Is Artificial Intelligence Even Possible?

He's used it on a different number of occasions in the past and as result, even though I knew the word, I felt I had to refrain from using it. (You know how touchy he can get.) Still it's a great word and regardless of whether or not it actually applies when he uses it, I always get a kick out of it.

Now that one I had to look up, and it too is a great word. May I use it in the future?

Only if you promise to use it ironically.

I'll try

It is a fine word. I found this quote:

The Turing test for artificial intelligence is whether we were able to tell if we were talking to a man or a machine. I think a machine without consciousness would never be able to fool an intelligent human interrogator.

What is being proposed is technological singularity. It relies on a machine becoming capable of: Quote is from Wikipedia
Keep a supply of dynamite on hand. Would that work?

I'll worry about the 'singularity' when machines pass the far easier Turing test.

In order to pass the singularity test (able to make improved versions of itself, etc.) it would have to be self aware would it not? That would pretty much be 'consciousness'.

Not feeling the need to put an IED under my PC.

"Singularity" is an unfortunate and overused word and as inappropriate in AI as it is in Cosmology. You needn't worry about anything that achieves the point of singularity because at that point it ceases to exist. If a black hole or the universe ever achieved singularity, they too would cease to have mass, gravity, rotational momentum, etc.
"Convergence" is in most cases more useful and accurate. As potentially mutually-reinforcing technologies merge they become more powerful as they are synergistic; unless of course they converge to the point of singularity.

I'm not bound to accept the Turing test. The problem throughout this thread has been a lack of consensus on what constitutes artificial intelligence. The rovers on Mars have artificial intelligence right now. With anywhere from a three minute to a 22 minute lag in radio signals from Mars to the earth, and those times doubled for query and response, the rovers would be constantly immobilized if they were not able to make path-finding decisions on their own. Consciousness be damned--this is not a theological discussion. The rovers on Mars right now are able to make their own decisions without consulting humans. You may not want to consider that to be artificial intelligence. You may not be well advised to risk your hard-earned cash by making intelligent investments.

But are the Mars rovers able to make decisions in circumstances that have not been anticipated by the programmers?

The answer is no, they are not.

The Mars rover path finding algorithms are well known. Nasa writes about them, they are based on what is called the Field D* algorithm. You can google it and read about it yourself. We use similar algorithms to guide robots on warehouse floors.

This is just a turn the crank mathematical process that the robot goes through given a goal set by humans. The robots aren't really making any decisions on their own while path finding.

Did you miss the part about "artificial?" Do you understand why it is called artificial intelligence? The holy rollers in this thread have been wringing their hands and gnashing their teeth at the thought that someone would create a machine which can think for itself independently, and therefore trivialize their creation fairy tales. That's not what artificial intelligence means. If you teach a child to read and write, and the child goes on to be a successful novelist, do you claim it doesn't count because the child was "programmed?"

Artificial intelligence is not just possible, it's in use right now. Get over it.

This is bullshit (and a little condescending since its purpose is just to take a cheap shot at other people on the thread).

There is a fundamental difference between what human children do as intelligence, and what current technology does as "artificial intelligence".

Humans are creative. They can figure out new ways to use knowledge creating solutions to problems that no one has encountered before. They can create their own goals and formulate new ideas.

Current technology doesn't do this. The algorithms used today by things like the Mars robots aren't creating any new solutions, or innovating or solving any types of problems that haven't already been solved by humans.

The current generation of "artificial intelligence" just follows a set of mathematical steps that humans have set up for it. There are a couple of technologies that make some claim to "learn" (e.g. neural networks and genetic algorithms), but even these aren't even the slightest bit related to how human intelligence works.

I believe that we will get to the point where software can truly replicate the creativity and innovation of human minds, and the ability to learn.. not just by tweaking networks... but by actually adapting on its own to solve new classes problems.

No one is successfully doing anything like this yet.

Please tell me Watson is reading from a script in his commercials and not ad-libbing.

-Get in a pissing contest with someone else, Max . . . go beat up a women or something. You apparently also don't get the part about artificial.

If you don't post bullshit, I won't have anything to piss on. You compared the current generation of artificial intelligence to a human child. That is nonsense. The fact that you did so as part of a childish slap at people you consider beneath you is the reason I pushed back.

You like to bully people who don't see things your way... but you can't take it back.

If you don't want a pissing contest.... let's see if you can refrain from responding.

I don't like bullying.

Watson is not. The software I work with is not even close to the level of power or sophistication of Watson.

Most current artificial intelligence (i.e. pattern recognition, search engines or speech recognition) works primarily with statistical models. The software is not trying to understand the input, the way a human would. Instead it is using statistical models to generate the response the human wants.

There is a difference between semantic understanding and statistical modeling. Human actually ascribe meaning to each piece of information we take in, or that we synthesize into a response.

When computers successfully gain the ability to process information with semantic meaning on a large scale, it will be a step toward true intelligence.

I don't know all the research that is going into Watson (and similar cutting edge projects). I do know that the artificial intelligence that is currently in commercial use deals with statistical models rather than semantic understanding.

... even so, the technology is clearly useful and often quite impressive.

Intelligence is more than powerful memory and speed of calculation. It includes creativity. All the creativity in the Mars rover was put there by the natural intelligence of the minds who designed its programming.

One personal slur after another, Mr. Respectful Discussion. Par for the course with you.