Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Scoops Novel]

The AI Revolution: The Road to Superintelligence (part 1) and 2

Elon musk seems to like it too.

[Levi]

The big question for me is what is going to come first. 

A)  AI's becoming super intelligent.
B)  Humans altering themselves to become AI's (uploading, augmentation, whatever).

I think A is more likely to come first, but we might be able to piggy back on advances to get B soon after.

(I miss the transhumanism thread)

[wierd]

There is no reason why both cannot be happening simultaneously-- In fact, that is exactly what I have been observing.

Take for instance, the artificial hippocampus that was used on rat brains a few years ago.
http://www.popsci.com/technology/article/2011-06/artificial-memory-chip-rats-can-remember-and-forget-touch-button

Happened right alongside modern "Deep learning" AI research.

[hops]

Too bad that sticking people's brains into things to create AI is probably unethical.

But hey, maybe by that time we can grow custom-made brains.

[wierd]

As long as the cells arent human, it's fair game.

http://news.discovery.com/tech/robotics/brain-dish-flies-plane-041022.htm

[Dutchling]

the future is now

[Sergarr]

I wouldn't have used neural networks for stuff like this - they tend to react unpredictably when they're operating in conditions different from what they were trained in. Also, 25 000 is a really small amount of neurons.

Pretty sure our brains also contains some stuff other than neurons that's necessary for full functionality of the brain.

[hops]

As long as the cells arent human, it's fair game.

http://news.discovery.com/tech/robotics/brain-dish-flies-plane-041022.htm

it's fair game now, but if history told us anything, ethical standards will continue to rise higher. In fact, I have a feelings that even robots will have rights in the future. I'm not saying it's a good or a bad thing, but it's definitely inconvenient.

[wierd]

That's a barrel of fish that I dont see getting opened any time soon, because opening it would upset an already festering mess we have right now.

Specifically, the whole abortion ethics issue, with Roe-vs-Wade, and all that not fun stuff we dont want to talk about in this thread because it should be discussed elsewhere.

However, as it relates to THIS issue-- with synthetic intelligence and research on cultured neural tissues--  If you ascribe a glob of some few thousand neurons on a silicon substrate as being alive, and with rights-- how can the same determination not be given to a developing fetus with substantially more than that?

See how the problem gets really sticky really quickly?

This is why human cells arent allowed in the research without extensive oversight, amongst others, and why I dont see the legal system allowing that kettle of fish to be untinned without some pretty damned alarming data to force it open with.

[Gentlefish]

Well you see it's easy. They were rat neurons, not human ones. Therefore it's okay.

[lemon10]

Yeah, abortion issues in regards to AI is pretty small fries compared to the more fundamental issue: We will eventually be creating beings as intelligent than humans with the goal of them being slaves. Willing slaves, but slaves nonetheless.
We are a very long ways from that being a real issue, but it's especially troublesome as there isn't going to be any way to draw a clear line.

[Bauglir]

I dunno, we pretty much do that with kids by encouraging them to be hard workers and strive to achieve various goals. It seems to me that, in all probability, raising an AI to human-level intelligence will follow a similar teaching and ethics-guiding process, and that as we do this we will face all the same dilemmas as in raising a child who will grow to be smarter and better than us.

[Frumple]

Yeah, abortion issues in regards to AI is pretty small fries compared to the more fundamental issue: We will eventually be creating beings as intelligent than humans with the goal of them being slaves. Willing slaves, but slaves nonetheless.

Or we could always just not do that. Most people don't make kids for the express purpose of future laborers.

'Course, the AI will still have the issue of being able to support themselves (i.e. keep whatever computer is running them powered and repaired), so they might not want to go be a musician or whatev' to avoid the digital equivalent of starving to death in the streets, but eh. They'll likely have better tools to pull that off than most humans have.

I mean, really, we've had a few thousand years to get the hang of making intelligent beings without being brainfuckingly immoral about it. Admittedly not the best track record, but it's not like this is a complicated subject in and of itself. Made complicated pretty often, sure, but "Don't be a colossal dickwad to your offspring" is actually not some grand philosophical conundrum.

[Criptfeind]

I don't get it. Are you saying we shouldn't make a legion of robot butlers to cater to our every whim (if that was possible)

Because if so, I don't want to subscribe to your moral philosophy.

[Bauglir]

Oh, by all means, create the legion. Just don't fill your robots with self-aware, hyperintelligent AIs.