reply to post by grinitagi
It depends on what you mean by "true" AI. It depends on how you define intelligence. We can already create neural networks. We have learning
machines like Watson. Search engines like Google can figure out what context you're looking for.
Ad networks track you, learn about you, and figure out what to advertise to you based on what it knows about you. There's software that can translate
what you say into any language. Police departments have computers that can predict when and where crimes are most likely to happen next.
There's software you can show an image or video and it will bring up other pictures that are similar to it. Even MP3 audio files or image compression
can be a form of AI, depending on what AI means to you. They can make files smaller by taking out sounds you can't hear and colors you can't see.
One of the most common forms of AI is one, that if it's working correctly, you'd never even notice. Email spam filters that look at billions of
emails a day and learns what is spam, and what isn't spam by reading them so they can protect your inbox from being flooded. But while they read the
email, they still don't understand what they're reading. Just matching symbols with other symbols.
We can even simulate the complete neural connections of small animals like a fruit fly or a cat. Not at the same speed, but all the connections are
there. It's just a matter of advancing.
We have AI all around us. It just doesn't work like the movies. There's only so many CPUs in the box. So, coders make the best trade offs they can
for each problem.
We already have "AI", the question is just subjective and depends on how intelligent it has to be before you consider it intelligent. I mean,
what's true AI? When it's smarter than a human? When it becomes self aware? If so we're not there yet, and may never get there. It just depends on
how you think about it.
The human brain is pretty complex and has trillions of connections and it doesn't work like computers do. Computers work in digital, very exact yes
or no type connections. The human brain is a little more fuzzy. Analog chemical connections that just need to connect.
A computer sacrifices speed and power for the sake of giving an exact yes or no answer to every connection. In other words, the computer is designed
to be correct first, then as fast as possible. This is because they were designed to do math, and the math has to be correct. Humans are the other way
around.
The human brain is designed to be fast by making billions of connections all at the same time, even if some of the answers are wrong. And compared to
a computer humans are extremely bad at math. But really good at making batcrazy fast decisions, even if they're really dumb decisions.
The human brain saves a lot of energy doing this. To simulate all the connections of a human brain with digital computer equipment at the same speed
as the brain, it is estimated that it would require an entire power plant to supply the needed electricity. And I know I don't have a nuclear power
plant in my skull.
Some people are trying out more analog type equipment that doesn't have to be as exact. It just makes a bunch of really fast really crappy
connections and takes an average. IBM is working on a simulation on a chip type thing that works like a human brain, but much much slower. If it'll
catch up in speed is anyone's guess.