Would an average year in the life of an em in Hanson’s Malthusian explosion scenario really be >0 QALY? Hanson has kinda defended this scenario because the ems would want to be alive but I don’t think that means anything. I remember reading about mice and painful wireheading (probably Yvain’s post) and how you can make mice want that kind of wireheading even though it’s painful. Similarly it’s easy to imagine how people would want to live painful and miserable lives.
Yeah, I read that, reconsidered my impression and it seems you are right. My memories about his opinion seemed to have become muddled and simplified from several sources like his Uploads essay where he says “Most uploads should quickly come to value life even when life is hard or short, and wages should fall dramatically.” (which doesn’t seem to be a value statement) that poor folks essay, this discussion here (in which he doesn’t commentate) and this video interview in which he constantly says that life will be okay even though we’ll become more and more alienated from nature.
But I don’t think my view of his opinion was 100% incorrect. The distinction between “valuing your life” and “wanting to live” is interesting. If you want to live, does that automatically mean that you value your life? I mean, I’ve had days when maybe 95% of the time I’ve felt miserable, and 5% of the time I’ve felt okay and in the end I’ve still considered those days okay. If I want to have more of those kind of days, does that mean I value misery? How do you assess the quality of life in these kind of cases, and in cases where the ‘misery’ is even more extreme?
In your first paragraph, you agree with me that it isn’t a value judgement, but then in your second paragraph, you go back to claiming that it is the foundation of his position. I think it is mainly a response to claims that uploads will be miserable. I think his position is that we should not care about whether the uploads value their lives, but whether we, today, value their lives; but he thinks that moral rhetoric does not well match the speaker’s values. cf
I would guess yes—but that might change depending on details. At the very least, if we decided on some way to measure QALYs (our current methodology is real simple!), and then tried to maximize that measurement, we’d at best get something that looked like pared-down ems.
Ultimately, how you choose between futures is up to you. Even if something has an objective-sounding name like “quality-adjusted life years,” this doesn’t mean that it’s the right thing to maximize.
Yes, wanting to live isn’t perfect evidence of a life worth living. But it sure looks like it provides some bayesian evidence.
Looking at whether the ems want more copies of themselves and want faster clock speeds should provide stronger evidence, and it seems unlikely that ems who don’t want either of those will be common.
Ems should have some ability to alter themselves to enjoy life more. Wouldn’t they use that?
If it provides bayesian evidence, shouldn’t there be something that would in principle provide counterevidence? I can’t figure out what that kind of counterevidence would be. Can you imagine an em population explosion where at some point no ems would want to make copies of themselves? I’ve got the impression that once an em population explosion gets started you can’t really stop that because those ems that want copies get selected no matter how miserable the situation.
Ems should have some ability to alter themselves to enjoy life more. Wouldn’t they use that?
Since in this scenario almost all ems work on a subsistence level and there’s a huge number of ems, if enjoying life makes them even slightly less productive I don’t think that kind of alteration would become very common due to selection effects.
As I understand it, the premise behind ems is that it’s possible to copy human minds into computers, but not to understand the spaghetti code. There won’t be an obvious way to just make workers happier.
Would an average year in the life of an em in Hanson’s Malthusian explosion scenario really be >0 QALY? Hanson has kinda defended this scenario because the ems would want to be alive but I don’t think that means anything. I remember reading about mice and painful wireheading (probably Yvain’s post) and how you can make mice want that kind of wireheading even though it’s painful. Similarly it’s easy to imagine how people would want to live painful and miserable lives.
Has he? I think his more typical defense is Poor Folks Do Smile.
Yeah, I read that, reconsidered my impression and it seems you are right. My memories about his opinion seemed to have become muddled and simplified from several sources like his Uploads essay where he says “Most uploads should quickly come to value life even when life is hard or short, and wages should fall dramatically.” (which doesn’t seem to be a value statement) that poor folks essay, this discussion here (in which he doesn’t commentate) and this video interview in which he constantly says that life will be okay even though we’ll become more and more alienated from nature.
But I don’t think my view of his opinion was 100% incorrect. The distinction between “valuing your life” and “wanting to live” is interesting. If you want to live, does that automatically mean that you value your life? I mean, I’ve had days when maybe 95% of the time I’ve felt miserable, and 5% of the time I’ve felt okay and in the end I’ve still considered those days okay. If I want to have more of those kind of days, does that mean I value misery? How do you assess the quality of life in these kind of cases, and in cases where the ‘misery’ is even more extreme?
In your first paragraph, you agree with me that it isn’t a value judgement, but then in your second paragraph, you go back to claiming that it is the foundation of his position. I think it is mainly a response to claims that uploads will be miserable. I think his position is that we should not care about whether the uploads value their lives, but whether we, today, value their lives; but he thinks that moral rhetoric does not well match the speaker’s values. cf
I would guess yes—but that might change depending on details. At the very least, if we decided on some way to measure QALYs (our current methodology is real simple!), and then tried to maximize that measurement, we’d at best get something that looked like pared-down ems.
Ultimately, how you choose between futures is up to you. Even if something has an objective-sounding name like “quality-adjusted life years,” this doesn’t mean that it’s the right thing to maximize.
Yes, wanting to live isn’t perfect evidence of a life worth living. But it sure looks like it provides some bayesian evidence.
Looking at whether the ems want more copies of themselves and want faster clock speeds should provide stronger evidence, and it seems unlikely that ems who don’t want either of those will be common.
Ems should have some ability to alter themselves to enjoy life more. Wouldn’t they use that?
If it provides bayesian evidence, shouldn’t there be something that would in principle provide counterevidence? I can’t figure out what that kind of counterevidence would be. Can you imagine an em population explosion where at some point no ems would want to make copies of themselves? I’ve got the impression that once an em population explosion gets started you can’t really stop that because those ems that want copies get selected no matter how miserable the situation.
Since in this scenario almost all ems work on a subsistence level and there’s a huge number of ems, if enjoying life makes them even slightly less productive I don’t think that kind of alteration would become very common due to selection effects.
Evidence that most ems are slaves whose copies are made at the choice of owners would seem relevant.
Making miserable workers a bit happier doesn’t seem to make them less productive today. Why should there be no similar options in an em world?
As I understand it, the premise behind ems is that it’s possible to copy human minds into computers, but not to understand the spaghetti code. There won’t be an obvious way to just make workers happier.
I expect faster and more reliable evaluations of Prozac-like interventions.
I also expect that emotions associated with having few cpu cycles are less strongly ingrained than those caused by lack of food.