The Computer Revolution Keeps Reinventing Itself As the Possibilities of Progress Continue to Increase
By BRAD GRAVES
Originally published May 12, 2016 at 2:40 p.m., updated May 12, 2016 at 2:40 p.m.
San Diego — With each passing year, computers get more capable. The rapid development of the technology continues to shape business — all business, even those far from the heart of the technology sector. The San Diego Business Journal recently approached several people in the region’s business community with a simple question: Could your core product of today work just as effectively if all you had at your disposal were the computing resources of five or 10 years ago?
Or even 20 years ago?
Almost always, the answer was no.
Their answers offer glimpses into just how far computer capabilities have come in the last few years.
2011 is as distant as Pluto. 1996 might as well be light years away.
Olin Hyde credits the easy availability of powerful cloud computing resources for making his seven-person business possible.
Five years ago he would have needed a supercomputer to do what he does, said Hyde, the CEO of LeadCrunch. Today, “we can scale as if we were IBM,” thanks to vendors such as Amazon Web Services Inc.
LeadCrunch, a business in residence at EvoNexus’ downtown incubator, generates B2B sales leads for small and medium-sized businesses. Hyde says his software can take a list of a client’s 25 best customers, review billions of pieces of information about U.S. businesses on the internet, and prepare a list of similar companies that the client’s sales department can approach with confidence. Hyde’s business began life under the name Englue.
LeadCrunch has grown, Hyde said, with the help of $100,000 worth of donated services from Amazon.
With such technology, “a very small team can do something that five years ago required a lot more people,” he said.
The business also leverages recent gains in artificial intelligence, which is advancing rapidly. “The state of the art five years ago is almost irrelevant today,” Hyde said of the artificial intelligence field.
Even a company working in a very old craft — shipbuilding — feels the rapid advance of computer technology.
General Dynamics NASSCO builds U.S. Navy and commercial ships at its yard in Barrio Logan. It produces construction drawings of those ships on software from Cambridge, England-based AVEVA Group PLC.
Bill Hale, the company’s senior director for engineering, helped implement the shipyard’s first three-dimensional product modeling software in 1995.
Such software produces 3-D computer models of ships. The advantages of building a virtual model include alerting designers to places where systems interfere with each other. Such warnings eliminate errors in the field. The software also breaks the huge process of building a 600-foot-long ship into manageable subassemblies. “You want to create very tailored instructions for the worker in the field,” Hale said, comparing the instructions to the four-page assembly manual you get when buying a backyard barbecue at a big-box retailer.
The software has improved greatly since the mid-1990s. Hale said the user interface is much better, and it runs on store-bought desktops rather than dedicated computers. Today’s system handles big files much better than its predecessors. The software also ties in better with the rest of NASSCO’s ERP (enterprise resource planning) software, informing other business functions such as the ordering of materials.
NASSCO’s engineering department has 240 people and is located in landlocked Mission Valley.
Qualcomm Inc. recently announced that it has paired up its Snapdragon 820 processor chip with artificial intelligence, extending artificial intelligence to smartphones, automobiles, drones and security cameras.
In a May 2 announcement, Qualcomm (Nasdaq: QCOM) said mobile devices equipped with its top-of-the-line Snapdragon chip and special software will be capable of deep learning without being connected to the cloud. Those devices will be capable of tasks such as scene detection, object tracking and avoidance, face recognition, text recognition, translation and natural language processing.
Deep learning might help a driver send voice commands to a car, said Gary Brotman, director of product management for subsidiary Qualcomm Technologies Inc. It might work with an automobile’s computer to discern whether a driver is dozing off, and sound an alarm. It might also cut down on false alarms triggered by security systems, teaching a computer the difference between a stalking cat, a swaying tree branch and a suspicious person casing a house.
Qualcomm has also adapted machine learning to identify malicious software.
What is driving machine learning to the mobile device? Computer hardware has advanced markedly during the last five years, Brotman said. Also contributing are advances in computer code and more data available for training models.
Qualcomm said it plans to roll out its Snapdragon Neural Processing Engine software development kit in the second half of 2016.
Edico Genome Inc. is part of the Torrey Pines Mesa biotech and biomedical cluster. The business created a microchip called Dragen, the world’s first bioinformatics processor designed to analyze next-generation gene-sequencing data.
Executive Gavin Stone said Edico Genome could not work on yesterday’s computers. In fact, it can’t even work on today’s. It had to go out and invent its own technology.
The marketing vice president compares conventional CPUs (central processing units) to minivans. They are good general-purpose vehicles, but not optimized for specific tasks. Edico Genome, he continued, makes chips adapted to genomics work, much as Formula 1 cars are adapted to racing.
The business uses the latest FPGA chip — which Stone described as a “blank slate” — as a foundation for its product, which is a board with its processor and other computer electronics. The initials FPGA stand for field programmable gate array.
Other businesses have put FPGA chips to work in specialties such as Bitcoin mining and internet search. FPGA technology is just now going mainstream, Stone said.
RippleNami Inc. offers a simple, cost-effective way to visualize data on maps, such as those offered by Google. The San Diego business relies on well-established technology, said CEO Jaye Connolly-LaBelle. Its programming language goes back to the 1990s, as do the concepts of internet connectivity and cloud storage.
RippleNami wants to grow in emerging markets, including Africa. To put roots down in Africa, however, it had to wait for a more recent technology event: the mass adoption of smartphones. Smartphones, rather than desktops, are the computers of choice in Africa, said Connolly-LaBelle. There were only 25 million users in 2001. By 2012, smartphones had reached critical mass with some 650 million Africans taking them up. Now more than 1 billion people use them.
In March, RippleNami inked deals with two African companies: one in financial services and the other in transportation. Both businesses plan to use RippleNami to offer customers certain location-specific data, such as aviation tracking information, weather and situational awareness news (information that may affect travel, for example). RippleNami also opened an office in Nairobi, Kenya with an eye toward generating more business.
Remember those tape players that visitors used to rent at museums? They gave patrons in-depth information about whatever exhibit they stood in front of. Guru provides a 21st century version of that service, sending location-based information to patrons’ smartphones rather than rented tape players.
In fact, Guru does them one better. People walking around San Diego’s Little Italy district with a Guru guide can hold a smartphone camera up to certain landmarks (Washington School, for example) and see historic pictures of the place. The technology is called augmented reality.
Image recognition technology that makes augmented reality possible is about 10 years old, said Guru CEO Paul Burke. Bluetooth proximity beacons as well as the rise of the “app store” concept are other recent events that make Guru possible.
“We couldn’t be doing what we’re doing without a lot of technologies coming together at once,” Burke said.