This seems like a very limited, anthropogenic view of things. Because Prof. Conway can’t imagine life evolving on a planet much different than our own, and he can’t imagine intelligent life evolving out of a different niche as ourselves, he concludes that it is necessary for alien life to resemble ourselves.
How very Twilight-Zone like thinking.
Along these lines, has any scientists attempted running evolutionary algorithms to see if we can simulate approximations of spontaneous evolution of human-like life forms?
Along these lines, has any scientists attempted running evolutionary algorithms to see if we can simulate approximations of spontaneous evolution of human-like life forms?
Not exactly like that, or as complex, but I’ve been planning to do a small-scale simulator of evolution. It would achieve its computational shortcuts by having much simpler chemisty to work with: just enough so that you can have diverse types of reactions and a favored direction for them, and some type of self-replication is possible.
But the goal in my case is not to learn about extant species or alternate evolutionary paths, but rather, to explore the interplay between life, intelligence, thermodynamics, and complexity. Issues like: under what conditions can a self-replicator become more complex while continuing to replicate? What thermodynamic conditions permit systems to stay very far from equilibrium (i.e. become dissipative systems)? What rules must “intelligence” adhere to, and what costs must it pay?
Yes, though not in affiliation with any university or other group. And it’s as much for my understanding as to come up with something novel.
Be patient though: I announced my intentions to do this ~4 months ago, and still haven’t done anything on the implementation side. I’ve just been reading books and papers about those topics and gleaning insights on how they relate.
I think it would be impossible for humans, but possible for a super-intelligence that could intelligently take shortcuts and cut corners without skewing its results. Even then, I think it would take dyson-sphere levels of computer power.
Modern computers can’t even simulate one protein folding in real time as far as I know, and simulating a whole ecosystem is unimaginably harder than that.
This seems like a very limited, anthropogenic view of things. Because Prof. Conway can’t imagine life evolving on a planet much different than our own, and he can’t imagine intelligent life evolving out of a different niche as ourselves, he concludes that it is necessary for alien life to resemble ourselves.
How very Twilight-Zone like thinking.
Along these lines, has any scientists attempted running evolutionary algorithms to see if we can simulate approximations of spontaneous evolution of human-like life forms?
Not exactly like that, or as complex, but I’ve been planning to do a small-scale simulator of evolution. It would achieve its computational shortcuts by having much simpler chemisty to work with: just enough so that you can have diverse types of reactions and a favored direction for them, and some type of self-replication is possible.
But the goal in my case is not to learn about extant species or alternate evolutionary paths, but rather, to explore the interplay between life, intelligence, thermodynamics, and complexity. Issues like: under what conditions can a self-replicator become more complex while continuing to replicate? What thermodynamic conditions permit systems to stay very far from equilibrium (i.e. become dissipative systems)? What rules must “intelligence” adhere to, and what costs must it pay?
I’d be very interested in seeing the results of such an experiment. Is this intended to be for AI-related research, by the way?
Yes, though not in affiliation with any university or other group. And it’s as much for my understanding as to come up with something novel.
Be patient though: I announced my intentions to do this ~4 months ago, and still haven’t done anything on the implementation side. I’ve just been reading books and papers about those topics and gleaning insights on how they relate.
I’ve been involved in that kind of research: in short, we lack the computer power to do it at the moment.
Do you have an estimate of how much would be required?
I think it would be impossible for humans, but possible for a super-intelligence that could intelligently take shortcuts and cut corners without skewing its results. Even then, I think it would take dyson-sphere levels of computer power.
Modern computers can’t even simulate one protein folding in real time as far as I know, and simulating a whole ecosystem is unimaginably harder than that.