A NewScientist article summarizes the memristor revolution so far and predicts great things for AI as a result. To summarize their summary. Leon Chua mathematically predicted the existence of a fourth basic circuit component in addition to the resistor, capacitor, and inductor. He named this mythical component a memristor. It was similar to a resistor but "remembered" current. Memristors appeared not to exist, so Chua moved on to other research. 30 years later, HP Researchers stumbled onto a real memristor while trying to make low-power switches (Missing Memristor Found PDF format). The race was on. Memristance could revolutionize electronics. But here the story takes a detour in the world of intelligence. Physicist Max Di Ventra was studying P. polycephalum, a slime mould that puzzled researchers because it acted intelligently and learned without the benefit of neurons. He realized the slime mould acted as a memristor, confirming a suspicion Chua had that memristance could explain how organisms learn. It turns out memristors behave like neural synapses. Researchers are now working on hybrid transistor-memristor chips that will be able to reproduce some of the brain's processes. For more, try the HP Memristor FAQ.