Surprisingly, there were lots of interesting tech pieces this week, though no theme emerged.
The Internet was in a virtual frenzy over developments, or rather lack of developments, with Apple’s mobile payment system, which I believe is doomed to failure. There were a number of developments in networking and computing technology as well as further signs of “blow back” associated with the Snowdon/NSA revelations.
Google’s Secretive DeepMind Startup Unveils a “Neural Turing Machine”
Neural networks are based on a simplified model of the function of a brain. Not a “human brain” per se because brains work pretty much the same way across species. Brains are very good at certain things (consider the complex behavior of a mosquito, despite a microscopic brain) such as pattern recognition and adapting to varying conditions, but not very good at other things like repeatability and math. This appears to be a novel type of neural network, however, it is not clear that such a design is useful for any real-world application.
“One of the great challenges of neuroscience is to understand the short-term working memory in the human brain. At the same time, computer scientists would dearly love to reproduce the same kind of memory in silico. Today, Google’s secretive DeepMind startup, which it bought for $400 million earlier this year, unveils a prototype computer that attempts to mimic some of the properties of the human brain’s short-term working memory. The new computer is a type of neural network that has been adapted to work with an external memory. The result is a computer that learns as it stores memories and can later retrieve them to perform logical tasks beyond those it has been trained to do.”
The Internet jungle drums beat mightily all this week after this news broke, prompting an unrelenting slew of increasingly hysterical defenses of Apple’s offering as well as condemnation (complete with allegations of conspiracy) of a rival technology. Frankly, it is hard to believe so many people give a damn which payment system, if any is successful. Such is the might of marketing in the Internet age. For the record, I believe the odds of Apple’s system being successful are next to zero: their market share is too low, especially among the ranks of, say, Walmart shoppers. The most likely successful payment system, assuming one emerges, is bound to not be tied to any smartphone OS or vendor.
“The fight for control of the mobile payments market is opening a rift between merchants and banks. Banks and credit card companies have enthusiastically supported Apple Pay, seeing it as a way to increase the number of purchases people make with their credit cards. But Apple has struggled to get merchants on board. A quick look at Apple’s website explaining the service highlights just 34 retail partners that support the system. Eight of those are different flavors of Foot Locker. One is Apple itself.”
This article is an excerpt. You can subscribe to the full version of The Geek’s Reading List here.
Big Factories Won’t Solve High Cost of Electric Vehicles, Carnegie Mellon Researchers Say
There is a popular misconception that battery prices have been plummeted, aided and abetted by the cult following around Tesla, jabbering idiots on Wall Street, and the traditional wishful thinking among the “green” energy crowd. Batteries are very simple things: containers with metal and chemicals inside. They are not significantly more difficult to manufacture than cans of soft drinks and, as a consequence, the process is highly automated and continuous. Similarly, the materials are industrial commodities made in enormous volumes and not amenable to significant improvements to the cost of production. Nonetheless, the expectation persists that an enormous factory will somehow make a difference – at least among those who know nothing about batteries.
“”Electric vehicle batteries are expensive,” Michalek says. “Federal and state governments have been subsidizing and mandating electric vehicle sales for years with the idea that increasing production volume will reduce costs and make these vehicles viable for mainstream consumers.” Tesla’s planned Gigafactory has a similar hope, promising major cost reductions at higher volume. “But we found that battery economies of scale are exhausted quickly, at around 200-300 MWh of annual production. That’s comparable to the amount of batteries produced for the Nissan Leaf or the Chevy Volt last year,” Michalek said. “Past this point, higher volume alone won’t do much to cut cost.””
This announcement made headlines, and it is pretty impressive. Plus, the multi-core approach seems to be the sort of thing which might be expanded and probably should be relatively straightforward to manufacture. However, the problem is not that fiber isn’t fast enough but that it doesn’t go to enough places, mostly because it is expensive to run the cable. So expensive, in fact, that cables often contain a multiplicity of individual fibers, even if most are left “dark”. So, what we really need is an incentive to string more fiber-optic cables to more places, not faster fibers.
“Researchers at Eindhoven University of Technology (TU/e) in the Netherlands and the University of Central Florida (CREOL), report in the journal Nature Photonics the successful transmission of a record high 255 Terabits/s over a new type of fiber allowing 21 times more bandwidth than currently available in communication networks. This new type of fiber could be an answer to mitigating the impending optical transmission capacity crunch caused by the increasing bandwidth demand.”