4 ms·

If it's possible for the Earth to have "one" consciousness while containing both the Cheetah and the Gazelle, the universe can reconcile us and a paperclip maxi

If it's possible for the Earth to have "one" consciousness while containing both the Cheetah and the Gazelle, the universe can reconcile us and a paperclip maximizer.

derefr 9y ago
If you assume that the Earth has a consciousness, then individual beings on its surface are more like cells in a body than like separate minds. The minds of the individual components aren't "part of" that consciousness, in the way that mental schema are a part of a brain; they're more like neurons, or other nerve cells elsewhere in the body, literal parts of the brain. You wouldn't derive your idea of what a brain is thinking from what an individual nerve cell "wants" (cells "want" ATP, mostly); you'd only derive it from what that cell does in response to stimuli from other cells, and how that maps to thinking at a larger level. A human being wants things that often involve the destruction of its various cells. We want to drink alcohol sometimes, for example. This is possible because our preference-function has no factor for the preferences of the individual cells killed in the process. They aren't a semantic part of our mind. They're just a literal part of our bodies. A hypothetical Earth-Mind wouldn't have a preference-function derived via coherent extrapolated volition of all Earth-life. Its mind would simply be the set of preferences the Earth seems to have if you're an alien trying to interact with "the Earth" as a coherent being, on some scale where the actions of individual cells (e.g. humans) don't matter as much as the actions of the biosphere as a whole. Maybe with action-and-reaction observed over a time-scale of billions of years, for example. That preference function might be observed to be, for example, "install replicating carbon-based life onto anything that comes into contact with it." Or "capture solar radiation into ongoing surface chemistry." Note that these are coincidentally things that some of the individual cells also want, but again, the cells in existence at any given moment and their wants aren't anything you can use to deduce the preferences of the mind. It's more what cells are kept in existence and given more resources over time—especially in response to specific stimuli—that allow you to make educated guesses about what a multicellular agent is "trying" to do. (With the proviso that those cells shouldn't also be clearly fighting the rest of the organism in doing so; a model that predicts that animals want to get cancer is a bad model.) Presuming humans terraform the planet completely, the preference-function of the Earth as a whole might seem to end up as "transform all raw materials into structures, items, and cells suitable as inputs to sustain the life of the 'human' cell-line." This is a higher-level thought-process that could be found, but it wouldn't erase the lower level; the Earth-Mind would still look to have those other desires, because humans are still carbon-based life and still do carbon-based life things. Just like our own brain cells, despite being the "dominant cell-line" of our bodies, still function as cells in a body. Humans are "embodied" by our biosphere just like our brain cells are "embodied" by other cell-lines. And so it still makes sense to talk to a human by talking to its body, rather than directly to its brain—and it still makes sense to talk to "the [terraformed] Earth" rather than talking to humans. On the other hand, the "observed mind" of a planet with a paperclip-maximizer on it wouldn't look anything like a natural-biosphere or human-driven-biosphere. It would just look like the single paperclip-maximizer's mind. A paperclip-mamixizer isn't part of a biosphere; it's a monocellular culture, the preferences of the whole identical to the preferences of the individual entities. And those preferences are completely orthogonal to the preferences of a biosphere. If Earth came into contact with a paperclip-maximizer, their interaction wouldn't look like a coherent mind at work; it would look like two minds each blindly attempting to achieve their own goals with no greater purpose in their collision. The point I hold out on is that, if you interact with the Universe as a whole, on a scale where things like biospheres are the "cells", and attempt to ascertain its preference-function, you won't get anything sensible out. Unlike Earth, where both the Cheetah and the Gazelle "serve" to tree-search a set of biological machines that started with Earth's first carbon-based replicators, an interaction between the Earth-Mind and the Paperclipper-Mind doesn't serve to "do" anything. The behaviour of the interaction wouldn't throw off byproducts that would be of use to anyone trying to predict anything. The Universe, as an experiment containing multiple orthogonally-preferenced agents, is purposeless; unlike Earth itself, there is nothing from which an alien standing outside of it could infer the hand of an intelligent designer—a mind at the Universe's scale that found it useful to set its "cells" in motion in the way the Universe has been set in motion. Which makes sense, I think; the "intelligent designer" of the Earth-Mind was the first carbon-based replicator, but there was no equivalent process that "took over" the Universe as a whole to infuse it with a self-similar cluster of optimization-processes.

If it's possible for the Earth to have "one" consciousness while containing both the Cheetah and the Gazelle, the universe can reconcile us and a paperclip maximizer.

derefr 9y ago

If you assume that the Earth has a consciousness, then individual beings on its surface are more like cells in a body than like separate minds. The minds of the individual components aren't "part of" that consciousness, in the way that mental schema are a part of a brain; they're more like neurons, or other nerve cells elsewhere in the body, literal parts of the brain. You wouldn't derive your idea of what a brain is thinking from what an individual nerve cell "wants" (cells "want" ATP, mostly); you'd only derive it from what that cell does in response to stimuli from other cells, and how that maps to thinking at a larger level. A human being wants things that often involve the destruction of its various cells. We want to drink alcohol sometimes, for example. This is possible because our preference-function has no factor for the preferences of the individual cells killed in the process. They aren't a semantic part of our mind. They're just a literal part of our bodies. A hypothetical Earth-Mind wouldn't have a preference-function derived via coherent extrapolated volition of all Earth-life. Its mind would simply be the set of preferences the Earth seems to have if you're an alien trying to interact with "the Earth" as a coherent being, on some scale where the actions of individual cells (e.g. humans) don't matter as much as the actions of the biosphere as a whole. Maybe with action-and-reaction observed over a time-scale of billions of years, for example. That preference function might be observed to be, for example, "install replicating carbon-based life onto anything that comes into contact with it." Or "capture solar radiation into ongoing surface chemistry." Note that these are coincidentally things that some of the individual cells also want, but again, the cells in existence at any given moment and their wants aren't anything you can use to deduce the preferences of the mind. It's more what cells are kept in existence and given more resources over time—especially in response to specific stimuli—that allow you to make educated guesses about what a multicellular agent is "trying" to do. (With the proviso that those cells shouldn't also be clearly fighting the rest of the organism in doing so; a model that predicts that animals want to get cancer is a bad model.) Presuming humans terraform the planet completely, the preference-function of the Earth as a whole might seem to end up as "transform all raw materials into structures, items, and cells suitable as inputs to sustain the life of the 'human' cell-line." This is a higher-level thought-process that could be found, but it wouldn't erase the lower level; the Earth-Mind would still look to have those other desires, because humans are still carbon-based life and still do carbon-based life things. Just like our own brain cells, despite being the "dominant cell-line" of our bodies, still function as cells in a body. Humans are "embodied" by our biosphere just like our brain cells are "embodied" by other cell-lines. And so it still makes sense to talk to a human by talking to its body, rather than directly to its brain—and it still makes sense to talk to "the [terraformed] Earth" rather than talking to humans. On the other hand, the "observed mind" of a planet with a paperclip-maximizer on it wouldn't look anything like a natural-biosphere or human-driven-biosphere. It would just look like the single paperclip-maximizer's mind. A paperclip-mamixizer isn't part of a biosphere; it's a monocellular culture, the preferences of the whole identical to the preferences of the individual entities. And those preferences are completely orthogonal to the preferences of a biosphere. If Earth came into contact with a paperclip-maximizer, their interaction wouldn't look like a coherent mind at work; it would look like two minds each blindly attempting to achieve their own goals with no greater purpose in their collision. The point I hold out on is that, if you interact with the Universe as a whole, on a scale where things like biospheres are the "cells", and attempt to ascertain its preference-function, you won't get anything sensible out. Unlike Earth, where both the Cheetah and the Gazelle "serve" to tree-search a set of biological machines that started with Earth's first carbon-based replicators, an interaction between the Earth-Mind and the Paperclipper-Mind doesn't serve to "do" anything. The behaviour of the interaction wouldn't throw off byproducts that would be of use to anyone trying to predict anything. The Universe, as an experiment containing multiple orthogonally-preferenced agents, is purposeless; unlike Earth itself, there is nothing from which an alien standing outside of it could infer the hand of an intelligent designer—a mind at the Universe's scale that found it useful to set its "cells" in motion in the way the Universe has been set in motion. Which makes sense, I think; the "intelligent designer" of the Earth-Mind was the first carbon-based replicator, but there was no equivalent process that "took over" the Universe as a whole to infuse it with a self-similar cluster of optimization-processes.