Mathematics in the brain and computers

A while back I heard a talk on how the brain calculates. The main idea from the talk was that the brain uses a different function to do accurate arythmatic as opposed to conceptual math and other types of computations.
For arythmatic the brain will access the information it needs from memory and apply it like an algorythm. This is a typical left brain complexity function. And this is the model used to build computers, and doing more of these faster and faster brings us to the computing power we have now.

But when the brain does other calculations, like when catching a ball, it uses estimations. This function is more like guessing and then narrowing in the get the approximate answer. With these functions the brain realises it does not need the exact answer, so its "computations" are only as accurate as it needs to be.

So, this got me thinking. Why are we modeling robots and certain applications to act with the memory and algorythm approach when we dont need a dead accurate answer (just yet). So I was wondering if a kind of fuzzy math is being applied to robotics and other applications, to sell accuracy for speed in the settings that allow for it.

Perhaps this is what they are aready doing. I often come up with ideas that have been used already - but they were authentic to me. Well, let me go try do some reading on this one.