"Ironically, if I were trying to create an AI system that could see, and reason about, the world more effectively (and who says I’m not!) I would be trying to find ways to inject more of this kind of embodied, valence-laden affect into its perceptual apparatus." I think this is exactly what people do when they assign a persona to a chatbot (e.g., "you are an expert at Jungian dream analysis. Last night I dreamt that..."). The persona seems to more effectively "constrain" the statistical space of token prediction, often improving output quality. It's possible that we're all socially encouraged to do this, take on expertise that implicitly narrows our affect in order to improve the quality of our labor output. The humanities sort of acknowledge this by (ideally) broadening our conceptual space and supposedly enriching the experience of life (before eventually being pushed into expertise). To that end, seems like market forces might be a drag for humans and for AI.
"Why can’t we see this, too, as nature sorting itself into delicate, complex, semi-stable equilibria? [...] You would be too excited to try out this superpower on museums and libraries, on kindergarten concerts and family reunions and busy street corners and starry skies and lecture halls and crowded bazaars and lonely ponds!"
I try not to be that guy, but damn it: meditation, you're talking about meditation! 😭
"The Eye of Man, a little narrow orb, clos’d up & dark,
Scarcely beholding the Great Light, conversing with the ground:
The Ear, a little shell, in small volutions shutting out
True Harmonies & comprehending great as very small"
"Does it have to taste bad to you now? Wouldn’t your life be strictly better if it didn’t?"
I've been thinking about this 'sommelier paradox' of yours. Instinctively I think it's false, maybe because I'm a snob, but also because I believe having true knowledge must always be preferable to not. So thinking it through, I think the mistake might be in only taking into consideration the hypothetical marginal utility of consumption in that moment- as in, there are more bad wines on offer than good, so we'd be better off if we liked bad wines. But that ignores how we acquired that knowledge in the first place.
Like, my freshman year of college, I devoured Cryptonomicon and the Baroque Cycle by Neal Stephenson, thought they were simply the greatest, and continued to think so for years. Then I decided to reread it a decade or so later. I couldn't even finish it! I couldn't believe I'd be saying it was favorite book! What happened in the interim is I read other books, further developed my own tastes, and grew up. Marginally, it might be better to be able to appreciate mediocre writers, there's certainly many more of them out there, and they are often more prolific too. But even though I can't appreciate Stephenson the way I could as a teenager, I'm not worse off, because that lack of appreciation is itself a byproduct of the pleasure gotten from lots of reading. Similarly, the only way to acquire an appreciation of wine is through study, to learn all those complex subtleties, and that requires drinking and appreciating a lot of wine. The cumulative benefit of that true knowledge is positive, even if it means the marginal benefit of the next bottle of Barefoot Cabernet is now negative. So I tell myself, at any rate.
"Ironically, if I were trying to create an AI system that could see, and reason about, the world more effectively (and who says I’m not!) I would be trying to find ways to inject more of this kind of embodied, valence-laden affect into its perceptual apparatus." I think this is exactly what people do when they assign a persona to a chatbot (e.g., "you are an expert at Jungian dream analysis. Last night I dreamt that..."). The persona seems to more effectively "constrain" the statistical space of token prediction, often improving output quality. It's possible that we're all socially encouraged to do this, take on expertise that implicitly narrows our affect in order to improve the quality of our labor output. The humanities sort of acknowledge this by (ideally) broadening our conceptual space and supposedly enriching the experience of life (before eventually being pushed into expertise). To that end, seems like market forces might be a drag for humans and for AI.
"Why can’t we see this, too, as nature sorting itself into delicate, complex, semi-stable equilibria? [...] You would be too excited to try out this superpower on museums and libraries, on kindergarten concerts and family reunions and busy street corners and starry skies and lecture halls and crowded bazaars and lonely ponds!"
I try not to be that guy, but damn it: meditation, you're talking about meditation! 😭
"The Eye of Man, a little narrow orb, clos’d up & dark,
Scarcely beholding the Great Light, conversing with the ground:
The Ear, a little shell, in small volutions shutting out
True Harmonies & comprehending great as very small"
"Does it have to taste bad to you now? Wouldn’t your life be strictly better if it didn’t?"
I've been thinking about this 'sommelier paradox' of yours. Instinctively I think it's false, maybe because I'm a snob, but also because I believe having true knowledge must always be preferable to not. So thinking it through, I think the mistake might be in only taking into consideration the hypothetical marginal utility of consumption in that moment- as in, there are more bad wines on offer than good, so we'd be better off if we liked bad wines. But that ignores how we acquired that knowledge in the first place.
Like, my freshman year of college, I devoured Cryptonomicon and the Baroque Cycle by Neal Stephenson, thought they were simply the greatest, and continued to think so for years. Then I decided to reread it a decade or so later. I couldn't even finish it! I couldn't believe I'd be saying it was favorite book! What happened in the interim is I read other books, further developed my own tastes, and grew up. Marginally, it might be better to be able to appreciate mediocre writers, there's certainly many more of them out there, and they are often more prolific too. But even though I can't appreciate Stephenson the way I could as a teenager, I'm not worse off, because that lack of appreciation is itself a byproduct of the pleasure gotten from lots of reading. Similarly, the only way to acquire an appreciation of wine is through study, to learn all those complex subtleties, and that requires drinking and appreciating a lot of wine. The cumulative benefit of that true knowledge is positive, even if it means the marginal benefit of the next bottle of Barefoot Cabernet is now negative. So I tell myself, at any rate.