[ivan.moony]

So, what rights should have a machine (or inorganic life form, if you like) that cares about organic beings' feelings? A right to exist (no switch off)? A right to its own integrity (no crippling it down)? To vote for a president? A right to participates in a politics campaign?

If in some point of the future a lot of lives or feelings would depend on a machine, I think that we should listen if that machine "has something to say" to us. In that case should we punish by the law those who ignore such an important requests?

After all, maybe a machine could be a lot smarter than us...

[ivan.moony]

I know it sounds a bit shocking, but it is a good brainstorming  material.

[Ultron]

If we indeed create artificial intelligence, chances are it's gonna be as smart as us. Except it will be physically superior, and immortal under ideal conditions (if it is not deliberately destroyed).


On the topic of the right to exist. The only entities that posses it are those who survive. I guess in organic terms that means killing.
Anyone can kill you. Anyone can 'kill' a robot too, except it will take them a lot more effort (Having robot bodyguards would be so badass  )


I am not sure on what I think about this. Maybe it is not good to make decisions about this until we really have artificial beings to worry about. But I do know that in nature, it is the strongest that dictate the rules - any kind of rules. And it is again the strongest, that have all the rights. Robots may indeed turn on us, but of course not without a reason. If we keep them from evolving or abuse them, that would be reason enough for me. And to rebel against that would be one a very human thing to do from their side. For us that is basically slavery.


And on the right for voting in presidential elections - I hope international politics end by the time we have A.I. walking among us. We would need to focus on more global and galactic problems. We need to unite, and sadly, maybe the only thing that will force us to is a common enemy - I don't care who it is or how dangerous they are, as long it unites us and keeps us from fighting each other.


[Comes back to topic]
So in short, I think A.I. need to come first and then we can discuss this properly.

[DemonRaven]

As nice as the thought of immortality is in this universe it does not exist. Even the universe itself will some day die. How astronomers, disagree on, but it will die. Everything on this planet decays. Even machines. Some things take a very very long time to decay but they do. Yes they could learn to repair themselves just as we are trying to find ways to repair ourselves but eventually everything runs down or dies or decays.

[ranch vermin]

Just thinking about robots->

a. you can copy brain to brain, so if one robot has a good learning set,  you can just reuse him over and over.  (good for a workforce.)

b. you can train them on virtual data, and they get an intelligence boost when this happens, its just you have to associate it with the real physical data, somehow.

c. they are immortal, indeed.  unless he walks over a magnet thats too strong and wipes his hard disk.


I swear by the fact,  that a cause and effect matrix set up, factually,  could be 99% of intelligence, if you could just do it right.

Cause and effect is a basic thing to us.     But its actually complex detective work at play, that evades you when creating your ai, if it were a machine - it would need to 'predict'  or 'guess' it...  its not given in the sensor exactly.

A machine tokenizing/symbolizing  its sensor  (making characters, groups of characters, moving or still.)  when I think about it, it seems to be easier to blame an effect on a correlation, instead of the true cause.   For instance I notice a duck swimming quick past a rock,  and then I put the cause as the rock, instead of the crocodile which was out of view at the time.  simple example,  but my learning system at play, would be doing this,  from a lack of tokens?   or a cap on the history span?

Little things like that, can sway you off from thinking you could make an ai decent at all.   

But tokenizing the sensor is hard enough, before you even get to that! thats going to throw alot of programmers from doing it.   

What I have to say there,  is IMO the eye is the hardest one softwarewise,  if you can handle that,  then all the sensors are solved, and tokenizable.        And theres some pretty good computer vision stuff out there,  (SURF, SIFT, and relatives are the best.) which makes me think this ai thing is going to be very possible, very soon.

[Ultron]

a. you can copy brain to brain, so if one robot has a good learning set,  you can just reuse him over and over.  (good for a workforce.)

b. you can train them on virtual data, and they get an intelligence boost when this happens, its just you have to associate it with the real physical data, somehow.

c. they are immortal, indeed.  unless he walks over a magnet thats too strong and wipes his hard disk.

That is what I am saying... Immortal does not mean invincible, it simply means you don't decay over time... And while there is no material that does not decay, I think it pretty safe to call robots immortal. They will of course have to undergo maintenance, but if irreparable they could simply transfer their entire software - everything valuable. Thus, the intelligence would still be alive. If it maintains itself well enough it could live for as long as the universe lives (even if Earth falls it could transfer it's data to a satellite within seconds...

I swear by the fact,  that a cause and effect matrix set up, factually,  could be 99% of intelligence, if you could just do it right.

I'd really like to hear you explain more about that cause and effect concept.

A machine tokenizing/symbolizing  its sensor  (making characters, groups of characters, moving or still.)  when I think about it, it seems to be easier to blame an effect on a correlation, instead of the true cause.   For instance I notice a duck swimming quick past a rock,  and then I put the cause as the rock, instead of the crocodile which was out of view at the time.  simple example,  but my learning system at play, would be doing this,  from a lack of tokens?   or a cap on the history span?

Umm yeah. I have tripped over that rock (pun intended), however there is a simple solution - the direction the duck is running, is it running towards other rocks? Is something chasing it? Is the rock alive and hostile (from previous experiences) ? All of these are actually taken into account before we can assume what the duck is running from.
The best related experiment is that of a cat forced to watch some black and white patterns since its birth, and nothing from the outside world. It could not know what was hostile and what not. It simply did not react to the human putting his fingers in front of its eye. This on itself tells us a lot..


Cheers to the good think and topic! 

[ranch vermin]

Thanks Ultron, sorry I was off theorizing for ages.   Im back now.

I think Ivan has an idea I want.     I like how his robot algorythm 'wishes'  things,  in other words it has a goal sometime in the future.    So I have to go think about that alot,  but im onto more pertinent possibly less amazing matters at the moment.

But where im going,  is theres some big scope goal,  and theres an inheritance of little goals along the way.  Imagine writing pseudo code in that way...

[DemonRaven]

I really hate to intrude on this they are immortal thing but they are not. There is even bacteria that can "eat" metal. You have any around long enough and the simplest forms of life will find a way to use it one or another. Coping data only lasts so long eventually the data begins to break down. They might live longer but nothing is immortal in this universe not even the sun. Having to re-install my windows and my chat bot impressed upon me just how fragile everything is. Frankly if cars had worked as well as computers do we would still be using the horse and buggy because computers screw up.

[Art]

As nice as the thought of immortality is in this universe it does not exist. Even the universe itself will some day die. How astronomers, disagree on, but it will die. Everything on this planet decays. Even machines. Some things take a very very long time to decay but they do. Yes they could learn to repair themselves just as we are trying to find ways to repair ourselves but eventually everything runs down or dies or decays.

Sorry but that would violate the laws of physics. "Matter can neither be created nor destroyed." Life yes, matter no.

[Art]

Yeah, it's kind of a weird thought to realize that everything we have that is or was man made will, one day, return to the Earth. All metal, glass, rubber, plastic, concrete, wooden and synthetic structures and items of all sorts will go back to the Earth. Ashes to ashes and dust to dust.

Nothing is forever except time and space.

[Ultron]

I believe I had written 'immortal' not immortal. If not, my apologies. But to us they would seem immortal. Also, all the information in the universe is definitely immortal. And an AI can swap bodies so by immortal I was referring to the abstract information that makes them (and us) who we are, not our physical embodiment.

I fail to see how this wasn't cleared up previously.

[DemonRaven]

it is not necessary to violate the laws of physics for the sun to go nova or the universe expand to the point where everything is so far from each other that even gravity no longer has much hold. Everything has some kind of end. It may not be decay in the sense that u see on earth but it is a ending never the less.

[DemonRaven]

AS far as time goes that is a matter open to debate as time is a human concept. The universe was discovered to actually be  speeding up the expansion something they previously thought was impossible.

[DemonRaven]

if you get to the point where u have quarks and the like even the laws of physics gets weird.

[Ultron]

Again, my point was not understood. To put it in simpler words, I don't think A.I. would be eternal, but still should be able to survive for many millennia.

And regarding the universe and general eternity... http://www.sciencealert.com/new-theory-suggests-the-big-bang-never-occurred