Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×

Comment Re:What we don't know... (Score 1) 564

Babies are unconscious then are some point they are. That's the mystery, and why my use of the word initial is appropriate. How do I trigger the process of consciousness, does it occur automatically once all the pieces are present, perhaps, or does it need a bootstrap or pilot light. Interfering with part of a brain to cause unconsciousness is super interesting, but by itself it doesn't give us all the answers. (nothing ever does)

If you want to talk about mechanisms, like recursion, then we're going into theory. Theories which I don't have the resources or capabilities to test on my own, so I don't really wish to speculate too deeply there. Maybe it emerges gradually, maybe it is quick like waking up. No idea, but it seems that part of what happens is testable. Should be an interesting future when we find out more.

Comment Re:What we don't know... (Score 1) 564

Oh I agree it's being worked on. But it sounds like a very familiar article, as in I think I read similar articles in the 80s, 90s and 2000s. But the mechanism that triggers initial consciousness is, to the best of my knowledge, still a mystery. It will one day be solved. maybe the article you read really does have it figured out, the ones I've seen were just speculation with theories that could not be realistically tested without interfering with the process.

Comment Re:What we don't know... (Score 1) 564

Oh and understanding what needs to be simulated and the initial state of the human brain. How is consciousness born? We've wondered that for centuries, we don't have the answer yet. Will we eventually know all of this and have the capability to duplicate human intelligence? I don't doubt it one bit! Will we be there in 30 years, at least down the path you suggest? Extremely unlikely.

Comment Re:What we don't know... (Score 1) 564

A cascade of AI's capabilities in a short period of time seems likely, but we haven't made much progress yet. And even with a 100 fold increase in processing power we haven't managed to take our simple learning models and make them 100 times more powerful. There is a real scalability problem right now. So it feels a bit like putting the cart before the horse to worry about these what ifs.

Comment What we don't know... (Score 4, Insightful) 564

Your cell phone is less capable of learning than a jellyfish. Although your cell phone can sometimes simulate very simple learning under extremely rigid frameworks for learning.

a human competitive AI in 30 years? seems unlikely given the almost zero progress on the subject in the last 30 years. But maybe we'll hit some point where it all cascades very quickly. Like if we could do a dog level intelligence it is not a far leap to do human level and super human level. But we have trouble with cockroach levels of intelligence, or even defining what intelligence is or how to measure it.

AI research for the last several decades have taught us how little we know about the fundamental nature of ourselves.

Comment Re:Did anyone care anymore? (Score 3, Funny) 105

Yup we didn't need print magazines in the 80's. Because downloading images at 2400 baud and displaying them on your 8 color computer was vastly superior to full color printing and inexpensive monthly delivery.

I used to print out source to do code reviews, because I was too impatient to wait for VGA projectors to be invented.

Comment Re:global warming? (Score 1) 123

it uses it therefor it traps. but I assume you meant the word "use" in some other way...

we are pulling hydrocarbons from deep underground and if we continue to either burn it or allow it to decompose, it will become CO2 and contribute to the greenhouse effect. Anything that could sequester carbon long term is better than the nothing we are doing today.

the eventual goal would be to stop pulling material out of the ground, and reverse the trend and put carbon back faster than we take it out. We're so far away from doing that that I don't believe there are any viable plans written up to accomplish it.

Slashdot Top Deals

"If anything can go wrong, it will." -- Edsel Murphy

Working...