Marc Andreessen's ears had to be burning, as industry analysts sought to unravel the implications of the big Uber self-driving car event in Pittsburgh on Wednesday. Software may not be eating the world, as the developer of Mosaic and co-founder of Netscape famously asserted in 2011, but it's taken a big bite of the car, and not just those of the driverless variety.
"There is a hilarious conceit in the industry," said Gartner research director Mike Ramsey. Every time an auto executive talks about the amount of software in their cars the number gets bigger. "I've heard 25 million lines of code, 50 million lines of code, 100 million lines of code, 150 million lines of code. I recently heard 200 million lines of code."
Indeed, the transformative, disruptive power of software and information technology was invoked at every turn this week as the industry mulled over what to make of Uber's offer to allow loyal customers to take a spin in one of its experimental autonomous cars: a modified Ford Fusion hybrid sedan loaded with software, tricked out with nifty hardware (e.g., 20 cameras, seven lasers, including a spinning lidar on the roof, two radars) and accompanied by a test driver to prevent accidents.
"One of the things that is hard to analyze is how far along Uber really is," in developing a road-ready self-driving car, said Ramsey. But that is hardly the point, he and others said. Hiring talent from Carnegie Mellon's robotics department, Uber launched its program to build self-driving cars a little more than a year ago in answer to what it perceived as a threat to its business model from Google -- generally acknowledged as the leader in self-driving cars. A year later, it "is not afraid to put people in their vehicle and give them the experience of what it will be like," he said.
Are we staring down the barrel of a Silicon Valley/technology company-induced disruption of the car industry? Will Uber do to Detroit what Amazon did to retail stores, smartphones to Kodak? And as Uber, Google, Tesla and the traditional car companies (Ford, Volvo and BMW have autonomous programs) compete ferociously to perfect the technology that will make self-driving vehicles road-ready, what are the downstream effects?
"This is one of the biggest industries on the planet, and it has a whole ecosystem that is going to change," said Eric Brynjolfsson, the Schussel Family Professor at the MIT Sloan School of Management and director of the MIT Initiative on the Digital Economy. "I don't think people think broadly enough about how self-driving cars will affect everything from car ownership to the nature of transportation in cities to supply chains and logistics."
Not to mention, our relationship to the car. In an era of Uber self-driving vehicles and automobiles as a service, the car in the driveway may become less a symbol of freedom than an expensive depreciating "motorized sculpture," said Frank Gillett, a principal analyst in the CIO group at Forrester Research. He foresees the day when autonomous transportation is overlaid with personalized services, including when to “bike to work for my health and fitness" instead of summoning a self-driving car, or where along his commute to buy the best mocha cappuccino.
"If you think you've seen change, you ain't seen nothin' yet. This is like sitting in 1995 at the dawn of the web and consumer access to the internet. We've got decades of unraveling what it means and how it changes how we work, and what we buy," Gillett said.
In the meantime, CIOs can be sure of a couple of trends that will have an impact on their industries over the next five years, Ramsey said. First, cars -- from hybrids to full electric cars -- will become more electrified, relying more on battery power; more of their functions will be controlled by motors and drive-by wire than by physical mechanical connections. Second, while cars are already loaded with processors, the processing power is generally built into individual components -- the windshield wiper actuator and braking system -- rather than controlled by a central computer, but that will change. "Cars are going to have more centralized brains and more automation," Ramsey said.
The two developments mean a "huge amount of new software in cars," Ramsey said -- and a big opportunity in the immediate future for the micro-processing players. "As in a lot more sales to the graphical processing units for Nvidia and Intel and NXP -- all the chip makers are already salivating over the potential," Ramsey said. But generating revenue from a vehicle whose electrical systems are connected to the Internet all the time will take many forms, he added, citing the jump in business opportunities with the advent of smartphones. "When you had a flip phone there were limited numbers of people who could make money off the phone, but when the phone could offer applications for users and GPS data and could take pictures, new businesses popped up."
As the capabilities of autonomous cars mature, CIOs can also expect that expensive technologies -- millimeter wave radar and lidar for topographical mapping -- will drop in cost and be incorporated into consumer products, including phones.
Who will own the autonomous future?
As for who will own the autonomous future, few are willing to place bets. The leader in self-driving cars -- Google -- has indicated it is not interested in building cars, said Forrester's Gillett, who sees it more likely selling its autonomous driving capabilities as a new utility.
"This is the central tension point right now," said Ramsey. He, for one, does not predict an overnight disruption of the car industry. While tech companies have an advantage in IT "talent, experience and agility," it remains "super hard" to build a car, and even harder to build one extremely well. Can Google or Apple make a car? No question, although they would be better off buying a car company "with their spare change," Ramsey said. The last major successful debut of a U.S. car company was Chrysler in 1925.
Tesla, the poster child for proving that Silicon Valley can build a car company, is also testament to how difficult it is to do that "even with a super strong brand." Founded in 2003, Tesla produced 50,000 cars last year -- its record, he said. "It will produce around 80,000 cars this year," or one-tenth of 1% of the global auto market.
"The idea of a single player [disrupting] the market on a global basis is kind of crazy," Ramsey said.
MIT's Brynjolfsson is not so sure. "I'd be careful," he said, noting Tesla's highly rated vehicles. "I wouldn't be complacent if I were a traditional automobile manufacturer. I'd keep my eyes open in how startups could use new technology to get pieces of that market."
Speaking of keeping eyes open, Forrester's Gillett made note of Uber's purchase a few weeks ago of Otto, a robotics company that's focused on improving efficiency of long-haul trucking. "Their premise is really interesting, which is, 'Let's forget about how to drive autonomously in the cities -- that is a hard problem. Let's try to figure out how to help large trucks, which have a higher platform on which to put sensors, drive on the highway so truck drivers can rest,'" he said. "They went and solved a simpler problem that has higher value."
CIO news roundup for week of Sept. 12
The Uber self-driving car event wasn't the only tech news this week. Here's what else made headlines:
HP acquires printer business. HP Inc. is buying Samsung Electronics' printer business for $1.05 billion, HP announced Monday. This acquisition positions HP to "disrupt and reinvent the $55 billion copier industry, a segment that hasn’t innovated in decades," according to a press release. The deal is subject to regulatory approval and is expected to close within a year. In other HP news, the company has entered a six-year agreement with Microsoft wherein HP's 6,500 salespeople and 20,000 service employees worldwide will be using Microsoft Dynamics' customer relationship management (CRM) software. "This brings us a cloud-based solution that delivers a more effective and efficient collaboration engine across our business,” HP chief operating officer Jon Flaxman said in a statement. Meanwhile, Hewlett Packard Enterprise (HPE) announced that it will extend its partnership with Arista Networks to deliver secure hybrid IT solutions and will also increase its focus on software-defined networking architecture, Sci-Tech Today reported.
Hybrid cloud dominating the enterprise. Organizations are switching to hybrid cloud models faster than predicted, according to a study conducted by the IBM Institute for Business Value. The study, titled Tailoring hybrid cloud: Designing the right mix for innovation, efficiency and growth, surveyed over 1,000 C-suite executives from various industries. Fifty-four percent of respondents stated they are adopting hybrid clouds to lower cost of ownership, 42% said it helps them boost innovation and operational efficiencies, and 40% said it helps them meet customer expectations. Adopting a hybrid cloud strategy has also helped organizations expand into new industries, generate new revenue streams and create new business models, the study found.
Galaxy Note 7 fiasco. Samsung is planning a software update to prevent Note 7 batteries from overheating by allowing it to charge only up to 60% of its capacity. Earlier this month, the world's largest smartphone maker announced it was recalling more than 2.5 million Note 7 smartphones worldwide after it received several complaints about the devices overheating and even exploding while charging. The recall, due to faulty lithium-ion batteries, came within two weeks of the company launching its latest top-of-the-line smartphone. The U.S. formally recalled the device Thursday, a day before the iPhone 7 went on sale. Authorities worldwide have advised users against using Note 7 devices on planes, trains and buses.
Reading closed books. Researchers from MIT and Georgia Tech have teamed up to devise an imaging system that can read closed books. They have designed a prototype imaging system that uses terahertz radiation, an electromagnetic radiation between the microwave and infrared spectrums. Researchers tested the prototype on a stack of papers that each had one letter printed on it, and the system was able to correctly identify the letters on the top nine sheets, according to a press release. Researchers said the system could be used to read antique books that are too delicate to open.
Assistant editor Mekhala Roy contributed to this week's news roundup.