NationStates Jolt Archive

Since there was a thread on MT mind-control, and I'm considering opening an international electronic company, I thougt I'd ask it:

How plausible is Artificial Intelligence and or somewhat autonomic robotics in MT? Like those seen in I, Robot, but without the weird rebellion and not quite as advanced. Of course, I know if it was vaguely possible they could only be sold for millions of dollars, and probably could only do simple commands. I wouldn't use them for police/military. Thoughts & comments?

Since there was a thread on MT mind-control, and I'm considering opening an international electronic company, I thougt I'd ask it:

How plausible is Artificial Intelligence and or somewhat autonomic robotics in MT? Like those seen in I, Robot, but without the weird rebellion and not quite as advanced. Of course, I know if it was vaguely possible they could only be sold for millions of dollars, and probably could only do simple commands. I would'nt use them for police/military (though there is some cloning in MT, I think the best 2 examples are Kraven's Capitol Police and Automagfreek's Sentinels). Thoughts & comments?
KC CP are not clones

Well, IRL the Blue Brain project is capable of simulating 1/2 of a mouse brain on supercomputers as of April 2007, and this simulated brain does display thought patterns associated with biological mice. They plan to scale it up and increase the simulation's depth over the next year, with IIRC a goal of human neocortex simulation by 2008. Since the neocortex is the part that is responsible for higher-level thinking, it is likely that a plausible artificial human intelligence would be functional by 2008 or so. However, since the brain is being gradually constructed, other regions of the brain would still need to be added and it probably wouldn't be capable of true conscious thought and communication for another couple years.

So, based upon the rapid growth in supercomputing and the trendline of this project (which it has in fact exceeded, so these numbers may in fact occur earlier than projected even by the BBP itself), I could see fully functional, conscious human-brain AI within the MT timeframe, maybe around 2010-2012 although possibly earlier depending on economic and governmental factors. However, it would be extremely expensive since it would require a significant number of supercomputers to operate; also, it would not really be any more powerful than a human brain, so its contributions would be limited at first but would increase rapidly after that point. It would cost hundreds of millions of dollars to field even a few dozen of these units in an MT setting, and their utility would be limited and inferior to humans in the same occupation.

Vetalia is currently in the year 2023, so we do have strong AI and citizen robots. Of course, this might also be earlier than other nations because we have focused on technology, especially in IT, and that has advanced that field faster than it might have in other nations.

KC CP are not clones

Thanks for the tip, post edited accordingly.

Any answers?

KC CP are not clones


OOC: Sentinels are no longer cloned either.

Thanks AMF, I'll remove that bit then...

OOC: Sentinels are no longer cloned either.

first time i heard of them sad to say, i just knew CP are not clones as i read how they are born still feel sorry for those wamon

*Points out to his first thread, which is still open*

http://forums.jolt.co.uk/showthread.php?t=488758

Perhaps what I'm trying to say is are AIs feasible (yes, expensive) in MT or at least PMT?

Perhaps what I'm trying to say is are AIs feasible (yes, expensive) in MT or at least PMT?

there are low tech ones arounds now

Perhaps what I'm trying to say is are AIs feasible (yes, expensive) in MT or at least PMT?

Generally, yes. Let's say that by 2010, there is a functioning simulated human brain. Based upon continued growth in supercomputer processing power, the cost of simulating one of these AIs would halve every 12 months or so, so by the year 2023 they would be 13 times less expensive than they were in 2010.

However, a truly economic AI program would probably not really be available until at least the 2030's; they could be used for government and corporate purposes, but would still be exceptionally expansive and out of individual reach.

So, if I opened up some international company that sold robots similar to those seen in I, Robot and claimed it MT, there wouldn't be any tech breach?

Sorry for being unclear.

PMT i say

Well, if you note, TWSP already has a robot that can "learn" so you could ask him whether to be MT or PMT

Generally, yes. Let's say that by 2010, there is a functioning simulated human brain. Based upon continued growth in supercomputer processing power, the cost of simulating one of these AIs would halve every 12 months or so, so by the year 2023 they would be 13 times less expensive than they were in 2010.

However, a truly economic AI program would probably not really be available until at least the 2030's; they could be used for government and corporate purposes, but would still be exceptionally expansive and out of individual reach.


Similar to how computer power tends to double every 18 months?

I'm not saying the company would be trying to incorporate the robots into the average man's home and or daily life, but sell several handfulls to national leaders or people with more money than sense that want to fork over 10 million or so for a robot that does chores and takes orders, and offers companionship.

Any more opinions?

Computational speed, as addressed by Moore's Law, (the 18 months thing), isn't the big problem. Some large distributed computing networks can match our estimates of the total operations per second of a human brain (1e13 to 1e16 ... which shows how little idea we actually have), now. Our absolutely pathetic understanding of the nature of reason is the problem.

The fundamental problem with making an AGI is that we do not comprehend how to make one. It's like taking an enormous heap of bricks and cement ingredients, and telling a child to build the Houses of Parliament. We know next to nothing about what makes a thought.

Of course, in NS, you could explain this tremendous hoover dam of a hurdle away by saying you got information from a more advanced society, or some such. There's more than a few tooling around in faster than light space ships, after all.


Also, the 'rebellion' in the "'I Robot'" film is the logical conclusion of Asimov's laws. See R. Daneel Olivaw for more on that particular topic. Suffice to say, the way to minimise human injury is to take control of humanity, how else do you stop the monkies bashing each other in wars, or doing other stupid things, or letting their 'underclasses' die of easily cured diseases? It would be logical if every one of these robot slaves you sell immediately steals all its owner's money and gives it to a particularly worthy charity, before absconding to do good works at the first opportunity.

As with most things in science, it's hard as hell to predict when a given discovery will take place. The "eureka" moment where somebody comes up with a truly conscious computer may come tomorrow, or may never come.

As long as you have had a serious pre-existing research program (which you say you did) and it's already had progress, it would be quite realistic to make the breakthrough in the present.

I'm curious: what do exactly you plan to use your AI for? What level of independance do you expect to give them?

Similar to how computer power tends to double every 18 months?

That would be the basis for it; the raw computational power is where a lot of the cost will come from. The simulated brain itself will be nowhere near as big a cost as accurately running it so as to produce conscious thought and behavior. Ultimately, it's the hardware that determines how much it costs and whether or not it can be fit in to a humanoid body.

I'm not saying the company would be trying to incorporate the robots into the average man's home and or daily life, but sell several handfulls to national leaders or people with more money than sense that want to fork over 10 million or so for a robot that does chores and takes orders, and offers companionship.

You'd need to be able to fit the kind of processing power needed to accurately simulate the human brain in to a humanoid robot first, and that would still take a fairly long time. Obviously, as you get in to the 2020-2030 timeframe and desktop computers can do this scale of real-time calculation, it becomes plausible that you could fit that simulated brain in to a robotic body.

Any time from probably 2040 or later is a given for fully human, conscious robots. Of course, by this point, you're not going to be selling them because they will be demanding human rights (and if your nation's like mine, it will give it to them).

For reference, Vetalia is in the year 2025 at present, but our development in this field is probably more around 2035 or so due to our focus on technological advancement. You could adjust your nation accordingly to fit your own cultural and economic situation.

Also, the 'rebellion' in the "'I Robot'" film is the logical conclusion of Asimov's laws. See R. Daneel Olivaw for more on that particular topic.

Eh, personally, I'm more of a fan of the Tercentenary style 3-law rebellion than the standard "we're gonna liberate the shit out of you" fare. After all, it's just about as effective, and only involves one human casualty (and you can't even prove it!). :D