What Does a Beer Taste Like After the Singularity?

[Luke212]

From Popular Mechanics Magazine http://www.popularmechanics.com/tech...y/4332783.html

Imagine a techno-futurist rapture (according to some theorists, in the near future) when computer processing and storage has improved to the point where artificial intelligence becomes smarter than human intelligence, and computers are able to improve and refine themselves at an accelerated pace.

That pace would be dramatic, possibly exponential, but the character and exact speed of the acceleration is almost beside the point, since the intellectual capabilities of such machines would be, by definition, beyond the comprehension of mere humans. In other words, one day, our machines will get smarter than us, and then they will be able to make machines that are even smarter than them, then those machines will make other machines that are smarter than the previous machines, and so on, making the conventional human mind more and more of an anachronism.

[Arbite]

Uploading your consoisness wouldn't do much i don't think. It would leave you as normal, but with a replica of yourself in the machine.
A good read on this is Greg Egan's Diaspora.

[aXis]

All of this hinges on the hypothesis that you can make a smarter computer from a dumber one. Or at least make a smarter computer than a human.

So far we haven't demonstrated anywhere near that yet, and the best we may end up doing is simulate a human brain in hardware. It may be faster, but not more gifted or insightful.

[Luke212]

Originally Posted by aXis

All of this hinges on the hypothesis that you can make a smarter computer from a dumber one. Or at least make a smarter computer than a human.

So far we haven't demonstrated anywhere near that yet, and the best we may end up doing is simulate a human brain in hardware. It may be faster, but not more gifted or inciteful.

depends what you mean by smarter. faster can mean smarter in some situations. particularly if all degrees of freedom are accounted for.

[power]

just because they are smart doesn't mean we can't smash them - stupid machines!

ps machine's aren't smart they are fast.

[aXis]

Originally Posted by Luke212

faster can mean smarter in some situations.

Maybe. Say you wanted to brute force a new computer design by simulating all of the possible combinations. Faster computers would help you do that.

but.....

What do you use as the inputs for possible combinations? Faster doesnt let you think "outside the box" so to speak. Currently there requires a human in the loop to provide imagination.

[Menthu_Rae]

Originally Posted by aXis

Maybe. Say you wanted to brute force a new computer design by simulating all of the possible combinations. Faster computers would help you do that.

but.....

What do you use as the inputs for possible combinations? Faster doesnt let you think "outside the box" so to speak. Currently there requires a human in the loop to provide imagination.

/dev/random and a sufficiently high processing speed - there's all the imagination you could ever need.

[aXis]

random what? silicon? gallium? nanotubes? elephants?

Unless you're going to simulate every possible atom/element, it's just not meaningful. And even then, with a sufficiently high processing speed, it still might take as long as the age of the universe.

[Maelstrom]

Originally Posted by aXis

All of this hinges on the hypothesis that you can make a smarter computer from a dumber one. Or at least make a smarter computer than a human.

So far we haven't demonstrated anywhere near that yet, and the best we may end up doing is simulate a human brain in hardware. It may be faster, but not more gifted or insightful.

We've already been able to use evolution concepts (survival of the fittest) to create more efficient code and hardware.

An example is this antenna:

http://www.spaceref.com/news/viewpr.html?pid=14394


For more on this sort of stuff: http://www.nasa.gov/centers/ames/res...lesystems.html

also: http://en.wikipedia.org/wiki/Evolutionary_algorithm and http://en.wikipedia.org/wiki/Genetic_algorithm

[Luke212]

using monte carlo to jump out of local minima is an example of computer imagination.

[aXis]

Yeah, I get your point, but I still think that's an optimisation process that's largely going to get us faster computers, not smarter ones. It does not truly think outside of the box - the generic algorithms produce novel results but still function inside of limited domain.


We still dont have anything even approaching a the AI algorithms that people talk about with singularities.

[Luke212]

Originally Posted by aXis

Yeah, I get your point, but I still think that's an optimisation process that's largely going to get us faster computers, not smarter ones. It does not truly think outside of the box - the generic algorithms produce novel results but still function inside of limited domain.

we can understand some environments (domain) well enough to set up the degrees of freedom. in these situations, computers can be smarter than a human.

if you dont understand the domain enough to setup degrees of freedom, the human may have an advantage.

We still dont have anything even approaching a the AI algorithms that people talk about with singularities.

we have the algorithm. the limiting factor is the environment. it takes a great deal of work to build an environment for the algorithm.

[Autti]

We currently can't create AI because computers are based on algorithms, the human brain is as well but its complexity is so large it won't be deciphered for a long time. If computers still being based on algorithms they would still act predictably as not be able to make "human" like decisions.