Information and Evidence of the Singularity

[Samir]

Post websites and other refences here. Feel free to discuss too.

The Singularity Institute for Artificial Intelligence
SIAI - Blog

Singularity Hub

Accelerating Future
Michael Anissimov - Blog

[Samir]

'Eureka machine' can discover laws of nature
Meet Adam and Eve: AI Lab-Bots That Can Take On Reams of Data

Robots that can do Science!

[GenXer1965]

I read the latest Blog entry on Copenhagen and though I must agree mostly that it has become a big industry and everybody is in it for the money. I must say that I believe climate change, with large human component due to Global warming gases is really happening. i.e. I don't let myself get totally cynical, throw the baby out wtih the bathwater. Secondly I think any "singularity" like in Terminator is just not possible due to the Peaking of all earthly resources, water soil, oil, coal, gold. Machines are made from metals and their movement fueled by fossil fuels. These are running out. We use the resources for wasteful purposes and not for limited R+D puproses to make the samrtest machines possible so tha any possible advances are bing flushed or thrown into the garbage heap. Tantal in Central Africa for 1 billion throwaway mobile phones, ltihium for laptop and car batteries, ever rarer copper depostis and deeper gold deposits. Biological life is dominant on earht for millions of years for a good reason. If rocks could walk and talk on theri own they would have done it without our help (we are those rocks in fact.)

[MnMark]

I don't see how computers could become more intelligent than people.

I write software for a living. I am not an expert on leading-edge software technologies. But having said that, I know enough to know that software is simply a set of algorithms, a series of commands. For a computer to begin to make itself more intelligent, intelligence would have to be simply another algorithm that could be written by a software engineer and programmed into the computer. Intelligence is not an algorithm. There is no set of steps A, B, C, etc, which you can write that will lead in a logical, mathematical way to the discovery of new, more intelligent ways of accomplishing something. The great vulnerability of science is that while it is a fantastic way of disproving or proving hypotheses, it has no explanation for how a scientist comes up with a hypothesis in the first place. The scientist simply studies a problem, gathers facts about the problem, sits and looks at the problem...and in what is really quite a mystical event, an idea pops into his head - "aha, why don't I try this!" There is no algorithm you can write to duplicate that process of inspiration, and since there is no way to write such an algorithm, there is no way to program it into a computer.

To reiterate: a computer cannot solve a problem through inspiration, like a human being. A computer can only solve a problem that a human being has programmed it to solve, which means it is a problem the human being already knows how to solve, which means the human being is always "smarter" than the computer. The computer can simply run through the grunt work of exercising the algorithm much faster than a human being. But don't confuse the ability to quickly execute an algorithm with intelligence.

Secondly, a computer has no will. It has no consciousness and no intentions, no desires. It has no consciousness that drives it to seek higher quality, nothing that gives it a reason to do anything. So there is no reason to think that, even if somehow you could write an algorithm that captures the essence of intelligence and program it into the computer, the computer would DO anything at all. It has no motivations, no emotions, no desires, no reason to DO something rather than do nothing. Living things act because they have a drive to grow, learn, expand, become ever more powerful and to bring ever more quality, as they perceive it, into their existence. Computers have no consciousness to motivate them to do that, and you can no more write an algorithm for consciousness than you can for intelligence.

So this whole "singularity" thing with regard to computers strikes me as science fiction. I am open to being convinced but first you have to explain how you'd capture the essence of "intelligence" in an algorithm, and not just computing speed. They aren't the same.

[John]

> I read the latest Blog entry on Copenhagen and though I must agree
> mostly that it has become a big industry and everybody is in it
> for the money. I must say that I believe climate change, with
> large human component due to Global warming gases is really
> happening. i.e. I don't let myself get totally cynical, throw the
> baby out wtih the bathwater. Secondly I think any "singularity"
> like in Terminator is just not possible due to the Peaking of all
> earthly resources, water soil, oil, coal, gold. Machines are made
> from metals and their movement fueled by fossil fuels. These are
> running out. We use the resources for wasteful purposes and not
> for limited R+D puproses to make the samrtest machines possible so
> tha any possible advances are bing flushed or thrown into the
> garbage heap. Tantal in Central Africa for 1 billion throwaway
> mobile phones, ltihium for laptop and car batteries, ever rarer
> copper depostis and deeper gold deposits. Biological life is
> dominant on earht for millions of years for a good reason. If
> rocks could walk and talk on theri own they would have done it
> without our help (we are those rocks in fact.)

The development of the Singularity will not be affected by lack of
oil, coal, gold, etc.

The development of super-intelligent computers is going on in labs
around the world. Our DoD has announced the first generation of
intelligent battlefield computers for 2015.

** Army's 'Future Combat System' calls for autonomous robot soldiers by 2014
** http://www.generationaldynamics.com/cgi-bin/D.PL?xct=gd.e050228#e050228

These are intelligent devices that will fight wars and kill people.
You can be absolutely certain that the Chinese, the Japanese, the
Koreans, the Indians, the Russians and the Europeans, and probably a
lot of others are all doing their own development in this area. The
first war among autonomous intelligent computers will probably be
fought in the early 2020s.

Sincerely,

John

[John]

> I don't see how computers could become more intelligent than
> people.

> I write software for a living. I am not an expert on leading-edge
> software technologies. But having said that, I know enough to know
> that software is simply a set of algorithms, a series of commands.
> For a computer to begin to make itself more intelligent,
> intelligence would have to be simply another algorithm that could
> be written by a software engineer and programmed into the
> computer. Intelligence is not an algorithm. There is no set of
> steps A, B, C, etc, which you can write that will lead in a
> logical, mathematical way to the discovery of new, more
> intelligent ways of accomplishing something. The great
> vulnerability of science is that while it is a fantastic way of
> disproving or proving hypotheses, it has no explanation for how a
> scientist comes up with a hypothesis in the first place. The
> scientist simply studies a problem, gathers facts about the
> problem, sits and looks at the problem...and in what is really
> quite a mystical event, an idea pops into his head - "aha, why
> don't I try this!" There is no algorithm you can write to
> duplicate that process of inspiration, and since there is no way
> to write such an algorithm, there is no way to program it into a
> computer.

> To reiterate: a computer cannot solve a problem through
> inspiration, like a human being. A computer can only solve a
> problem that a human being has programmed it to solve, which means
> it is a problem the human being already knows how to solve, which
> means the human being is always "smarter" than the computer. The
> computer can simply run through the grunt work of exercising the
> algorithm much faster than a human being. But don't confuse the
> ability to quickly execute an algorithm with intelligence.

I gave a detailed answer to this question in my book chapter:

** Chapter 7 - The Singularity
** http://www.generationaldynamics.com/cgi-bin/D.PL?d=ww2010.book2.next

There you'll find software algorithms can implement computer
intelligence and do all the things that you say can't be done.

As for the "mystical event," that chapter also gives a couple of
examples of such mystical events, and shows how a computer might
solve them.

One of them is Andrew Wiles' discovery of the proof of Fermat's Last
Theorem, one of the most mystical accomplishments of all time. There
was indeed a point where an idea popped into Wiles' head exactly as you
say: "Aha, why don't I try this?"

However, there's nothing mystical about this. Wiles was applying
something that he had learned years before. He had put together two
pieces of knowledge in the way that you might put together two jigsaw
puzzle pieces. And that's the point. Computers can solve jigsaw
puzzles by brute force a lot faster than humans can.

> Secondly, a computer has no will. It has no consciousness and no
> intentions, no desires. It has no consciousness that drives it to
> seek higher quality, nothing that gives it a reason to do
> anything. So there is no reason to think that, even if somehow you
> could write an algorithm that captures the essence of intelligence
> and program it into the computer, the computer would DO anything
> at all. It has no motivations, no emotions, no desires, no reason
> to DO something rather than do nothing. Living things act because
> they have a drive to grow, learn, expand, become ever more
> powerful and to bring ever more quality, as they perceive it, into
> their existence. Computers have no consciousness to motivate them
> to do that, and you can no more write an algorithm for
> consciousness than you can for intelligence.

What you're talking about here is goal-seeking, something that's very
easy to program in a sufficiently powerful computer. If a computer's
goal is, say, to kill all the humans, then it will accomplish that
goal by putting together bits of knowledge until it figures out how
to do it.

> So this whole "singularity" thing with regard to computers strikes
> me as science fiction. I am open to being convinced but first you
> have to explain how you'd capture the essence of "intelligence" in
> an algorithm, and not just computing speed. They aren't the same.

Computer speed is required because the algorithm that I developed
requires a very fast computer. I estimate that by the 2020s,
computers will be fast enough for a full-fledged implementation of my
algorithm.

John

[GenXer1965]

Ok so suppose they get intelligent robots. They will have the very same problem we have. Finding food and repair of wear and tear over time. Their Food will be some enrgy source and they will have to repair themselves. I see a Matrix or terminator situation. They control us to use us as batteries or some sort of enrgy source and then they use our scrap yards for metals to repair themselves. The children's film Robots comes to mind here. I would love to write a Scifi where Peak Oil and the singularity overlap causing a race for the last resources between humans and smart robots. Cool!

[xakzen]

I read the latest Blog entry on Copenhagen and though I must agree mostly that it has become a big industry and everybody is in it for the money. I must say that I believe climate change, with large human component due to Global warming gases is really happening. i.e. I don't let myself get totally cynical, throw the baby out wtih the bathwater. Secondly I think any "singularity" like in Terminator is just not possible due to the Peaking of all earthly resources, water soil, oil, coal, gold. Machines are made from metals and their movement fueled by fossil fuels. These are running out. We use the resources for wasteful purposes and not for limited R+D puproses to make the samrtest machines possible so tha any possible advances are bing flushed or thrown into the garbage heap. Tantal in Central Africa for 1 billion throwaway mobile phones, ltihium for laptop and car batteries, ever rarer copper depostis and deeper gold deposits. Biological life is dominant on earht for millions of years for a good reason. If rocks could walk and talk on theri own they would have done it without our help (we are those rocks in fact.)

Here is a related article I saw a few months ago which should be really disturbing to anyone who has seen the Terminator movies: http://www.foxnews.com/story/0,2933,533382,00.html

The original report said it would eat animal remains including dead bodies on the battlefield. It has since been updated by company reports to refute this. Sounds like one obvious machine aided design flaw fix may already be in the making.

[OLD1953]

As costs go up, recycling increases. Gold, for example, has been 99%+ recycled for millenia. If you have any gold items, it's a certain bet that some of those gold atoms were present in gold worn by Cleopatra.

Peaking production does not mean instant reduction in production. We've got a long (decades long) taper off period with oil.

The side effect of lithium production will be interesting, Great Salt Lake will become less salty, as will a number of other saline sinks.

http://www.jstor.org/pss/3955088

http://www.planetark.com/dailynewsstory ... /story.htm

One does wonder about the impact on the brine shrimp.

[burt]

I don't see how computers could become more intelligent than people.
...
So this whole "singularity" thing with regard to computers strikes me as science fiction. I am open to being convinced but first you have to explain how you'd capture the essence of "intelligence" in an algorithm, and not just computing speed. They aren't the same.

Just my answer:
What you need is efficiency, not human intelligence, so I do not agree when you compare the 2 (even if I believe that we can mix bio-tech. with computers, but this is another subject, a super man, WITHOUT conscience).

In less than one century robots will be able to reproduce themselves (without human intervention) and be self-programmed, BUT, they need a lot of energy, and that could save the humanity (other remark on the same blog). The other things that could save us, are the hackers that could put some Chaos in the process, BUT governments are already employing hackers and in a shot-time they will be employees like anyone (just to attack the “others’” computers).

Why will human need more efficiency (only the black side, we could use technology for the best of the humanity, but I think it is impossible, for 2 reasons -1- the type of era we are living and the lack of ethic that ANY government has -2- the only existence of this black side of the man):
-- To kill its neighbour
-- To manipulate <election> so that "normal people" think that they are free (and a mass is much, much more stupid that one human, so I'm pretty sure that we will have the sufficient knowledge about social psychology to manipulate ANY election in less than 10 years (2020, at the LATEST), maybe today, but I think it is not the case, we are on the test phase)
-- To program computers (a lot of works have been done on that, we need only more power to make them available)
-- To decide when to automatically kill old and “inefficient” people in the hospitals (the wording will be: we are going to take care of you...but...) (I'm personaly almost certain it will be used to cut the deficit of the pension funds in such a way that the medical person won't understand it).
-- To control the media, so that you are manipulated even without knowing it. The control of Internet is underway now to "protect us".

Open a window that make possible MORE POWER to certain people and humanity WILL choose this way... No escape. On my point of view this IS a generational process, "wanted" in a Crisis era.

One time the computers will be more efficient than us, a spiral will begin without any hope of getting out of it, and the next step will be when we (as humans) will finish as the little dogs of the robots.
Believing that we can dominate robots more, and much more, efficient than us (on WHATEVER SUBJECT you may think about, even choosing the subjects you are allowed to think about, the education programs and the war programs) is just living on another planet.

So even if it is not what you call "intelligence" the result will be the same.