We shouldn't fear our future robot overlords, we should welcome them. They're not as corruptable as the human ones. So long as we create them properly, we have nothing to fear, and a whole lot to gain.

The problems of this world are bigger than we can hold in our minds all at once. We just need to build bigger minds.

We'll be fine.

It's difficult to compare the human brain to a binary computer. We honestly don't know how it works yet. So we need to phrase the question as how much power would we need to simulate a human brain. The number according to many would roughly come to around 90 Peta FLOPS. For those who don't know the progression is Mega, Giga, Tera and Peta. I won't bore you with enormous numbers so in short 1 Peta FLOPS would be in rounded numbers 1.126 quadrillion FLOPS where 1 Giga FLOPS would be rounded to a mere 1.074 Billion FLOPS. That means to simulate human in computer math you need 101.34 quadrillion FLOPS to simulate a human brain. So by these numbers the average human brain is around 3500 times more powerful than my PC. This makes my PC 5 times more powerful than a politician.

Remember this I highly speculative because we don't know how the human brain works. We can only plug in what we know it does and see what it takes a computer to perform that task. Today the most powerful computer in the world is sitting in China. It's called the Tinehe-2 and it is running at around 34 Peta FLOPS or about 1/3 that of what's required to simulate a human brain. So in theory we could have a desktop computer that has that kind of power in less than 20 years.

The power isn't the only factor the memory and storage and then the big one a program that can throw AI into mix may take a little longer.

Having witnessed the evolution of computers over the years from Super computers down to the PC I can't rule out the possibility that we will have computers that we will call friends and have meaningful relationships with in a few years. We may even have androids one day that we will keep for companionship. I'd frankly be surprised if we didn't have them. As to their capabilities that will be determined by the programmers that will constantly be updating them. It's doubtful that one person will write all of the code for AI. Once it starts I'm sure that the lawyers will spill from every Crack and crevice to attack and defend the research. Time will tell. In short I can't confirm that our computers will become smarter than we are but I can't rule it out either. The reason I can't rule it out is because by the standards of the original PC my current is impossible. They never thought the PC would become what it is today. In the 60s a TV show called Star Trek gave birth to the flip phone. Not many people use them anymore because the smart phone is both a triquarter and communicator…

A truly self aware piece of software is qualitatively different from pac-man. The simple fact that our hardware is improving in no way indicates that we'll be dealing with sentient machines soon. Thinking that a quantiative change (faster computers) will lead to a qualitative change (intelligent machines) simply does not make sense to me.

Not saying we'll never have intelligent machines. Just saying that when we will have intelligent machines, its because we've adopted a fundamentally different design approach, not because our computers are really, really, really fast.