Several pieces of tech coverage came across our desk in the last week that we thought important enough to share.
Light-Based Memory for Purely Photonic Computing Systems
The first was from the journal Nature, the premier scientific publication in Britain. Interested readers can find the abstract of the paper here (Integrated all-photonic non-volatile multi-level memory). Researchers from the Karlsruhe Institute of Technology (KIT) in Germany, and Oxford and Exeter Universities in the U.K., succeeded in creating the first “integrated all-photonic non-volatile multi-level memory” for computing. We’ll break that down and explain why it’s significant.
Readers are likely familiar with optical fibers, which transmit data by means of light rather than electrical signals. The use of optics in data transmission is widespread because it allows many more data to be transmitted much more quickly. There’s a problem, however: the interface between optical and electrically based components in a system. That interface leads to what’s known as the “von Neumann bottleneck”—a limitation to the system’s speed that arises from the need to transfer data from memory to the processor. If memory could be stored optically, rather than electrically, there would be no need to “translate” between the two systems, and processing speed could be dramatically increased.
Also, until now, optical data have been volatile—they’re lost when the power goes off. The team from KIT, Oxford, and Exeter created a purely optical memory system that is non-volatile, so even when unpowered it will retain the data for decades. They exploited the physical characteristics of a well-studied metal alloy comprised of germanium (Ge), antimony(Sb), and tellurium (Te), “GST” for short. When subjected to high-energy light, this alloy switches back and forth between being a structured crystal, and being amorphous—that is, its phase state changes. Then when low-energy light is passed through it, the system can detect how much light is absorbed, and “knows” whether it is in a crystalline or amorphous state. That’s the equivalent of a conventional memory circuit being in the “on” or “off” position—a “0” or a “1” in the language spoken by computer processors.
The system may turn out to be far superior to traditional electrical memory storage, since it allows the possibility of using materials that will take on several different crystalline structures in response to different levels of light-pulse intensity—storing many bits of data in a single nanoscale element, rather than just one. (That’s why the paper’s title refers to “multi-level” memory.) In fact, one of the researchers in an interview gave his view that this technology will eventually allow neurally modeled computing systems, mimicking the structure of the human brain, because of their potential speed and complexity.
Google, NASA Lend More Clout to Quantum Computing
On the subject of data storage and processing, we’ve written in these pages before about quantum computing, which aims to exploit some of the stranger aspects of quantum mechanics to produce computers that are orders of magnitude more powerful than current technology allows. In quantum mechanics, particles can be “superposed”—which means that they are not in one state or another, but can be in many states simultaneously. In theory, quantum computing would be powerful enough, for example, to instantly break and disrupt the most robust encryption technologies currently known. Bitcoin advocates are therefore dreading the prospect of its arrival, because it would definitively wreck the technology that underlies the “crypto-currency.”
So far, the only company producing a quantum computer is privately-held D-Wave Systems out of Canada. It’s a controversial company, having sustained criticism from technical analysts who don’t believe that quantum computing is actually occurring in its devices, or that they show any superiority to conventional computer systems.
Now, however, Alphabet (the company formerly known as Google, NASDAQ: GOOG) and NASA disagree. They’ve spent the past two years working together with the latest D-Wave system at the Ames Research Center. And they just extended their contract with D-Wave for another seven years, giving them access to new generations of the technology as they become available. The company’s management says they are looking forward to meeting the challenge of outperforming GOOG’s own data centers, and observes that this is a significant step to overcoming the quantum computing skeptics.
Certainly, GOOG’s engineers are top shelf, so the news made us sit up and notice.
Vindication for Nikola Tesla
We read an article in the Wall Street Journal (you can find it here: Soon, Power Will Be Delivered to Your Device by Air) describing an event that has become typical over the past generation: a technological prophecy from long ago becoming a reality in the present.
This one refers to the turn-of-the-century Serbian-American inventor Nikola Tesla, most famous for inventing alternating current, triumphing over Edison’s advocacy for direct current in household electrics, and probably saving many lives as a result. After this, though, Tesla pursued more outlandish technological innovations that didn’t bear fruit in his lifetime—he died a pauper. One was a tower that would use the Earth’s magnetic field as a means of communication over vast distances. That tower was never completed, and was dismantled and sold for scrap before Tesla’s death.
The tower had another purpose—broadcasting electricity wirelessly over great distances, obviating the need for wires and batteries. A century later, that looks like it’s actually starting to become a reality.
Many readers have cell phones which they can charge by laying them on a mat—no need to plug them in. As the Journal article describes, several private companies are exploring technology that will take that a step further—with devices that will be able to “trickle charge” any battery in the room. The main advances, according to the article, have come because devices have become more and more miserly in their use of power. Now a 2-watt power drip from an electricity broadcaster in your living room could charge your phone in a few hours.
Since most people spend the bulk of their time at home and work, this will mean that you’ll never have to worry about forgetting to charge your device. It will also potentially be transformative for the Internet of Things (IoT), where countless tiny, simple computing devices need power. It could open a world of peel-and-stick thermostats and cameras, not to mention all the other IoT advances in supply chain management and automation that we’ve often discussed in these pages.
Investment implications: Some of these technologies are far from commercialization. As usual, our approach will be to monitor the scientists and institutions involved as they proceed towards actual implementation in real-world environments. We also monitor privately-held firms in areas of particular interest, so that when they plan to come public, they and their technologies are already familiar to us. These innovations also provide a note of macro optimism that the pace of technological innovation is not slackening.
For more commentary or information on Guild Investment Management, please go to guildinvestment.com.