Artificial Intelligence

meepm00pmeep

macrumors 6502
Original poster
Sep 20, 2006
279
0
Toronto
How long do you think technology will take to finally create A.I that can think for itself... and do you think this is a good idea?
 

MultiM

macrumors 6502
May 9, 2006
451
4
TO. I've moved!
Not long from now, if someone hasn't already. I can't believe that a REAL intelligence is actually running FOX news. Anybody with any real intelligence couldn't, could they?
 

sushi

Moderator emeritus
Jul 19, 2002
15,658
3
キャンプスワ&#
meepm00pmeep said:
How long do you think technology will take to finally create A.I that can think for itself... and do you think this is a good idea?
Limited AI is already here.

To the level where it is like an individual, I would say that we are many years from that.
 

XNine

macrumors 68040
I seem to remember scientists creating a jumping lamp (no, not the one from Pixar) that would learn it's limitations and what it could/couldn't achieve, and adjust it's jumping according to that. So every time it jumped and fell over or what have you, the lamp would learn that it shouldn't do that because the results were negative.

As long as they don't call a main computer SKYNET, I think we'll all be okay...
 

beatsme

macrumors 65816
Oct 6, 2005
1,204
1
meepm00pmeep said:
How long do you think technology will take to finally create A.I that can think for itself... and do you think this is a good idea?
I dunno if you've seen this, but I did and it really made me think. I'm wondering just how far away we are from usable AI.
 

patrick0brien

macrumors 68040
Oct 24, 2002
3,238
0
The West Loop
I feel we'll see true - self aware A.I. - e.g. human mind in-a-box when Quantum computing gets to an X number of qubits.

As quantum computing was being explored, it was discovered that a computer based around quantum computing only would not be particularly useful. One still needs the backbone of binary computing we have today (and coincidentally, the brain is a form of binary as well), therefore quantum would be a GPU-style coprocessor on the main board.

As this idea was being explored, somebody discovered the human mind had bundles of nerve cells in the hippocampus(IIRC) that act exactly like quantum computers. Putting two and two together, if one were to take a binary-based neural network and place a quantum computing at its core, may just yield consciousness.

This, of course is theory, and it is unknown at what amount of qubits would be required to acheive the proper level. We do know, however, that we are approaching brain parity with the binary, memory, and storage side of the house however.

My guess..? I think we'll see the first real A.I. in about 15 years.

Then things get interesting...
 

yellow

Moderator emeritus
Oct 21, 2003
16,033
1
Portland, OR
Well, I heard the G5 PowerBook was going to include an AI ROM, and that's the REAL reason that it never came out..
 

whooleytoo

macrumors 604
Aug 2, 2002
6,560
628
Cork, Ireland.
It depends on what you mean by "AI". If you mean something that can simulate human behaviour, it's difficult to say as there are varying degrees of "human intelligence" to match!

If you mean create a sentient, self-aware consciousness, I reckon never; unless we start delving into organic computing.
 

Sdashiki

macrumors 68040
Aug 11, 2005
3,511
8
Behind the lens
Its very hard to imagine a computer being humanlike....

I have seen a docu where a "specialist" said:
"I hate the brain is like a computer analogy. Because, for instance, we can teach a computer to read. It can read the letter E, for example. And report back, it sees the letter E. Whereas a 3yr old child can see an upside down E, a graffiti style E, an Old Gothic English E, and still understand that it is the letter E in all those forms. A computer can not do this."
imagine all the possibilities of the human experience JUST like that example, and figure out how to make a binary computer "think".
 

wimic

macrumors regular
Aug 24, 2006
235
0
calgary, alberta
Terrible Idea IMO

I think it's a terrible idea... and I think that it's something that humans are going to eventually accomplish.

The only thing that separates human beings from robots (which are essentially a more efficient form of being) is our capacity to feel emotion and make decisions.

Think about it - technology has advanced so far as to be able to create machines to replace workers (making many jobs obsolete), it has enabled humans to function with virtually no effort (dishwashers, etc.) and it has even been used to simulate human reproductive organs (ie. an incubator).

I think that technology is great until it starts in intrude on what it means to be human being and undermine it. To duplicate the wonderment a human being, in my opinion, is essentially saying that life is simply the organization of atoms merely oriented in a certain way. We're selling ourselves short and taking for granted how miraculous life really is.

If technology is able to produce a robot that can take in all the input in needs, process every variable and output a decision, that would romotely resemble that of a real human being - then what distinguishes us from the robots? If a robot was able to feel - then why are we so special? I think we should stop while we're ahead.

Who's to say that these supposed "robot-humans" wouldn't DECIDE to overtake us as a human race... what then?
 

thedude110

macrumors 68020
Jun 13, 2005
2,478
2
I pretty much gave up on humanity on May 11, 2005 when we introduced the "breakthrough" of self-replicating robots.

I'm not even going to save for my retirement. Just going to wind up a slave to the robots, anyway ...
 

n-abounds

macrumors 6502a
Mar 6, 2006
562
0
thedude110 said:
I pretty much gave up on humanity on May 11, 2005 when we introduced the "breakthrough" of self-replicating robots.

I'm not even going to save for my retirement. Just going to wind up a slave to the robots, anyway ...
'Twas the beginning of the end for humanity. I'm sure that some Terminator-type person will time-travel back to that instant and destroy the lab that created self-replicating robots. Although I guess that would've already happened.
 

atszyman

macrumors 68020
Sep 16, 2003
2,437
1
The Dallas 'burbs
n-abounds said:
'Twas the beginning of the end for humanity. I'm sure that some Terminator-type person will time-travel back to that instant and destroy the lab that created self-replicating robots. Although I guess that would've already happened.
It happened in 2003....

Unfortunately in order for Dr. Alfred Flannigan to invent the time machine in 2037 requires the break through work of Dr. Murray Oblink which was published in 2017 and based on the work of Dr. Fritz Hanso's research into the self replicating robots done between 2003 and 2004. Unfortunatly Dr. Hanso's life was tragically cut short in 2004 when his cat peed on his magsafe connector of his MacBook Pro burning down his house. Thus the time machine will not be invented and we're screwed....

or not...
 

clayj

macrumors 604
Jan 14, 2005
7,474
181
visiting from downstream
sushi said:
You can always buy a rice cooker that uses some sort of AI to cook the rice for around $700-$1,000 over here! :D
That's really more of a primitive form of expert system... a system that really, REALLY understands a very specific process (like cooking rice or navigating an automobile), but knows NOTHING about anything else.

I think true AI will be much longer than 15 years in the making... the CYC guys down in Austin are doing pretty well at growing an AI, but it turns out that programming common sense and general knowledge into a computer is a lot of work.

Even when one is built, I don't think it'll be anything like Skynet... more likely there will be big restrictions on what an AI is allowed to connect to (if anything; see Greg Bear's books for how AIs would be disconnected from the world so as to prevent them from running amok, or William Gibson's Turing organization from Neuromancer).
 

sushi

Moderator emeritus
Jul 19, 2002
15,658
3
キャンプスワ&#
clayj said:
That's really more of a primitive form of expert system... a system that really, REALLY understands a very specific process (like cooking rice or navigating an automobile), but knows NOTHING about anything else.

I think true AI will be much longer than 15 years in the making... the CYC guys down in Austin are doing pretty well at growing an AI, but it turns out that programming common sense and general knowledge into a computer is a lot of work.

Even when one is built, I don't think it'll be anything like Skynet... more likely there will be big restrictions on what an AI is allowed to connect to (if anything; see Greg Bear's books for how AIs would be disconnected from the world so as to prevent them from running amok, or William Gibson's Turing organization from Neuromancer).
Yep, the rice cooker is a very limited example for sure. Hard to believe how expensive they are! Yikes!

I agree with you that it will be a long time before we see true AI.

As I stated earlier:
Limited AI is already here.

To the level where it is like an individual, I would say that we are many years from that.
How many years is anybodies guess. My guess is that the limitation is on the software/logic side rather than the processor/memory side. But they do go both hand in hand.

Not being in that field, I have not idea how many years we are talking about. Back in my college days, a classmate of mine emulated a small set of nerves with a computer. It was pretty cool. As he felt, to get true AI we will have to somehow emulate the true functions of our nervous system and brain. Not an easy task for sure. Exciting to think about.
 

meepm00pmeep

macrumors 6502
Original poster
Sep 20, 2006
279
0
Toronto
i think creating an intellegence that can process data and think/act for itself is a big mistake that mankind will inevitably realize once machines (organic or mechanic) realize that we are obsolete and un-necessary...

Reality is yesterdays science fiction
...
 

Metal Hippy

macrumors newbie
Feb 9, 2010
5
0
STARFLEET PERSONNEL FILE: Data


Final Rank:
Lieutenant Commander

Last assignment:
Second Officer/Science Officer, U.S.S. Enterprise NCC-1701-E (was promoted to First Officer, but died prior to position taking effect)

Full Name:
Data

Date of birth:
Permanently re-activated Feb. 2, 2338 (initial activation unknown)

Place of birth:
Omicron Theta science colony

Parents:
Created by Dr. Noonien Soong and Dr. Juliana O'Donnell Soong Tainer

Education:
Starfleet Academy, 2341-45

Marital status:
Single

Children:
One (deceased)

Quarters:
Formerly, Enterprise-D: Deck 2/Room 3653

Date of death:
2379

Place of death:
Destroyed with Reman ship Scimitar in line of duty

Service Awards:
  • Starfleet Command Decoration for Valor
  • Starfleet Command Decoration for Gallantry
  • Medal of Honor (with Cluster)
  • Legion of Honor
  • The Star Cross



Additional notes - Specifications

Weight:
100 kilograms including the following:
24.6 kilograms of tripolymer composites
11.8 kilograms of molybdenum-cobalt alloys
1.3 kilograms of bioplast sheeting
Upper spinal support: stress-resistant polyalloy
Skull: cortenide / duranium​

Storage:
88.8 petabytes (93 million gigabytes)

CPU Speed:
Total linear computational speed rated at 60 teraFLOPS. However, through bidirectional sequencing (which compensated for signal degradation) and conversion of his main interlink sequencer to asynchronous operation, Data's computational speed was thereafter no longer limited by the physical separation of his positronic links and, thus, became unlimited.

Source: Google​

Yep, I'm bored! :D

[Edit] Wow - I just realised how old this thread is! :eek:
 

localoid

macrumors 68020
Feb 20, 2007
2,447
1,724
America's Third World
Ray Kurzweil: The h+ Interview

Topics include: Consciousness and quantum computing, purposeful complexity, reverse engineering the brain, AI and AGI, GNR technologies and global warming, utopianism and happiness, his upcoming movies, and science fiction.

How Long Till Human-Level AI?

When will human-level AIs finally arrive? We don’t mean the narrow-AI software that already runs our trading systems, video games, battlebots and fraud detection systems. Those are great as far as they go, but when will we have really intelligent systems like C3PO, R2D2 and even beyond? When will be have Artificial General Intelligences (AGIs) we can talk to? Ones as smart as we are, or smarter?

Self Tracking: The Quantified Life is Worth Living

What would you do with a complete memory of your entire life? Would you relive your first kiss? Figure out what triggered your recent migraine? Remember the name that goes with the familiar face in front of you?

In other words, wouldn’t it be great to have a backup of your brain?
 

Gelfin

macrumors 68020
Sep 18, 2001
2,166
4
Denver, CO
That's really more of a primitive form of expert system... a system that really, REALLY understands a very specific process (like cooking rice or navigating an automobile), but knows NOTHING about anything else.
Anybody want some toast?