I just finished reading Humans Need Not Apply: A Guide to Wealth and Work in the Age of Artificial Intelligence a fascinating book, written in a very lively manner. It raises some issues that we will likely face in the near future (and some potential solutions). An excellent review is available in the Economist.
Then I saw this CES2015 demonstration where a mass-market manufacturer is approaching robotic vision and scene modeling, using a Tegra K1’s 192-core GPU. What was really striking was that at the same CES show, NVIDIA announced their Tegra X1 mobile chip which achieves 1 Teraflop performance. It was compared to the ASCI Red, the world's fastest supercomputer until late 2000. Excluding the many issues of 16 vs 64 bit FP, architecture and memory, what is fascinating is that the ASCI Red consumed 850 kW and occupied 150 m2 of space, versus the 10W, thumbnail size of the X1.
The pieces keep falling ... maybe they will self-assemble into place.
