Oh noe.. Planetarian is taking over REAL LIFE

[Sephi]

Originally Posted by Un1234l

alright, maybe bad title.

However, it is very interesting to see how far this technology will advance within our lifetime...

How much do you guys think this costs/will cost?

Can you imagine yourself buying one?

In fact, what would you guys use them for? I personally wouldn't find any use for them, unless they can accompany me on an instrument. That would be pretty cool.

Face of that robot looks a bit freaky imo. Probably due to the strange way the eyes sometimes moves.

I can imagine buying one if it can cook the dishes i like, and do the cleaning.
And i don't even care how it looks than. For all i care it looks like Bender from Futurama, as long as it doesn't sprout the same foul language as Bender than it's fine with me

[Tran225]

I love engineering and want to major in this field and probably live in japan doing engineering for computer but i really hope i will not see the day where i would see robots walking the earth like its nothing special lol kinda just freaky

[onehp]

I'm looking for John Connor

[WanderingKnight]

Well, AI is certainly impossible unless we find out what algorithm or set of algorithm a sentient being's brain utilizes to learn things. And still, it's pretty likely that such set of algorithms will be enormously complicated--and that it would vary not only amongst species, but also amongst individuals of the same species.

And even so, one would have to actually breed an AI and subject it to the necessary stimulus in order to get the response we want out of it. In fact, you could almost argue that once that boundary is crossed, you could say you're actually dealing with human beings, since their output would be exactly the same (or at least pretty similar) to what you would expect a human being would do were he subject to the same input.

In short: AI is not really feasible, at least not with our knowledge. Quantum computing won't do much, either, since at most it would enhance our computational power many times, not our ability to input data to a false AI. At most, what we could get are highly specialized AIs that could somehow "learn" from a limited set of input data--once extraneous data comes into play, the thing simply can't process it further. It could learn, but only with a specific set of expected input data.

[ClockWorkAngel]

At best with AI we'll get is something like this;

A great computer with an infinitely large capacity to receive and perceive: statements, body language etc. and a list of responses. The list would be huge, almost infinitely large and the need to judge which is the best response would be huge too. But it'd never be "human", it's never actually be at the level of Intelligence, it is afterall a program. With our technology, we're not even possible of creating the most primitive AIs.

[Vexx]

Originally Posted by ClockWorkAngel

At best with AI we'll get is something like this;

A great computer with an infinitely large capacity to receive and perceive: statements, body language etc. and a list of responses. The list would be huge, almost infinitely large and the need to judge which is the best response would be huge too. But it'd never be "human", it's never actually be at the level of Intelligence, it is afterall a program. With our technology, we're not even possible of creating the most primitive AIs.

hmmm, you're thinking about traditional programming -- with that, yes, its a limited response (though its not as big as you think to fool most people most of the time).

You should take a look at neural simulation programming ... which has interesting promise and some fascinating results. In actuality though I'm going to assume that rather than simulate neural systems -- we're going to short cut and just use organic brain tissue as the "learning core" in advanced systems. At some point, we'll accidentally create something that doesn't want to be turned off.

[ClockWorkAngel]

Problem with that is once you put the organic brain tissue is that could you count it as being somewhat alive? Of course we as humans have no attachment to individual cells, tissues etc; it is the sum of the parts (Brain + organs + etc) which gives us the inability to kill it without saying "You killed a living thing". Its best if we don't develop such an attachment to something that would be created using brain tissue (Would we use donor ones or grow them?)

It sounds interesting (neural simulation programming) and it may infact be that step up we need for AI. In a way we could one day be like that; being able to transfer our thoughts etc (essentially electric signals, though infinitely complex) and thus essentially being able to be immortal. But what I worry about is that would it be the case that the being/thing before the shut off = the one once you "reboot".

[cuteycindyhoney]

I'd like an android driver for my car. That would be something!