I think we will be torn by a desire for advancement (and the fear of it) and a longing to return to simpler times. Think of the small house movement. Think of people wanting to live off grid.
Plus I doubt we will make it to 2100 without some major worldwide calamity.
Wesley Clark,
Stuff I don’t understand:
What will motivate these changes? Political altruism? General economic benefit? Natural progression toward some goal? DARPA?
Is there a social or evolutionary advantage to living longer?
How will the general population obtain income in a society that is not based on labor?
How will our limited intelligence be able to recognize the greater wisdom of computers?
Why would an earthly economy support interstellar travel? Due to time dilation the ship could not return in the lifetime of the investing capitalists. The faster you go the worse it gets.
Stuff like that.
Crane
[QUOTE=Crane]
What will motivate these changes? Political altruism? General economic benefit? Natural progression toward some goal? DARPA?
[/QUOTE]
All of the above. Basically, it’s natural technological and societal progression stuff.
There are both advantages and disadvantages, as with all things. Just like there are advantages to society and humanity of people living longer than the 30 average life expectancy in the past.
Probably by changing the definition of what ‘labor’ is, would be my WAG. Today, ‘labor’ could be playing a game on Twitch while talking about it.
We already recognize the advantages they give us. Myself, I think what will actually happen is more a fusion of human and computer, where humans use computers more tightly integrated to do the stuff they are really good at and help the humans do the stuff we are really good at.
You don’t have to ship everything here for it to work. The capitalists don’t have to stay here either. There are vast, almost infinite resources outside of this one planet. Also, all our eggs in one basket is probably not going to be optimal for the continuation of the species into the future.
XT,
Beyond simple repetitive tasks, computers are only good at searching and taxonomy. We like them because they are tireless and predictable.
Expert systems are not intelligent. They are not more than elaborate look up tables. True AI systems have to be taught. Like humans they do not know anything they have not been taught. They can, as can humans, make inferences based on their experience. The inferences made by humans vary depending on their education. Militarists, politicians, theologians and scientists make different inferences based on their individual backgrounds. Human societies use peer review or broad public opinion to validate their inferences.
So, how would we educate a computer to make decisions/inferences that are altruistic and absolutely correct? Who, or what, would review them? How could they evaluate the results of an intelligence that is greater than their own.
Crane
[QUOTE=Crane]
Beyond simple repetitive tasks, computers are only good at searching and taxonomy. We like them because they are tireless and predictable.
[/QUOTE]
But they are VERY good at those things, so much so that they are tremendous labor savers. I see that becoming more prevalent in the future.
No, this is incorrect. Well, they aren’t ‘intelligent’ in that they aren’t strong AI if that’s what you mean. But they are learning systems that become better and better as they are used more. Calling them simply ‘elaborate lookup tables’ is incorrect.
So do expert systems today in most cases.
Not sure where you are going with this.
Why would we necessarily want them to be altruistic? How do we educate humans to be ‘absolutely correct’?? What or who reviews, oh, say Trump? How do we evaluate his intelligence? Why would it matter if their ‘intelligence’ is greater than ours? I get the feeling you are trying to say something about strong AI(?) and maybe machine overlords or something but not sure where you are going here. There are certainly some pitfalls we need to watch out for wrt emergent strong AI, but I don’t think we need to worry about it taking over beyond what we are already going to be handing over to it to do to offload off of our own shoulders. Like I said, I think the future is a fusion between human intellect and strong/weak expert systems AI, each doing what they do best and complimenting the other. You can already see how expert systems are making a huge difference in many professions and in advancing technology, science, the arts and hell, even playing games (both virtual and real life).
Those are some odd objections. Computers aren’t just tireless, they are insanely fast. Yes, AIs have to learn but they can do it in a fraction of the time a human can. Initially we would evaluate them the same way we evaluate human decision makers, with humans already expert in the field. Probably we would eventually use a trusted AI to evaluate, speeding up the process more and sure, maybe plug in some poll numbers to guide the decisions - depending on what the particular AI is going to be working on.
There’s a bit of a problem with the math here 
To know if human lifespan can be stretched to 200 years, we’re going to have to watch it for 200 years
If any new anti-aging techniques are invented, say, by 2050- whoever is given the new treatment will only be 49 years older by 2099.
And the research is going to be a long, long process, with lots of problems, and deep ethical issues.
Remember Dolly the sheep–the first mammal to be cloned and genetically engineered?–She had great genes, and scientists were optimistic for a year or two. But then she died totally unexpectedly, a very early death from “natural causes”–her genes.
There’s a lot of… recency bias? Picayune-ism? Something like that… in the OP. To us, television and the Internet are game-changers, but people who saw the dawn of the telegraph would just say they’re extensions of that.
XT,
I agree that the transition has to be cooperative systems. Like co-piloting in automobiles rather than full automated driving.
But consider a case like health care. Given the statistics of coverage base, income spread, minimized management costs and desired outcome, a computer would create a system closely related to single payer. Probably with multiple underwriters. Congress and special interests would not accept the result, so the problem is political, not technical.
In all cases the solution is implied by the way the problem is stated. How would a computer improve decision making?
Neuronal systems can find the centroid of complex data but it’s not magic - still a case of: if garbage in, then garbage out.
Crane
Reminds me of a comedy bit I saw. “Imagine if the internet was invented before the phone? Yeah, it’s amazing we can type things and send that text around the world but with the phone you can just pick it up and TALK to someone!”*
*joke was pre-VOIP which kind of kills it.
Well, what you’re talking about has a lot to do with how and at what pace we would hand over decision making to computers. But there’s a tonne of ways they’ll infiltrate decision making whether it gets votes or not. Fortune 500 companies aren’t counting votes. It will probably (if all goes well) hit a tipping point where special interests will be at a loss to stop it. Even if there’s some political roadblocks, like in the UHC debate, we could easily say “ok computer, we can’t do UHC, what’s the best way to run employment based healthcare?”. If it gives an answer, and it works, it becomes all that much harder to ignore the occasional AI comments that “but you know, UHC would still be better”.
Also,
Healthcare Decision AI: “My political analysis bot friend says if you use the following phrases and arguments, people will vote for UHC”
is probably in the future.
Good points.
It would be fun to observe the board room responses to computer generated results. Kind of like the responses they routinely give to engineering.
It’s still a matter of weighting. Given to a computer, the Challenger decision to launch would have been ‘NO’ unless the pending Reagan speech had been a heavily weighted ‘YES’ input.
Crane
If I understand it correctly, Deepmind was given some tools and an environment and an objective. It tried out different possibilities until it figured out how to make a series of figures navigate mazes and obstacle courses.
If this video doesn’t simultaneously delight and unnerve you, we’re very different people.
This seems very, very wrong but I can’t quite put my finger on why.
It’s not really wrong, in the way that a modern combat rifle is really just an extension of a muzzle loaded black powder matchlock…or a modern composite bow or ax is really just an extension of their stone age antecedents. Looked at that way it’s all really the same thing.
I’d guess the folks who were about for the dawn of the telegraph would still be astonished by cat videos and porn, however. 
The fable I think about for understanding technological change is this: challenge Archimedes to invent a machine that will enable me to shout and someone 500 miles away to hear my words. Not only wouldn’t he be able to invent it, he’d likely be able to prove mathematically why no megaphone could physically amplify a human voice to that degree: the challenge, he’d argue, is scientifically impossible.
Give Samuel Morse the same challenge, and he–almost certainly the lesser intellect–could figure out how, at least, it’d be possible.
This is the sort of sea-change that technology can offer.
The horizons true AI will open may give us ways to solve problems we considered scientifically impossible. I’l make no bets.
The boardroom discussion would be amusing. I brought up the AlphaGo experiment earlier. In game 2, AlphaGo made a winning move that it figured would only be used 1/10,000 times by a pro human player. Without 10,000 board members, AlphaInvestment’s more extreme advice would probably get a rough ride until it proved itself. Lol.
The real fear, imho, is that an AI government might gain trust through a million reasonable and intelligent decisions but, because of underlying learned goals and parameters that we don’t quite grasp in time, leads to something horrible simply because we didn’t teach it that certain things are too horrible to risk or accept.
[QUOTE=Crane]
But consider a case like health care. Given the statistics of coverage base, income spread, minimized management costs and desired outcome, a computer would create a system closely related to single payer. Probably with multiple underwriters. Congress and special interests would not accept the result, so the problem is political, not technical.
[/QUOTE]
It would certainly use statistical analysis and predictive modeling, not gut instinct or intuition to craft a plan. I’m unsure why you think that, having the data in front of them showing that it would save money, the board or Congress or whatever would ignore the data at that point. Especially if the AI doing the projection and model has a proven track record. To date, that hasn’t happened, and we are still in the realm of gut feelings and instinct, but if that wasn’t the case I don’t see the issue.
The same way it works today. A lot of the earlier paradigm on, say, Wall Street with how and when stocks were bought or moves were made were based on gut instinct. Today, much of that has been supplanted by expert systems that use predictive modeling. And the results have been pretty dramatic. It’s one thing to let your political prejudices rule when there is a question about the results, but when real money is on the line? Basically, it will be a monetary Darwinism thingy going on, with those who can’t adapt becoming the equivalent of the old European aristocracy that eventually went down because they couldn’t change or adapt. Too bad for them.
Depends if you put garbage in and if your system is garbage. But a lot of expert systems have proven themselves quite a lot better than the older systems used. This goes for law enforcement, business modeling, finance, network/system/cybersecurity…a host of different human activities and a building track record.