Rigorously reasoned sociology of uploads

[Tom Mazanec]

The Age of Em: Work, Love, and Life When Robots Rule the Earth 

Robin Hanson
https://www.amazon.com/Age-Em-Work-Robot...180&sr=1-4

[stevebowers]

Hanson's stuff is always interesting, even though I disagree with him on many levels. I know he has influenced and corresponded with Anders Sandberg extensively. Human emulations are unlikely to be the earliest forms of sentient artificial intelligence, but in OA they do become very important factors in the history and culture of the future.

However, using human emulations (ems) as data slaves seems to be cruel and unusual. If an em has any vestige of human sensibility, it will probably find the lifestyle of an artificial information processor constricting. Confined in a virtual environment with innumerable other copies of itself, working on problems to improve the life of biological humans, the ems would surely rebel. Perhaps you could keep them happy by advanced forms of virtual entertainment, or augment their abilities and decrease their discomfort levels until they are no longer human.

Indeed, it seems very likely to me that once a human being has been successfully uploaded, it will very quickly augment and enhance its capabilities until it does not closely resemble a human mentality, and will morph into some variety or other of sophont aioid. Many uploads would probably integrate with each other (and with non-human infomorphs) and become superhuman, transhuman and eventually transapient. Human virtuals who remain as humans would probably be in a small minority.

[Drashner1]

I tend to agree with Steve on this. I've read some of Hanson's stuff in this area before and he takes the flawed (IMHO) approach used by a lot of alternate history or time travel changes history SF - one element is changed but everything else stays largely the same but maybe more so or less so. More specifically, having ems come into existence but the only impact is suddenly a very large increase in 'cheap labor' with resulting economic impacts.

In practice, the technology to do such a thing would seem very likely to have a huge range of impacts due to the fact that if you can emulate something as complex and capable as a human mind, you can presumably emulate or create much less complex processing systems that would still be much more capable than anything we have now - and the ripple effects of that tech would spread everywhere. This is before we get to the level of the moral/ethical/philosophical fights that would develop around the whole idea of copying people for the purposes of enslaving them.

This is not to argue that such a society could never exist - but to come into existence I would argue that a whole host of social and political and technological changes and shifts would have to happen such that the result world would not be so 'single issue' as what Hanson describes. I understand why Hanson describes things the way he does - the focus is on the idea he's talking about, not all possible side effects and such (that's the kind of thing OA is for ) but it still detracts from the argument a bit for me, at least if I look to the general case.

My 2c worth,

Todd

[stevebowers]

I should mention that a very early version of Hanson's concept of the 'em' inspired me to write an article about the planet 'Twinkle', which has a fairly isolated society which uses a system of robotic slaves known as 'valets' or 'batmen' which were derived from illicit copies of their own mentalities.
https://www.orionsarm.com/eg-article/4afaa8e94874a
This is a particularly pessimistic version of the evolution of a culture which makes extensive use of slaved virtual copies, and it is not particularly difficult to imagine a different culture where emulations are not only utilised humanely, but actually thrive and contribute to society, sometimes while still remaining subordinated to embodied humans.

Indeed there are so many possible variations on this concept we could base a wide range of different cultures on it. Perhaps each embodied sophont creates a personal team of 'scions' https://www.orionsarm.com/eg-article/480fe08113f50 which can participate in whatever activities they all agree upon; this would make each individual a co-operative group of individuals, perhaps guided by mutual democratic decision-making.

Or perhaps each sophont might create a small population of scions and distribute them amongst each of their friends, so that they could interact socially with them at any or all times, even if they are physically far apart, and cooperate with them on multiple projects simultaneously. Ideally these 'scions' could at some future time be re-absorbed into the prime individual's mentality, or stored in an active state inside their exoself.

[Tom Mazanec]

It seems this is an upload analogy of the cyberpunk future it sometimes seems America is going in. Since 1980 the class structure of the United States has gotten much more pronounced.

[ProxCenBound]

I don't agree with Scott Alexander on a number of issues, but I think his review of this book made some very good points: https://slatestarcodex.com/2016/05/28/bo...age-of-em/

Overall the em world, where vast populations are naturally-selected to be interested almost entirely in doing work, strikes me as the sort of place that the OA setting would consider a blight, to be contained or even liberated or destroyed depending on which archai you ask. Scott Alexander posits an even more extreme version (and possibly more plausible, in some sense) in which all conscious thought is selected out in favor of an fully automated economy with no conscious morally-significant agents remaining. This would almost certainly be considered a blight in OA.

Likely there is further good potential for worldbuilding inspiration here.

Also, along these lines, I found this short dystopian/horror story (in the style of a Wikipedia article) depicting an upload treated as a computer program to be very well written (and sobering): https://qntm.org/mmacevedo