Some algorithms are smarter. The computer system that figures out how to book flights for millions of people every day is quite intelligent when it comes to that specific task.

The problem is, that intelligence is more like an insect's brain in that it doesn't generalize to any other higher level problems. We're still trying to figure out how to organize knowledge so the intelligence is transferable. Doesn't seem too difficult to do in the next 30 years given currently available tools.

I think the point OP is making is that the underlying technology and "thihnking process" behind most AI hasn't changed much in the last 50-60 years. We look at a task like figuring out how to book millions of flights and think "That seems like it would be very difficult for me to do, therefore a computer must be intelligent if it can do it". But that isn't true for all definitions of 'intelligent'. The basic optimization process, once figured out (by humans) is fundamentally the same and is just being applied by the computer at greater scales and speeds. If computers were getting smarter, for some definitions of the term 'smarter', they would be identifying problems to be solved on their own, or at least they would be coming up with novel solutions without any human involvement.

A pen and paper can be used to make powerful calculations, but we don't generally attribute intelligence to the tools themselves. The promise of artificial intelligence was that computers would eventually become capable of taking independent action and doing things that humans couldn't predict or do themselves. To say that we're still trying to figure out how to do that is a bit of an understatement. The time frame for fulfilling that expectation is completely unknowable, at best.

have you heard about deep learning? Computers are coming up with novel solutions without human involvement. It was done before in the 70s and 80s with neural networks, but it only worked on simple tasks. Now the main limits seem to be computing power (and learning how to train it).

Ah yes. Lucy's infamous "in the next 30 years" moving goal post.

When it comes to CPU cycles or training data, as in so many other things, quantity has a quality all of its own[1].

1] http://www.quora.com/Who-said-Quantity-has-a-quality-all-its...