Well guys, it's almost time to bow down before our robot overlords

[AshyPhoenix]

http://www.gizmag.com/brain-computer-simulation/25349/

This is simply fascinating. While it's not an actual artificial intelligence, it's a massive step forward in that field.

What would you do with your robot assistant?

[G-Lexington]

Put him in charge and ask him what he wanted ME to do.

Lex

[Nishin]

He shall do all my paperwork, and chores...

[Axxess]

Like many things in the beginning, it's too primitive to be of much use now. It needs to be further developed and refined before it'll become a true asset.

Although, I'll admit I'm jealous that its handwriting is better than mine

[ixthrock]

at the doubling rate for technology...
we'll be uplifting silicon life in no time

[Harke the Boeotarch]

I grew up in the eighties...



"...but best of all they'll liberate us from mundane and time-consuming household chores."

Mind if I don't hold my breath?

[Ram]

What would you do with your robot assistant?

Can I sleep with him?

[Kulindahr]

This has to converge with two other things to give us something truly powerful: expert systems, and learning systems. Once it can learn on its own, file that and build up its own knowledge, then this ability will be far more valuable.

We have primitive versions of that now, for example robot vacuum cleaners that decide when the floor needs cleaned, and robot lawn mowers that not only determine when the grass needs mowed but reports on conditions such as weeds appearing, poor soil nutrients, and moisture content -- in fact at least one of the newer robot mowers can be linked to a built-in lawn watering system, and that watering system can have canisters with various lawn chemicals such as nitrogen, iron, etc. for fertilizing, moss control, etc. When the mower decides the lawn is too dry, it cues the sprinkling system, and if it has decided that one or more of the chemicals is needed, those get fed into the water line. According to engineering/tech news, future models will hold weed killer so they can spray as they mow, right on the weed as needed.

And however impressive that is, it's still primitive, because the mower is just following built-in parameters. But combine that with a learning system and the mower can expand on its base -- for example, learning that a shoe in the yard is to be gently pushed out of the way while a pile of dog poop should be avoided and the human of the house alerted. Throw the neural system in the article on top, and we'd have a robot that starts as a lawn mower and grows to be a grounds-keeper.

Once they get better developed, what I want is a driver. If I can get the RV I dream of and roam the country, I'd be able to spend travel time watching the scenery and/or writing, or catching naps so I'll be ready for the next hike.

[goldenmoth]

I hope it knows Asimov's three laws!

[Kulindahr]

I hope it knows Asimov's three laws!

It's interesting that I've seen discussions between AI geniuses discussing how we may need those laws, but have never heard one word about how to build them into real robots. If you consider the level of judgment they require, I'd expect some in-depth attempts at addressing that issue.

[Hard-up1]

I'd program him to dispense gumballs. If he is truly intelligent, he should have to go through a menial labor stage like so many of the rest of us have, to be more tempered.

[penayforay]

It's interesting that I've seen discussions between AI geniuses discussing how we may need those laws, but have never heard one word about how to build them into real robots. If you consider the level of judgment they require, I'd expect some in-depth attempts at addressing that issue.

not to mention how the robots might just plain misinterpret the three laws and just kill everyone.
then theres also the question of well even if you manage to program the three laws in...can't some asshole just program it out?

[goldenmoth]

Not if you build the entire positronic brain around it. Plus robots don't misinterpret laws or directly disobey them. They can't. It's their programming. A computer can't decide to open Excel just because you click on Word and it prefers Excel. The three laws are sound with the occasional, rectifiable bug.

[Kulindahr]

not to mention how the robots might just plain misinterpret the three laws and just kill everyone.
then theres also the question of well even if you manage to program the three laws in...can't some asshole just program it out?

It would be in the hardware, wired in rather than programmed. No matter what software games you played, there wouldn't be a way around it.

[Harke the Boeotarch]

Not if you build the entire positronic brain around it.

Positronic brain ey? The gamma radiation alone would probably annihilate humanity...

[penayforay]

It would be in the hardware, wired in rather than programmed. No matter what software games you played, there wouldn't be a way around it.

Is this set in stone?
even if were in the hardware, couldn't you replace the hardware?

Not if you build the entire positronic brain around it. Plus robots don't misinterpret laws or directly disobey them. They can't. It's their programming. A computer can't decide to open Excel just because you click on Word and it prefers Excel. The three laws are sound with the occasional, rectifiable bug.

I think the three laws are kind of open to interpretation. What constitutes harm to a human?
As opposed to opening word/excel which I dont think is open to interpretation at all?

[Kulindahr]

Is this set in stone?
even if were in the hardware, couldn't you replace the hardware?

You mean the entire brain?

I think the three laws are kind of open to interpretation. What constitutes harm to a human?
As opposed to opening word/excel which I dont think is open to interpretation at all?

There are two entire collections of short stories, and several novels, dedicated to exploring this. Of course they're also very good stories.

[penayforay]

You mean the entire brain?

I mean...why not?

There are two entire collections of short stories, and several novels, dedicated to exploring this. Of course they're also very good stories

I think Asimov had one story named "Sally" where you see the first one being broken. Only they have cars with brains killing someone I believe.
I feel like the three laws would tell the robot how it SHOULD function. but what if it were to malfunction?

[Kulindahr]

I mean...why not?

Probably because it would be cheaper to just build your own robot.

I feel like the three laws would tell the robot how it SHOULD function. but what if it were to malfunction?

Asimov handled that one too: build the Three Laws in in such a way that for them to fail, the brain would become useless anyway.