Singularity is gona get U all Humans !

http://news.pacificnews.org/news/view_article.html?article_id=575

 

Log in or Sign up to hide all adverts.

Even the best computers have NO intelligence or conciousness.

 

Log in or Sign up to hide all adverts.

There is a great book called "The Singularity Is Near" by Ray Kurzweil that I suggest every person read.

I worry far more about the "Grey Goo" effect that many theorize about.

~String

 

Log in or Sign up to hide all adverts.

And here's that Vinge article in full:

http://mindstalk.net/vinge/vinge-sing.html

So yeah, um, we're doomed. South Korea, Japan and others are feverishly drafting robot ethics charters but they won't be worth the paper they're written on when Skynet assumes control. The end will be painful but mercifully fast.

 

And it was u who voted for the last option, right ?

:roflmao:

 

computers are only as smart as the people who prgamed them

 

U mean if they were programmed, right ?

http://adaptiveai.com/research/index.htm



BTW i seriously doubt if u really made 110+ posts, did u hack SciForums ?

 

No, that would have been me. I'm highly doubtful but one can only hope. The silicon Gods may save us yet:

Please Register or Log in to view the hidden image!

http://en.wikipedia.org/wiki/Transhumanism​

 

From human point of view, sooth your self.

From my point of view, humans are pests or a Cancer of this planet.

 

In the next post, #7, Singularity gave a reference and pointed to one non-programmed approach to AI. (It was too long to read, but in my skim of it, I saw no mention another type, already in widespread use*, not just a "paper design.") I refer to what I, and some others, prefer to call a "connection machines," but many more call them: "neural networks."

About 20 years ago one was trained on records of loan applications at a bank. (Including whether or not the previously granted loans were repaid on schedule.) It is now better than the human staff in selecting which new loans to grant. (Humans still interview the applicant of course, but then feed the answers to the machine and it decides to grant, or not, the loan. The applicants good looks, low cut dress, etc. do not enter into the decision.) It, like all three or more layer "connection machines," was / is never programmed, only given a “learning set” of input information closely related to the type of problem to be solved.

You need to broaden your concept of what a computer is. It need not follow a Von Neuman architecture with software running a written program.

Another "singularity" that can wipe out both humans and robot high on AI is a real singularity, like the small black hole of my book, if it passes close to our solar system. I.e. Earth may be expelled from the HZ (habital zone).
--------------------------
*Many complex chemical process plants, like a pulp to paper mill, are better controlled by connection machines than humans as there are just too many variables and the interaction of them is not understood (by humans)** well enough to program or even process "intutitively," as done until recently by a seasoned / experienced "old timer," who might even taste the batch in an effort to adjust things.

**It is your choice - do you want to say the connection machine "understands" better? If not, are you sure any human "understands" anything - after all the brain is just the best, most-adaptive, connection machine in existance (on Earth at least.)

 

Close - but not close enough.

Please Register or Log in to view the hidden image!

Even though you claim it wasn't "programmed", that "learning set" was most certainly a set of parameters - which is just a form of programming. A machine can sort and correlate data but unless it's told what to do with the results nothing will come of it. And that's certainly not an example of Artificial Intelligence.

Please Register or Log in to view the hidden image!

There are several marks of true intelligence, one is being to handle a situation/input that it has never encountered before. Another is having an original thought. So far, no machine has ever done either of those and I honestly doubt if one ever will.

 

Never say never: http://news.bbc.co.uk/2/hi/technology/6600965.stm

 

"Genetic and evolutionary programming do have their uses - they are powerful tools that can be used to solve very specific problems, such as optimization of large sets of variables; however they generally are not appropriate for creating large systems of infrastructures. Artificially evolving general intelligence directly seems particularly problematic because there is no known function measuring such capability along a single continuum - and absent such direction, evolution doesn't know what to optimize. One approach to deal with this problem is to try to coax intelligence out of a complex ecology of competing agents - essentially replaying natural evolution."

So does this scare u ?
http://msdn.microsoft.com/msdnmag/issues/04/08/GeneticAlgorithms/default.aspx:mufc:

 

That article sounds like - they think to create a car , they must make the human like legs first

Please Register or Log in to view the hidden image!

.

 

oops, i meant to say anonymous.

 

Either (1) you think humans, H, are also "programmed" or (2) you do not understand how connection machines, CM, work.

Both Hs and CMs lean from their experiences, but both must have some external guide to correct them when their response is not acceptable / correct. When informed that their response was not correct, they both make internal modification in the "connections." (For Hs this a change in the brains synaptic connections. For CMs, this is the transfer weights for one "layer" to the next.) Neither H nor CMs are "programmed" as that term is usually used. Both learn from their mistakes and improve their performance on not only the training task set but only all never before seen similar sets. For example, separating photos of women from men, loan re-payers from dead beats, etc.

I attended a lecture by Terry Sonowski's on his "nettalk" CM about 25 years ago. It learned to read out loud (actually how to drive a TI voice synthesizer) when present with strings of letters (actually the ASCI codes for them). It was fascinating to hear it at various stages of its learning process. - Just like a child learning to speak. For example, at first, it only said "ma ma" or "da da" (constant+ vowel a) in response to every input string of letters. Later it went thru a period of "over regularizing" the past tense, just like a child, and said things like "he drowned," (instead of "he drown"), but later leaned correctly most of these exceptions.

Just like most of a child's learning, it never was given even one line of "program code" - not any instruction was ever given to it. (Children do get instructions also.) Both Hs and CMs are given an intital set of connections for them to modify to improve their performance and some systematic way to make and evaluate the effect of the changes they make, mainly trial and error. (Both tend to follow the rule "If it ain't broke, don't fix it." In the CM case, this means that when an error occurs on the learning set, the change in connection strengths is greatest in the connections that, in that case, were most influential in producing the erroneous result. - Perhaps the new connections will work better, perhaps not, but they keep trying till they are better on all the learning set of examples.)

Your choice:
Either both Hs and CMs are “programmed” or neither is. - Both Hs and CM learn operationally by the same "trial and be correct when wrong" way.* - You are really just discussing the definition of "programmed." As I use this word, it means giving a set of instruction to be followed. What do you mean by it?
--------------------------
*For Hs, the mother is a source of some corrections, but physics (nature) supplies more. - E.g Walk wrong - fall down. Touch hot thing -feel pain. Etc.

 

Computers taking over the world? I disagree.

I have it on excellent authority that IBM, Dell, Mac (yrrch), Toshiba and indeed all other computer companies have been for years equipping their products with an inherent and unbeatable failsafe device to prevent the conquest of the planet via sentient silicon beings.

This device in question is termed a "plug". It is located on the back of the computer. And it works like thi

 

No, I didnt vote. Computers have no consciousness, will, or even the ability to think.

Vastly complex calculation yes, but that doesnt qualify as intelligence.

 

"direction" (which is alternative word for "programming") is NOT required, either by humans, dogs, even amoebas, and connection machines. All that is required to "shape" their behavior is feedback that tells them when their response is "wrong" and when it is "correct" plus a set of trials on which to learn. (Using mild electric shock to tell it when it was wrong, and amoeba can learn to move either toward or from a light etc.)

I admit that at present the "learning set" is very restricted to one or two areas in most cases, I think. (not up to speed- all I post here is from memory at least 20+ years old) Tasks such as recognizing men from women, printed letters, different words, car made by companies (Fords from Toyotas, etc.), etc. where the sample space is small and well defined, but see no reasons, in principle, to say "programming" (or 'direction") is required. Humans, many other (even very simple) animals and Connection machines all learn without any. All they require is feedback telling them when wrong and some means of changing their "input to output transfer function" until they perform better on that sort of task, which in principle, perhaps 500 years form now, could be "behave just as the typical human would under all circumstances."

BTW, your idea that evolution requires "direction" {from God, the first programmer?} is entirely wrong. The whole point of Darwin was that "direction" is NOT required. What evolution requires is (1) some means (genes in the case of Earth's life forms) to hand down learned changes from one generation to the next and (2) some way to select (for transfer to the next generation with better than average probability) the set of these means (genes on Earth), which tend to produce the "correct" response (or eliminate the "error responders" before they can reproduce)

 

Humans are doomed for now, since out of 12 only 4 think they are in danger.

YAAAHHOOOOOO !