Shane, again, the issue is not differentiation. The issue is classification. Obviously, tiny smiley faces are different from human smiling faces, but so is the smile of someone who had half their face burned off. Obviously a superintelligence knows that this is an unusual case, but that doesn’t say if it’s a positive or negative case.
Deep abstractions are important, yes, but there is no unique deep abstraction that classifies any given example. An apple is a red thing, a biological artifact shaped by evolution, and an economic resource in the human market.
Also, Hibbard spoke of using smiling faces to reinforce behaviors, so if a superintelligence would not confuse smiling faces and happiness, that works against that proposal—because it means that the superintelligence will go on focusing on smiling faces, not happiness.
Retired Urologist, one of the most important lessons that a rationalist learns is not to try to be clever. I don’t play nitwit games with my audience. If I say it, I mean it. If I have words to emit that I don’t necessarily mean, for the sake of provoking reactions, I put them into a dialogue, short story, or parable—I don’t say them in my own voice.
Shane, again, the issue is not differentiation. The issue is classification. Obviously, tiny smiley faces are different from human smiling faces, but so is the smile of someone who had half their face burned off. Obviously a superintelligence knows that this is an unusual case, but that doesn’t say if it’s a positive or negative case.
Deep abstractions are important, yes, but there is no unique deep abstraction that classifies any given example. An apple is a red thing, a biological artifact shaped by evolution, and an economic resource in the human market.
Also, Hibbard spoke of using smiling faces to reinforce behaviors, so if a superintelligence would not confuse smiling faces and happiness, that works against that proposal—because it means that the superintelligence will go on focusing on smiling faces, not happiness.
Retired Urologist, one of the most important lessons that a rationalist learns is not to try to be clever. I don’t play nitwit games with my audience. If I say it, I mean it. If I have words to emit that I don’t necessarily mean, for the sake of provoking reactions, I put them into a dialogue, short story, or parable—I don’t say them in my own voice.