Do you think that simply carrying out a computation can be immoral? I think no, and the difference between experimenting on a real person and a person in your simulation is that in the second case nothing would exist if you didn't spend your resources to run the simulation, (or in the original context spend resources to revive a discarded brain).
> Given that the possible negative outcome is as close to the biblical interpretation of hell as I can imagine, and one of the benefits of being an atheist is not having to worry about that
Interesting, i have never thought of this as a benefit, to me the hell seemed better because there you do not die, and logically have a chance to get out, because of love and forgiveness preached by some parts of the bible.
Yes. If something can experience pain, I think it's immoral to cause pain when it's unnecessary. If something is sentient, I think it's immoral to stop sentience without reason.
> Do you think that simply carrying out a computation can be immoral? I think no, and the difference between experimenting on a real person and a person in your simulation is that in the second case nothing would exist if you didn't spend your resources to run the simulation, (or in the original context spend resources to revive a discarded brain).
Yes. If you run a simulation that works on sentient AI's, then IMO mistreating them is immoral. I view them as life, and creating life comes with responsibilities.
Importantly, I think this view is important otherwise morality breaks down entirely. There's a chance we're all in a simulation right now. Does that mean nothing we do to each other can be immoral? What if we could prove we were in a simulation? That that free people form moral constraints?
> Interesting, i have never thought of this as a benefit, to me the hell seemed better because there you do not die, and logically have a chance to get out, because of love and forgiveness preached by some parts of the bible.
I'm not a bible scholar, but to my understanding, you go to hell for "eternity". Given the mystical nature of the whole supposition, I don't think it's useful to believe it existence is correct but the unescapability is not.
exactly, but it doesn't help if we say that you have to indefinitely keep computing the future states of the ai because you have started doing it, or that you are not allowed to add two numbers because that operation completes the calculation of suffering for the ai.
The only solution i can see is to base morality on private property: If i kill someone in the same simulation as me i take something that is not mine, and it is not ok. If i shut down the computer with a simulation, or start a simulation to see how people behave during genocide, i do not do any harm to anyone that exists or could exist independently from me, so it is ok.
And how does that work with children? It definitely follows some of the older thinking on the subject, but not any "modern" thinking on it. If I and my wife decide we don't want kids any more, are we free to end our children? I could claim that they can't exist indepdently of me (at least up to a certain age). It's also possible that an AI in a simulation might be able to provide enough benefit to someone to receive income, and enough to pay for its own resources. I think "could exist independent of me" is fairly ambiguous in this context.
Ultimately, this is all philosophical, and I don't think there's really a right or wrong answer, just people - walking talking examples of emergent phenomena - trying to find meaning. I actually don't think it's possible for a person for be completely consistent in behavior because of this, and at best we can hope to at least be consistent in how we perceive our actions. Thus, it doesn't surprise me that I find some of your views inconsistent, and that you likely find the same of me. We ultimately have different perceptions of what's being explained, and thus what seems entirely logical and consistent to one of us seems askew to the other. Exploring those views does have merit though, as it allows for a more comprehensive view of the subject.
In the case when child is truly incapable to exist without the mother, (before it is born), even most of the modern thinking allows to end it. But i think this will change when we have technology to grow people from cell to child in a lab.
With children it is still easy to argue that if parents do not want them they can't kill them because they are unique and the society always needs more of them. But ai is not unique, you can have an ai interacting with real world, providing for itself, and use an exact copy of that ai in an experiment (even the ai could experiment on a copy of itself!).
I think if a society decides to not allow any experiments on ai/brains in a jar, it would have hard time to enforce this, and if it succeeds enforcing the ban it will be out-competed by a neighbour that allows such experiments.
I'm not going to disagree with that. I argued the same earlier in this discussion in a more general sense.
> But ai is not unique, you can have an ai interacting with real world, providing for itself, and use an exact copy of that ai in an experiment (even the ai could experiment on a copy of itself!).
Which is an argument for duplicates, but 1) as a copy gets different inputs and diverges, it is unique, in the same way the you of this second is different slightly than the you before you started reading this sentence (and throw in Ship of Theseus for flavor). 2) even if the copies aren't viewed as diffierent, what about the last copy? If you've grown an AI, deleting the last copy is eliminating something something unique, if there was any possible input that cannot be replicated.
> I think if a society decides to not allow any experiments on ai/brains in a jar, it would have hard time to enforce this, and if it succeeds enforcing the ban it will be out-competed by a neighbor that allows such experiments.
The same could be said of any human experimentation that is banned. And that might be right. It's only been a relatively short time it's been in place in any enforceable way, in historical context.
That only follows if you think that about your life right now. If you think it's not a big deal right now, then it's not a big deal if you're a brain in a jar. If it is a big deal right now, then it is a big deal if you're a brain in a jar.
Pain is useful for living organisms as a variable to use in our fitness functions (minds) that allows us to avoid situations that decrease our survival rate. Ir may be mostly useless for a brain in a jar, but that doesn't necessarily mean it should be ignored.
One can make an argument that complete rationality should discard feelings and work only on whether something is useful or not. I think that argument is incorrect when taken to that extreme, as it looks at a shorter horizon than the evolutionary scale which has produced us, to good effect. I would argue that the social aspect of our species is what has allowed us to not only thrice, but completely dominate the environment we are in. To ignore our empathy is to ignore what has brought us to this point.
We can say that a brain in a jar in pain should not be catered to because it has nothing to provide (which also might be incorrect), but by that reasoning we should also say that the person in hospice with weeks or months left to life with no relatives should not be catered to. Should we cause pain to or end the existence of each of those without any regard to their own feelings? We could just as easily use the person in hospice for testing. Amazing scientific discoveries could be unearthed.
But if we agree that emotion has played a positive role in the development of the species, then "hellish" things and the avoidance of them is not pointless computation.
> But until that the choices we have here are nothingness and an experiment with unknown outcome. Given this choice i don't see any experiment that is worse than the nothingness.
Given the known outcome of coming back after death for an existence of even known time, say one week, where that time comprised of pain and suffering exclusively, I would opt to not have that time, thus opting for nothingness. Given an outcome I would not choose, and an outcome I would choose, it them becomes an equation of what is the likelihood of each, and how much to I want or want to avoid each.
Given that the possible negative outcome is as close to the biblical interpretation of hell as I can imagine, and one of the benefits of being an atheist is not having to worry about that, I'm not in a hurry to opt into something that then provides that possibility. That's what I was getting at before.