IBM’s new NeuroSynaptic chip was recently announced and is making a big PR splash in the print and online media (more here and here). The chip’s processing power is based on the parallel implementation of essentially identical ‘neuronal’ structures that mimic major sections of the brain’s neocortex. We reported on one such promising structure – hierarchical temporal memory - developed some years back by Jeff Hawkins (On Intelligence). The IBM chip achieves higher transistor density levels, but most remarkably has an architecture that consumes miniscule power compared to standard chips with equal numbers of transistors. The claim for the NeuroSynaptic chip is that we are now a skosh or two away from truly intelligent machines in the form factor of a human brain that don't require a dedicated utility to supply its electricity. I will get real excited when IBM starts reporting on how easily a brain from such a chipset can learn to do everyday things in the real world.
Common Core is the latest jumbo sized Community Kool-aid served in Nevada County. Opening this morning's (9aug14) Union, we see multiple articles (here and here) and commentaries (here) on the joys of the new federal education standards. RR readers long ago learned how the heavy hand of federal education funding programs now mandate the adoption of this fast track to a deeper national educational deficit. The naifs still claim that CC is the best thing since sliced bread for our kids who will now learn less at a slower rate, and be forced into college catch-up courses when they leave high school. The remediation of CC kids will be required for many majors before they can start the prescribed four-year course of study. But hey, education will then be much more expensive, employing additional legions of registered Democrats, who will take more classroom time to teach the little darlins to think correct thoughts. All of it part and parcel of the ongoing fundamental transformation. (For more, search RR for ‘Common Core’)
Data scientists are highly paid and in short supply. Data scientist is the new label for a class of STEM workers who are skilled in handling massive databases, searching them for patterns, and developing algorithms to use such pattern data to generate information that is used for making decisions in all kinds of fields from medicine, through energy production, to investing and finance. We learn that STEM degrees in data intensive fields of science and engineering can get a twenty-something upwards of $300K a year for openers. (more here)
Over the recent months (years?) there has been much discussion here (search RR for ‘STEM’) about the availability of STEM workers, the number of STEM jobs, importing STEM workers, and so on. According to my lights, STEM fields and skill sets are now so numerous and diverse that not all areas are equally likely to ‘guarantee’ employment to the graduate. It pays to do some considerable research and get in on the ground floor in a promising STEM field during school. And when in doubt, take lots of courses in applied math, physics, algorithmics (computer science), and data handling.
Romney's return in 2016. I felt kinda lonely coming to the conclucion that Mitt might not be a bad candidate again in 2016 given the absolute disaster that Barack Hussein Obama has made of the Office of the President. I have reported the recent writings of American pundits on that possibility. Now we are starting to hear people overseas looking back at the big mistake our electorate made (again) in 2012. Here is an interesting report from a very popular British outlet.