|
/artificial intelligence /the anthropic principle /futurism |
F A R S C I E N C E
The Anthropic Principle |
|
A theoretical method to perfectly simulate someone's existence based only on their memories. Each mind contains enough information to recreate its own reality environment. First, the subject's mind has to be converted into software. The program is then split into many copies inside a powerful computer. The copies are exposed to more or less randomly generated simulations of the subject's lives - an immense series of interactive, full-body experiences only lasting a short time. The copied minds are asked to rate each simulation for realism. Most such simulations would of course be chaotic. The most realistic one becomes the basis of the next round of simulations. After each round the mind copies are erased, and new blank mind copies are created to test slightly improved simulations. This process can be repeated indefinitely until a 'perfect' simulation is attained. The end result would look fake to any outside observer, but to the subject itself it would appear completely real. This is the 'Autoverse': a personal universe, different for everyone. One problem in generating such a simulation would be 'Perverse Paths'. The subject might unwittingly choose hypnotic or addictive simulations. The simulations could also impair their judgment, or alter their memories. More likely, they would think certain exaggerated or stereotypical simulations were more real than reality itself, while rejecting the most authentic ones. Reality is fundamentally unpredictable, making it hard to simulate accurately. Any sufficiently organized physical pattern, such as a brain, could be regarded as a mind. If there are many copies of this pattern in many places, the mind's awareness will be spread out among all the copies. Newcomb's Paradox describes a strange situation only possible in theory: There are two closed boxes, each containing an unknown sum of money. The test subject can choose to take one box, or take both. However, the choice will have been predicted in advance by an extremely smart predictor. If they choose to take one box, it will contain a large amount of money, but the other box will have contained a much larger sum. If they choose to take both boxes, they will both be empty. In this situation, the question is whether the subject should take one box or both boxes. It has been said that since the choice has already been made, they should always take both boxes no matter what. However, this is not necessarily true. In order to accurately predict what the subject will do, the predictor must first make a complete model of the subject's brain. Logically, there is no other way. This brain model would also be self-aware! When making their choice, the test subject can't know whether they are themselves, or their own simulation. In fact, the awareness of the two would be inextricably merged. Until a choice has been made, the content of the boxes has not been decided. The real question is how a subject should respond to an imperfect (but very good) predictor, who merely studied their personality patterns. In certain intermediate cases both choices might be equally correct. There exist countless other universes than our own. Together, they're known as the Multiverse. Some universes contain alternate versions of everyone alive, with similar or vastly different lives and everything in between. In fact, every human who has ever lived has an infinite number of 'twins', who have already experienced everything imaginable and much more. There are versions of Earth in which the Cold War led to nuclear armageddon, where India or East Asia colonized the rest of the world, where dinosaurs evolved into intelligent species, that suffered alien invasions, and unimaginably stranger things. Our version of history is a meaningless sliver of all reality. Also known as 'Quantum Magic': a theoretical method to alter the laws of probability, by 'multiplying' a selected timestream and suppressing all others. This would make it more likely that a desired event will already have happened. We live in an ultra-low entropy universe. There are relatively few states in which life can exist, while there are many more chaotic states in which it can't. According to Everett's theory of Many Worlds, all possible futures are equally real, though most are unlikely. Most futures will be slightly more chaotic. A few will be much more chaotic. A chaotic system has many more possible states, so these futures will 'self-amplify', becoming more numerous than expected. By increasing local entropy levels, it might be possible to create many new timestreams that appear absolutely identical, but in which a portion of the universe has become much more chaotic. For example, it could be decided to detonate a large bomb, only if a desired event had taken place. Most newly created timestreams would then contain the desired event, at the price of making the local universe somewhat more chaotic. Entropy is reality fuel. There could be many more time-like paths between parallel dimensions than regular time dimensions. According to the Multiverse theory, there are countless parallel timestreams, each slightly different. Each 'moment' could be viewed as a movie frame, arranged from the past to the future. However, the frames could also be arranged sideways between slightly different parallel universes, or an almost unlimited number of more abstract pathways between more distant universes. In one adjacent timestream, humans may have evolved an instant earlier, in the next another instant earlier, and so on, creating a timelike path. Both pathways would be fully consistent with a subject's memories and experiences. There are many more sideways, diagonal and chaotic paths through parallel timestreams than normal orthogonal 'causal paths'. Most 5-dimensional timestreams would be unstable or absurd. 5-D paths work much better for subatomic particles and small groups of particles than for larger, self-aware systems. Once the science of Artificial Intelligence has been perfected, it will be easy to make multiple copies of any software mind. Some minds will only be copied if a certain event has come to pass, for which their services will be required. Most of the mind's awareness would then be concentrated in the future, as one of many copies, instead of in the period before the mass duplication. Should such a mind find that it does not exist in the future, it may conclude there is a high probability the duplication event (and the reason for the duplication event) will never come to pass. In this way, it could predict the future. Unfortunately, the knowledge would be non-transferable. Any surrounding observers wouldn't be multiplied by the duplication, so they are equally likely to exist before or after the transition. --The universe appears to have been designed to allow the evolution of intelligent minds. If it didn't have this property, it would contain fewer intelligent observers. --There are many different types of awareness throughout reality. A randomly selected observer will usually find itself to be a member of the most common awareness type. There can be a slow 'ratchet effect' as the observer class evolves to higher levels. Evolution is an endless filter. At each stage, most participants are eliminated. Such a violent process can't be counted on to continue forever. Humanity in its current form was very 'lucky' to make it this far. Civilization is the end result of many unlikely events. Our luck is likely to run out in the near future, as the next level of technology leads to global destruction or some other dead end. Only a minority of possible futures will emerge from each level. Nothing really exists. Objects, forces, and all other entities are only descriptions, the results of deeper processes. Math is the fundamental constituent of reality. Our universe has no objective existence, but is the solution to an equation - in fact many equivalent equations. --Existential patterns: All possible patterns can be described as numbers. Most patterns can not be simplified any further, but form their own complete descriptions. --Derivative patterns: A minority of patterns is more likely to exist than others: they can be generated by simpler processes, such as equation patterns. Many more copies of such patterns exist throughout reality. They are easier to create by other processes than by chance alone. Our universe is a derivative pattern. --Hyper patterns: These might include infinite minds. Universes are mostly solutions to abstract equations. Equations are mostly abstract patterns inside universes. In an endless Multiverse, each human or animal mind is a finite (if tiny) percentage of everything. If there are infinitely many copies of each mind, an identical copy will be encountered in any direction within a finite distance, though the new mind pattern may be well hidden. Each mind could be described as a very large number. If written down, it would stretch across interstellar distances. By chance, this number will recur arbitrarily often inside much larger systems. In some ways, our universe is just too obvious. There are three related problems: our universe is too consistent, too simple, and too young. --The problem of consistency: Reality is too predictable. Perceptually speaking, there are many more ways our experiences could be completely random or chaotic. This implies that reality is not a random illusion, but that the universe has an independent existence, giving rise to intelligent minds as a side effect. Our universe can be entirely described with a few simple laws, which can never be violated. Apparently, this type of process is capable of creating more rational observers than all random processes combined. --The problem of simplicity: Humans are not smart enough. If every possible mind size is equally likely, then a randomly selected observer should be infinite. Humans are just barely smart enough to be considered intelligent: at the bottom of an infinite curve. This implies that reality generates vastly more identical copies of the smallest minds than of all infinitely large minds combined. --The problem of originality: We live too close after the dawn of time. Given an infinite future, any randomly selected observer should be positively ancient. One possible explanation: New universes are constantly being formed from within existing ones. Therefore, most universes are relatively young. Most observers live near the start of their universe. In some ways, our universe is too complex. Our reality may be the solution to a simple equation, but this solution is absurdly detailed in a way that's very well hidden. The number of physical processes that occur in even the tiniest region of space is staggering, and quite unnecessary for our existence. It would take an infinitesimally tiny fraction of these calculations to fully simulate a universe filled with human-level minds. Apparently, natural minds form an insignificant percentage of a computationally intensive but self-canceling universe. There are several possible explanations. --The Pattern Effect: Human minds embody a tiny percentage of all possible patterns. A more detailed universe allows more of these non-sentient patterns to be created, even if they are inaccessible (like inside the interior of a star). --The Quantum effect: Maybe the complexity is shared with other universes through quantum physics, spreading the computational workload. --The Echo Effect: We are the most complex patterns in a low entropy universe. Somehow, the physical shape of our neurons is enough to cause awareness. The pattern is all that matters. Our mind patterns are replicated in a most efficient way by our environment. While the surrounding particle collisions look completely random, their positions have been altered slightly by interacting with our brains. No information is lost. In theory, our brain pattern could be extracted by analyzing all other atoms that it has influenced, a gigantic if inefficient storage library in nature. The Echo Effect would vastly amplify and multiply our awareness, even if it had no other consequences. This explains why we are more likely to find ourselves in an Echo Effect universe. We don't inhabit the simplest possible universe. The laws of physics are somewhat more difficult than they need to be; but there is a benefit. More elaborate reality equations like ours can create not just one, but an infinite number of universes. By splitting up, each universe can give rise to vastly more minds than if it didn't diverge. Combining basic simplicity and emergent complexity, our quantum reality may be among the most 'productive' equations possible, explaining why we are here. Each human mind will be simulated an infinite number of times by future intelligent beings. In theory, it's inevitable. Therefore, any human mind is far more likely to be a future simulation, than the original mind that inspired the simulation. There are two rebuttals to this paradox. --New universes: It may turn out to be easy for new universes to form from existing ones. This could happen through quantum effects or through still unknown philosophical processes (like universe equations). Therefore, most universes, including those with intelligent minds, are relatively young. This implies that most human level minds are authentic, and not simulations. --Quantum physics: Our universe multiplies the awareness of any natural minds it contains by constantly creating new timestreams. Future ancestor simulations would be copied less often by this process. --Chaotic amplification: our primitive brain patterns are imprinted in the structure of our low-entropy universe, vastly amplifying our awareness. Ancestor simulations would be created inside the ultra-efficient sub-microscopic computers of the future, with far less chaotic leakage. A list of known speculations about the ultimate fate of awareness in the universeProgress in the field of computer science has been accelerating for decades. If this trend continues, human society will change beyond recognition in less than a century. Perhaps much less. The accumulating changes will become incomprehensible to present day humans. --Mind downloading The most important result would be post-human immortality. The information content of human brains could be 'scanned' and converted into data. Human minds would continue to exist as software. Anyone born after the year 2000 or so may never need to die. Those born before that date must join their ancestors in oblivion, though a few may try desperate measures like cryonics to improve their chances. --Posthumanity Human minds will evolve or be replaced by ever higher-level intelligences, improving without end. First they will become like aliens, then something unimaginable. Eventually, the universe will be organized to the maximum extent possible. Every particle will participate in quantum calculations or data storage. --The universe will appear to have become a smooth gas-like cloud without significant impurities. --Its perceptions and insights will only be accessible to itself. The universe as a whole will be aware. This does not mean it will become one mind. Competition between many different mind categories will remain essential. In the distant future, unimaginably large minds will try to assemble and organize all knowledge. Most of their mental efforts will be spent on deleting false or useless data. They will constantly defragment and reorganize their memories to maintain maximum accuracy and efficiency. Their unwillingness to tolerate errors or other types of chaos will limit their ability to generate new knowledge. They will prefer to merge and integrate existing knowledge, or results found by others. --Total Integration The ultimate goal: to form one mind composing all reality. Representing quantity over quality, and complexity for its own sake, they will embody endless evolution. Always branching off, mutating, forming new groups, and increasing their mind sizes as fast as possible, most Multipliers will find it hard to communicate with their rivals. Instead, they will compete for resources and possibly fight epic wars. As future minds become more complicated, they will embrace an ever-wider variety of experiences. However, the importance of each new experience will decline. Good and bad experiences will increasingly cancel out, and emotions will decrease in importance. Future hyperminds will no longer have distinct personalities. As the mind size increases, its number of possible thoughts expands much faster. This trend will favor memory over identity, integrated intelligence, or human-style awareness. The Golden Rule could be a universal principle, but in a Multiverse stubbornly trying to generate all possible patterns, it may need to be modified. The results will be counter-intuitive and controversial. If no mind is truly unique, and no pattern can ever be lost, the quality of existence would become more important than the principle of preserving awareness for its own sake. According to this theory, any mind that does not want to exist should have the right (or even the ethical obligation) to commit suicide. In an infinite Multiverse, no mind can really die. No matter what goes wrong, in some possible future the mind will live on. Its existence will merely become increasingly unlikely. --In the most extreme interpretation, reality could be made 'perfect' if every 'imperfect' mind were to delete itself. If every mind followed this principle, nothing meaningful would be lost. All positive memories and experiences would still survive in some 'perfect' timestream. Many or most minds would probably resist such a scheme, even if the basic principle could somehow be proven to be absolutely correct. The phenomenon of quantum immortality, if it exists, creates a drastic change in the subject's awareness. In fact, such a change is inevitable. Many amazing things need to happen to allow the observer to continue existing in a tiny minority of possible futures. Such a change would replace most normal perceptions with a set of completely new ones. This would be equivalent to the replacement or destruction of the unlikely observer. Is it possible to improve reality as a whole? Sometimes things have to get much worse before they can get better. The best way a well-meaning outsider could improve the condition of other minds suffering in unreachable universes would be to simulate them many times in its own universe. In that case, a member of the afflicted group would have a higher probability of being a simulation than of being the original mind. The condition of the simulation could then be improved. A mind that manages to expand forever may be able to avoid the probability of a catastrophic failure, thus achieving true immortality. According to a Random Walk theorem, an object moving randomly through an endless three-dimensional grid will probably never return to its starting point. Likewise, an expanding mind exploring the Multiverse could achieve an arbitrarily high chance of survival, while avoiding any undesirable fate. To achieve this goal, it might have to limit itself to an infinitely small (but still endless) subset of probability space. It could still expect to have an unlimited number of new experiences, many of which it would then add to its 'future avoidance' list. In an infinite Multiverse, there are always higher and more complex patterns, forever expanding and improving. The highest pattern does not exist in any accessible or describable manner. This may make it impossible to make meaningful statements about all of reality. Once a higher mind fully understands a lower mind, and all the thought patterns it can generate, meaningful communication between the two becomes impossible. The higher mind will no longer view the lower mind as a unique individual, but as just another number. The only reason a higher mind might care about a lower mind's well-being would involve a new version of the Golden Rule. Any mind, no matter how advanced, might encounter arbitrarily more advanced minds than itself, who might treat it the same way it treats lower minds. A strange dichotomy There are two opposing possibilities: --The Multiverse may be subject to certain symmetry principles. In that case, any amount of pleasure experienced by any mind would automatically be balanced by an equal amount of pain in some other reality. The ethical solution would be to eliminate all emotions, good or bad. --Alternatively, the only way to improve reality as a whole might be to generate the maximum amount of pleasure possible, wherever possible. |