> To change topic, progress on memristors seems painfully slow.
I think that's a little unfair. They were first shown just 7 years ago, and have already been used to build (extremely) simple neural networks.
Transistors took ~6 years before they were sold in commercial products, and you can already buy a memristor on a chip.
When transistors were invented, ENIAC was around and had ~18,000 vacuum tubes, and was the kind of thing governments built. In 2008 you could wander into a shop and pick up something like this: http://ark.intel.com/products/37147 which has 730 million transistors.
I think it's understandable that it'll take a bit longer for memristors to overtake transistor based processors.
I think that's a little unfair. They were first shown just 7 years ago, and have already been used to build (extremely) simple neural networks.
Transistors took ~6 years before they were sold in commercial products, and you can already buy a memristor on a chip.
When transistors were invented, ENIAC was around and had ~18,000 vacuum tubes, and was the kind of thing governments built. In 2008 you could wander into a shop and pick up something like this: http://ark.intel.com/products/37147 which has 730 million transistors.
I think it's understandable that it'll take a bit longer for memristors to overtake transistor based processors.