3 ms·
We differentiate treatment by intelligence of species for a good reason No we don't. We keep dogs, we* kill pigs and eat them. *) 'We' meaning a generalised '
by interfixus 10y ago
We differentiate treatment by intelligence of species for a good reason
No we don't. We keep dogs, we* kill pigs and eat them.
*) 'We' meaning a generalised 'you'. I don't.
- nojvek 10y agoThere's a great book called Homodeus they goes in depth about this for a chapter Basically it's come down to humans deciding the fate of so many animals. We keep our cows and pigs in very narrow places and just enough systems for their survival but no regard for their emotional needs. Because we are the dominant spicies we act according to our selfish interests. A smarter intelligent spices will probably do the same. If we somehow end up being useful for them, our fate might be the same as chickens, pigs and cows.
- literallycancer 10y agoYes, of course. I have extremely strong preference for a world where other humans will cause suffering and/or death of non-human life, if necessary, to cure me of a disease or improve my quality of life. Therefore, I support such measures to help any other humans, hoping that they would do the same for me. It's self preservation, plain and simple. Non-human life is useful and worthwhile only insofar as it improves human life. As it happens, that includes, some would say luckily, trying to preserve a whole lot of lifeforms because we don't yet know whether they will be needed in the future. I can't find any reason to consider this immoral or wrong in any way. I think it's beyond morality - merely an inevitable side-effect of existence.
- interfixus 10y agoOut of honest curiosity: What is your take on human life, then? A sittuation, let's say, where I cause you suffering in order to improve the quality of my life? Or vice versa?
- literallycancer 10y agoI consider my utility roughly equal to that of close friends and family and will try to trade it favorably, such as spending 30 minutes to save them an hour. This is applicable to suffering as well. With strangers, the expected return would have to be better, probably by a factor of ten or so. One of the problems with this approach is that estimating expected utility gains (especially other people's) is hard. And of course you get the weird edge cases with people (think serial killers) deriving pleasure from hurting other people, so you have to use some kind of objective measure that applies at least to most of the population, rather than just trusting the other person that "it really does it for them".