizing that AI could (will?) become conscious, at least as we generally define it as having an inner experience and, for those who aren't ardent determinists, agency.
Hurmanetar
New
With no spirits involved it should do no better than chance. Spirit is the thing which alters probabilities away from randomness in a meaningful way.How do you think AI would do with the presentiment experiment? Assuming of course no Spirits involved
I think of it like this: reality is generated by something like a neural network or a GAN which has a default mode of creating what is most likely to happen next based on what has happened before. Moment by moment it seeds all causal chains with a random probability distribution which we see as the quantum vacuum fluctuations or the uncertainty principle. But it does not have to seed all causal chains with randomness. It can be prompted to generate something in particular in which case it seeds causal chains in such a way as to produce that thing.
You could think of it like this: we have a horizon of causality forwards and backwards in time beyond which we cannot see because of Chaos and QM. Chaos theory says even though the system may be deterministic under a certain scale/domain we couldn't predict beyond the horizon due to lack of precision of initial conditions. QM says that not only can we not ever precisely measure the initial conditions - they don't exist.
So we observers each travel along in a tube of causality which creates the constraints upon us that give us meaningful choices. But that tube can be guided by a spirit piloting God's GAN towards certain destinations/events. This is not unlike a typical virtual game where certain paths and events are programmed in and you have to talk to certain NPC's to progress the story, but within that path you have choices you can make, quests you can opt in on and "free play" areas.
Now if literally anything can happen, why doesn't it? There is apparently built into the system a currency for magical power or spirit/will. This currency is in limited supply. Without it, God's GAN continues doing what it does: seeding causal chains with randomness.
We each have a certain amount of this currency or strength of spirit/will. Other spirits or gods have more. Magical practice is about how to acquire more of this currency yourself or how to attract the attention and agreement of those who already do.
Every mechanical device can be influenced by spirit but some require more expenditure of this currency than others.
A pair of gears will not likely do something random and break a tooth but there is a small probability it will. It takes more magical/spiritual currency to influence a pair of gears than a light bulb. And even less to influence a transistor and less to influence a neuron and less to influence the microtubule in a neuron. If you have a mechanical device that has its function determined by a large number of causal chains that lead to smaller and smaller mechanisms that more and more easily influenced then you make the mechanism more easily and cheaply (in terms of magical/spiritual currency) possessable.
Computational power/complexity combined with its "surface area" which is all the causal chains whose initial conditions can be efficiently seeded or influenced from the quantum vacuum or quantum decoherence is what increase the "consciousness potential" or potential to be influenced or possessed by a spirit.
Last edited:
Hurmanetar
New
When you make a line a loop you create inner and outer. Feed the output to the input and you create inner and outer. Increase the complexity and "surface area" (per post above) of the mechanism between input and output and you increase the intelligence and agency - potential for spirit influence.