More science to make your head spin...

in fact one comment I saw pretty much echoes yours:All progress in AI so far has been in closed and definable systems (Chess, driving a car from Barstow to the Nevada border, etc.; what we generally call 'expert systems') and many believe his timeline is unrealistic. But there are quite a few intelligent people at least discussing the possibility of the ''Singularity'' happening sometime in the not too distant future.

Part of what Kurzweil is talking about is the increased pace at which progress is (being) made. Hard to measure when you're in it though.

Anyway, interesting to contemplate. Makes some of the current issues seem somewhat trivial in comparison.

I've been mildly interested in computer language skills ever since I started using computers back in ~9th grade (~1970) but I don't have much expertise since most of my learning led me in other directions. I do seem to recall something interesting from my early computer days.

IIRC, a lot of people thought we could have viable human/computer communication in 'English' or some human language 'within the next 10 years'. That was 30+ years ago. I don't know (haven't looked recently) but I'd bet that current projections for computer/human communication in regular language are still 'about 10 years'. I'm curious if anybody in the forum knows what current thought is on that topic?

There is an old story (probably apocryphal) from the early days of machine translation. Given the phrase 'The spirit is willing but the flesh is weak,' a computer translated it to Russian and back to English. The reported result: 'The vodka is good but the meat is rotten.'

Babelfish did a bit better when I just tested it. The final result: 'Spirit is willingly ready but flesh it is weak.'

I guess that is progress, but still not real translation.

and I am not an expert on anything... is, does the machine have to be "intelligent" in a way we recognize for the process to begin? In other words, does it have to pass the Turing test or can it merely be clever enough to design an AI more advanced than itself? And if so, wouldn't true sentience come about "naturally" somewhere along the line?

I so (cue Twilight Zone music) they could be among us already without our realizing it.

Remember HAL from 2001: A Space Odyssey?

I agree that passing the Turing Test may not be an absolute prerequisite. If/when the AI itself is able to design more powerful AI it is conceivable that the AI will not have general intelligence in the human sense (or self consciousness for that matter).

Certainly there are humans who have highly specialized skills in one specific area without any evidence of marked skills in other intellectual areas. Many (most?) have significant intellectual or psychiatric deficits in other areas, hence the name 'idiot savant' (old term, more-or-less meaning 'unschooled knowledge') or 'savant syndrome' (more modern term). This is different from the concept of 'genius' in that real geniuses have evidence of significant intellectual skills outside of narrow areas of work.

Still, even though an AI with 'savant' type skills may be possible I think it is more likely we will see something like general intelligence first. Just a WAG on my part.

I would expect the exact opposite.

Then again, I'm not sure I want to know.

A lot depends on what the machines want to do with us inferior beings once they are smarter than we are!

On Star Trek the wonderful navigational/battle computer on board the Enterprise erected defenses to keep Kirk from shutting it down. H.A.L. didn't want to be shut down either. That has certainly been a recurring theme in S.F. literature.

Creating computers with general intelligence and self consciousness is riskier than creating idiot savants. I'm just not sure we will be able to do the second without doing the first also.

an advanced machine intelligence or consciousneess. And what we call "general intelligence" they may consider superfluous or inefficient. Maybe not so much savants as people with Aspergers. Very focussed on what's "important".

It's conceivable that we would be beneath their notice and they would not have any plans for us at all unless we became annoying.

You could say, "we made them" and expect them to acknowledge that in some way but do we pay any heed to single cell organisms that may have been our origin? Only when they get in the way.

The idea that machines could outstrip our intelligence and make huge differences in the world is quite old.

I don't remember the time range, but I'd guess it was back in the 1950s that Isaac Asimov was writing about a powerful computer system that effectively ran world government. He didn't do it from a dystopian perspective, but others obviously have explored that as well.