I guess around 15 to 20 years.

I guess around 100-250 years.

They already are. Have you never heard about Boris Spasky and the IBM Chess playing computer. It beats him nearly every time.

What do you mean how long?

It's hard to determine because all computers to date operate in such a different way to the human brain that they're not easy to compare. I would expect however that within 50 years they will be able to function as a very unimaginative human.

When humans begin being able to only add and then only with ones and zeros.

I think they might be smarter, right now. They don't have politicians. They don't borrow money and not pay it back. They don't kill each other "for sport." They don't have a tax structure. Oh, to be a machine! [Please take note of the sarcasm.] +5

My advisor says "Ask again later"

Machines are made by human them how cames tey are smarter....?????

They are smarter they dont make decisions based on emotion.

The most likely answer is never. It is not a matter of time or technology. Its a matter of neurology. Humans can reflect, and use experience and common sense. computers can only hold vast amount of information and query results based on probabilities and a predefined set of rules. Computers have more answers but have no intelligence what so ever. They will never write an original story/song/painting or come up with a new invention.

'Smarter' is unfortunately rather vague term. You would have to define 'smartness' which is, however, not easy at all. There are intellectual tasks where machines already outperform most (if not all) humans. Playing chess, solving many optimization problems, and so forth. On the other hand, there are tasks which human brain can solve with efficiency and/or speed unmatched by any machine - image/sound processing and analysis, symbolic language processing, real-time motoric control. Therefore, human brain needs to be viewed as an extremely efficient but at the same time extremely specialized tool. In other words, it does what it does very very well but at the same time it fails badly doing things it has not been 'designed for'. This includes tasks that are ridiculously easy from computational point of view, like solving a system of linear equations. To sum it up, contemporary machines already possess a great deal of intelligence yet it is intelligence of a very different kind. Comparison is thus basically ruled out. Due to this, giving any timeframe on when these two completely different kinds of intellingence will be (or perhaps were?) 'on-par' is impossible. Let's hope we'll manage to get the most of both without losing who we are in the future :).