Um, that's utter crap.
I beg to differ. clock for clock, AMD blows Intel out of the water. Clocks-per-instruction (CPI) is a useful metric when it comes to comparing CPU horsepower across differing architectures, and AMD is the clear winner when you do. In an understandable decision, Intel listened to their marketers, who told them, "we need higher clocks than our competitors, because that's what people want to buy." So Intel chose long instruction pipes so they could get higher clock frequencies, and then focussed on minimizing the inherent penalties (like having to flush said long pipeline on a branch instruction.) Intel's branch prediction algorithms are excellent, but they can't erase the branch penalty caused by those long instruction pipelines. That penalty is why an AMD running at half the clock speed of an Intel processor (and therefore costing 1/2 to 1/3 as much!) does as much or more actual computing (CPI) than the Intel.