3 ms·

What dream characters reveal about the astonishing dreaming brain (2021)

1 points by Gooblebrai 2y ago

DemocracyFTW2 2y ago
An attempt to put the ideas I've gained from reading the article into words: The brain cannot know, by just 'looking at itself' (i.e. doing static analysis) what its own reactions would be in every conceivable situation—it has to actually experience such situations and observe the outcomes. This inability could be down to the brain's fundamental organic inability to look at individual synapses and the network that neurons have formed; it could also be that such networks could in theory be completely understood but are, in practice, too complicated such that the brain's computational power is insufficient to do the necessary calculations. In fact, both seem to be the case and would suffice to explain the necessity for running simulations. However there's a third component, and this is the truly interesting one: the Halting Problem which tells us that a sufficiently complex network of rules (such as we deal with when writing and debugging programs), although perfectly deterministic, cannot be, in the general case, completely predicted except by running the network / the program, for example by running a test case, a simulation. At least for the conscious and almost-conscious parts of us, building models of reality complete with orientational factoids and life-like actors and so on is an essential part of what enables us to navigate reality. The brain, then, having evolved the ability to form simulacra that behave in plausibly realistic ways, apparently also evolved the ability to form simulacra that are rather temporary and isolated from others, in essence forming small 'snow-globe worlds' that can at opportune times (during rest) get shaken and prodded at, scenes formed from imaginary play-doh, malleable and flexible. That is the brain at play if you will, trying out new things, getting better at old things, and generally spending time with itself.

1 points by Gooblebrai 2y ago

An attempt to put the ideas I've gained from reading the article into words: The brain cannot know, by just 'looking at itself' (i.e. doing static analysis) what its own reactions would be in every conceivable situation—it has to actually experience such situations and observe the outcomes. This inability could be down to the brain's fundamental organic inability to look at individual synapses and the network that neurons have formed; it could also be that such networks could in theory be completely understood but are, in practice, too complicated such that the brain's computational power is insufficient to do the necessary calculations. In fact, both seem to be the case and would suffice to explain the necessity for running simulations. However there's a third component, and this is the truly interesting one: the Halting Problem which tells us that a sufficiently complex network of rules (such as we deal with when writing and debugging programs), although perfectly deterministic, cannot be, in the general case, completely predicted except by running the network / the program, for example by running a test case, a simulation. At least for the conscious and almost-conscious parts of us, building models of reality complete with orientational factoids and life-like actors and so on is an essential part of what enables us to navigate reality. The brain, then, having evolved the ability to form simulacra that behave in plausibly realistic ways, apparently also evolved the ability to form simulacra that are rather temporary and isolated from others, in essence forming small 'snow-globe worlds' that can at opportune times (during rest) get shaken and prodded at, scenes formed from imaginary play-doh, malleable and flexible. That is the brain at play if you will, trying out new things, getting better at old things, and generally spending time with itself.