The singularity?

[Sharptooh]

Is it imminent?

http://techcrunch.com/2011/02/16/ai-paradox/

[Shadow]

Watson is not AI, he is just a very well designed database. Intelligence requires randomness, and we can't (yet) produce true randomness with our computers.

[Gen. Volkov]

Actually Watson is just very well designed search engine. Yes, he has a massive database, but the real breakthrough was parsing human speech into something intelligible to computers. Still, as Shadow said, not AI. Also not proof the singularity has arrived.

[Sharptooh]

Actually Watson is just very well designed search engine.

Finally Google has competition  :D

Watson is not AI, he is just a very well designed database. Intelligence requires randomness, and we can't (yet) produce true randomness with our computers.

I fell for the media over-hyping something, I guess I should look into things more, also I didn't know true randomness within a computer was impossible, is it difficult to explain why?

[Shadow]

How do you make a computer choose a random number? Computers are deterministic - if you tell it to do something, it does the same thing every time. THe best you can do is come up with an algorithm that generates numbers that are even distributed across a range, a sequence of numbers. But any such sequence eventually has a pattern, just a very long one.

One major breakthrough with quantum computing will be true random number generation, since measurement of the quantum state of identically prepared systems produces unpredictable results, in a predictably probabilistic way. So you can get true randomness from measurement to measurement, while still knowing the statistical spread of your results.

[windhound]

Software randomness is basically impossible and probably will remain impossible.  Computers can only do what you tell them to do, 'guess' is not in their vocabulary.  There are very good random number generators, they use things like the current time (to a very small value) to generate a number.

Hardware randomness, on the other hand, exists. 
http://www.youtube.com/watch?v=7n8LNxGbZbs  is an example.  The website it belongs to is currently down, might be back up later.  But there's a neat writeup on it, basically the the dice ride on a roller coaster, 'round and 'round, a camera photographs them as they pass and an image processor determines the value of the dice.  The dice tumble in such a way that their values are not predictable, and so give a true random dice roll.

There's also machines that generate random numbers based on the decay of radioactive materials - http://www.fourmilab.ch/hotbits/

Generally, software random numbers are 'good enough' though. 

Anyways, Watson is an interesting bit of research, I'm glad to see IBM is still putting money in Research and Development.  So many other companies have stopped due to the benefits of such R&D being 5-10 years out, not within the next quarter.

[Gen. Volkov]

Anyways, Watson is an interesting bit of research, I'm glad to see IBM is still putting money in Research and Development.  So many other companies have stopped due to the benefits of such R&D being 5-10 years out, not within the next quarter.

I don't think IBM will ever really stop putting money into R&D. It's what made them what they are today.

[windhound]

http://blog.reddit.com/2011/02/ibm-watson-research-team-answers-your.html
IBM did a Q&A for reddit, neat read.