TORONTO — As an undergraduate at Cambridge University, Geoffrey Everest Hinton thought a lot about the brain. He wanted to better understand how it worked but was frustrated that no field of study — from physiology and psychology to physics and chemistry — offered real answers.
So he set about building his own computer models to mimic the brain’s process.
“People just thought I was crazy,” said Dr. Hinton, now 69, a Google fellow who is also a professor emeritus of computer science at the University of Toronto.
He wasn’t. He became one of the world’s foremost authorities on artificial intelligence, designing software that imitates how the brain is believed to work. At the same time, Dr. Hinton, who left academia in the United States in part as a personal protest against military funding of research, has helped make Canada a high-tech hotbed.
Dictate a text on your smartphone, search for a photo on Google or, in the not too distant future, ride in a self-driving car, and you will be using technology based partly on Dr. Hinton’s ideas.
His impact on artificial intelligence research has been so deep that some people in the field talk about the “six degrees of Geoffrey Hinton” the way college students once referred to Kevin Bacon’s uncanny connections to so many Hollywood movies.
Dr. Hinton’s students and associates are now leading lights of artificial intelligence research at Apple, Facebook, Google and Uber, and run artificial intelligence programs at the University of Montreal and OpenAI, a nonprofit research company.
“Geoff, at a time when A.I. was in the wilderness, toiled away at building the field and because of his personality, attracted people who then dispersed,” said Ilse Treurnicht, chief executive of Toronto’s MaRS Discovery District, an innovation center that will soon house the Vector Institute, Toronto’s new public-private artificial intelligence research institute where Dr. Hinton will be chief scientific adviser.
Dr. Hinton also recently set up a Toronto branch of Google Brain, the company’s artificial intelligence research project. His tiny office there is not the grand space filled with gadgets and awards that one might expect for a man at the leading edge of the most transformative field of science today. There isn’t even a chair. Because of damaged vertebrae, he stands up to work and lies down to ride in a car, stretched out on the back seat.
“I sat down in 2005,” said Dr. Hinton, a tall man, with uncombed silvering hair and hooded eyes the color of the North Sea.
Dr. Hinton started out under a constellation of brilliant scientific stars. He was born in the United Kingdom and grew up in Bristol, where his father was a professor of entomology and an authority on beetles. He is the great-great-grandson of George Boole, the father of Boolean logic.
His middle name comes from another illustrious relative, George Everest, who surveyed India and…
Experts from Yale and Oxford University recently released their research on how AI will transform modern life by reshaping transportation, finance, health, science, and the military. Key findings pointed toward a 50 percent chance that AI will outperform humans in every job in 45 years’ time.
How can we anticipate and manage trends in AI, and will creative and analytical skills really be the key to job success in an AI world?
Recounting a personal scenario, my boys had a “future day” at school. One dressed up as a chief robot maker and the other as a soccer player — which left me wondering, which career would be right in a world of artificial intelligence?
When the internet came of age, I would tell anyone who asked, “Don’t do a job that can be put down a wire” — meaning, don’t train to do a job that someone on the other side of the world can do for half the price. But with AI, it’s no longer that simple, according to the survey.
This shift is worrying for two reasons: It’s going to happen relatively quickly, and most schools are ill equipped to handle it. But before exploring the future of work, let’s dispel some myths and look at what won’t happen.
Over- and underestimating the effects of AI
During a recent panel discussion at the Financial Conduct Authority, time and resistance were both flagged as key challenges to the effective deployment and adoption of AI technologies.
Resistance — or “AI friction” — is a particular challenge within the older demographic, where people may play with the tech but find ways not to use it because deep down they prefer human interaction.
Driverless cars and deliveries are another example of AI friction. That’s a job threat for three million people in the U.S. alone, as indicated in a report published in December 2016 by the White House that looked at the ways AI will transform the U.S. economy. The industry has scrutinized this and broken down the evolution of autonomous cars into phases. The initial phase is what we see now on the roads, which is essentially an advanced cruise control, and the final phase is a completely autonomous car.
The friction in this case occurs around phase three or four, where a human and a computer both drive the car. In this phase, when the computer does not know what to do, it hands control over to the human. On the surface this sounds logical; however, when simulated, it is revealed to be intensely dangerous. Imagine being in a driverless car and you have “switched off” from the task of driving — then a car pulls out and the AI hands over control. Few drivers will snap to attention quickly enough to avoid a collision.
The progression and extensive testing of these vehicles presents a problem that cannot be easily fixed, meaning there’ll be a gap between the smart cruise control phase and the introduction of a truly autonomous vehicle.
The AI friction in this scenario also relates to the three million people in the supply chain who could be replaced by AI. To understand if this impact is realistic, we need to consider the logistics. Most logistics models start with the inbound goods arriving on a train or ship. This is called a multi-modal shipment because it’s full of standardized containers that can be lifted directly from the ship onto the back of a truck, thereby changing the mode of shipment. Autonomous vehicles are not really going to affect this first leg.
The shipment is then typically long-hauled from the port to a warehouse or distribution center. These…
In Amazon’s warehouses, there is a beehive of activity, and robots are increasingly doing more of the work. In less than five years, they will load self-driving trucks that transport goods to local distribution centers where drones will make last-mile deliveries.
Soon afterward, autonomous cars will begin to take the wheel from taxi drivers; artificial intelligence will exceed the ability of human doctors to understand complex medical data; industrial robots will do manufacturing; and supermarkets won’t need human cashiers.
The majority of jobs that require human labor and intellectual capability are likely to disappear over the next decade and a half. There will be many new jobs created, but not for the people who have lost them — because they do not have those skills. And this will lead to major social disruption unless we develop sound policies to ease the transition.
The industry behind these advances — and reaping huge financial rewards from them — has been in denial. Tech entrepreneur Marc Andreessen, for example, calls the jobless future “a Luddite fallacy”; he insists people will be re-employed.
But now others, including Facebook’s Mark Zuckerberg, Tesla’s Elon Musk, and Bill Gates, are acknowledging a skills mismatch with the potential for mass unemployment. They advocate a Universal Basic Income (UBI), a payment by the government that provides for the basic wants and needs of the population.
But these tech moguls are simply kicking the can down the road and shifting responsibility to Washington. UBI will not solve the social problems that come from loss of people’s purpose in life and of the social stature and identity that jobs provide. And the politicians in Washington who are working to curtail basic benefits such as health care and food stamps plainly won’t consider the value of spending trillions on a new social-welfare scheme.
SAN FRANCISCO — Uber is spending millions of dollars to make self-driving cars on its ride-hailing network a reality. Now Lyft, one of Uber’s biggest competitors, is striking a series of partnerships to do the same.
Lyft on Tuesday announced an agreement with nuTonomy, a self-driving car start-up, that will eventually bring thousands of nuTonomy’s autonomous vehicles to Lyft’s ride-hailing network. The partnership will initially focus on research and development related to the customer experience of summoning an autonomous vehicle, Lyft said.
“Our ultimate responsibility is to bring the best autonomous vehicles to Lyft’s millions of passengers,” Logan Green, Lyft’s chief executive, said in an interview. “And since it’s very early in the development life cycle of autonomous vehicles, we’ll explore many partnerships to learn with and from partners to help figure out what passengers want.”
The agreement is part of Lyft’s broader move into autonomous car-sharing. Mr. Green has long postulated that the future of transportation will be less focused on private car ownership.
But Lyft is behind others in making inroads into that…
For all of the visions of robots taking over the world, stealing jobs and outpacing humans in every facet of existence, we haven’t seen many cases of AI drastically changing industries or even our day-to-day lives just yet. For this reason, media and AI deniers alike question whether true broad-scale AI even exists. Some go as far as to conclude that it doesn’t.
The answer is a bit more nuanced than that.
Current AI applications can be broken down into three loose categories: Transformative AI, DIY (Do It Yourself) AI, and Faux AI. The latter two are the most common and therefore tend to be what all AI is judged by.
The everyday AI applications we’ve seen most of so far are geared toward accessing and processing data for you, making suggestions based on it, and sometimes even executing very narrow tasks. Alexa turning on your your music, telling you what’s happening in your day, and how the weather is outside is a good example. Another is your iPhone predicting a phone number for a contact you don’t already have saved.
While these applications might not live up to the image of AI we have in our heads, it doesn’t mean they’re not AI. It just means they’re not all that life-changing.
The kind of AI that will “take over the world” — or at least, have the most dramatic effect on how people live and work — is what I think of as Transformative AI. Transformative AI turns data into insights and insights into instructions. Then, instead of simply delivering those instructions to the user so he or she can make more informed decisions, it gets to work, autonomously carrying out an entire complex process on its own, based on what it’s learned and continues to learn, along the way.
This type of AI isn’t yet ubiquitous. The most universally-known manifestation of it is likely the self-driving car. Self-driving cars are an accessible example of what it looks like for a machine to take in constantly-changing information, process and act on it, and thereby completely eliminate the need for human participation at any stage.
Driving is not a fixed process that is easily automated. (If it were, AI wouldn’t be necessary.) While there is indeed a finite set of actions involved in driving, the data set the AI must process shifts every single time the passenger gets into the car: road conditions, destination, route, oncoming and surrounding traffic, street lanes, street closures, proximity to neighboring vehicles, turning radiuses, a pedestrian stepping out in front of the car, and so on. The AI must be able to take all of this in, make a decision about it, and act on it right then and there, just like a human driver would.
This is Transformative AI, and we know it’s real because it’s already happening.
PITTSBURGH — When Uber picked this former Rust Belt town as the inaugural city for its driverless car experiment, Pittsburgh played the consummate host.
“You can either put up red tape or roll out the red carpet,” Bill Peduto, the mayor of Pittsburgh, said in September. “If you want to be a 21st-century laboratory for technology, you put out the carpet.”
Nine months later, Pittsburgh residents and officials say Uber has not lived up to its end of the bargain. Among Uber’s perceived transgressions: The company began charging for driverless rides that were initially pitched as free. It also withdrew support from Pittsburgh’s application for a $50 million federal grant to revamp transportation. And it has not created the jobs it proposed in a struggling neighborhood that houses its autonomous car testing track.
Blame is being pointed in many directions. While Mr. Peduto had trumpeted his relationship with Uber’s chief executive, Travis Kalanick, he didn’t get any commitments in writing about what the company would provide for Pittsburgh. That became an issue in Pittsburgh’s Democratic mayoral primary this month, with Mr. Peduto’s challengers criticizing his relationship with Uber and one calling the company a “stain” on the city. (Mr. Peduto won the primary.)
“This was an opportunity missed,” said Michael Lamb, Pittsburgh’s city controller, who has called on Uber to share the traffic data gathered by its autonomous vehicles.
The deteriorating relationship between Pittsburgh and Uber offers a cautionary tale, especially as other cities consider rolling out driverless car trials from Uber, Alphabet’s Waymo and others. Towns like Tempe, Ariz., have already emulated Pittsburgh and set themselves up as test areas for self-driving vehicles. Many municipalities see the experiments as an opportunity to remake their urban transportation systems and create a new tech economy.
Yet Pittsburgh shows the clash of private-versus-public interests that can result. The lessons are college course level “101,” said Linda Bailey, the executive director of the National Association of City Transportation Officials.
Uber “is a business, and they want to make money,” she said. “With Pittsburgh, we learned we need to present the city’s needs upfront.”
Uber said it was open to a deal with Pittsburgh but had yet to see a draft of proposed commitments the city is seeking from the company. Uber said it planned to share some data collected by its autonomous vehicles with the city this year, though Pittsburgh officials say the data Uber shares with other cities is insufficient.
The company, which still has allies in Pennsylvania’s state and county government, said it had created 675 jobs in the greater Pittsburgh area and had helped local organizations like a women’s shelter, among other moves.
“Uber is proud to have put Pittsburgh on the self-driving map, an effort that included creating hundreds of tech jobs and investing hundreds of millions of dollars,” the company said in a statement. “We hope to continue to have a positive presence in Pittsburgh by supporting the local economy and community.”
Pittsburgh’s frustrations with Uber are encapsulated in the Hazelwood neighborhood along the Monongahela River, where the company opened a driverless vehicle testing track last year. From the second floor of the neighboring Center of Life church, the track is in full…
But Intel knows that we’ll have to get data in and out of those cars at rates that are much faster than today’s LTE mobile networks can handle. And that’s why Rob Topol, general manager of Intel’s 5G business and technology, believes that 5G wireless networking will be like the “oxygen” for self-driving cars.
Intel is making 5G modem chips to transfer data at gigabits a second over wireless networks in the future, perhaps as early as 2020. Topol believes this wireless networking will enable self-driving cars to communicate with connected infrastructure. That infrastructure will help the cars process sensor, safety, and information for the car and return the results quickly to the cars.
“What we were showing with that demonstration is a capability called V-to-X – vehicle to infrastructure, vehicle to pedestrian, or vehicle to vehicle,” Topol said. “You’re utilizing the other objects and using a network to give the car vision beyond things it can’t see through the mechanisms in the car itself. Something that’s happening around the corner or further ahead.”
We talked about 5G and its connection to autonomous driving in a recent interview. Here’s an edited transcript of our talk.
Image Credit: Dean Takahashi
VB: Where are we on the timeline for 5G?
Rob Topol: As you’ve seen, there’s been lot of work going on in 5G over about the last year and a half. It started with developing the air interface, how it works between the device and the core access network. It starts with the trial specifications we’ve been doing, where you try out different things like modulation or channel coding, essentially building this new radio.
The timeline is that you do many of those trial specifications over about a year to two years. All that work you do in the field, testing with partners and building a recipe, you take them to the standards bodies and submit them as contributions. Intel and other companies will submit their ideas and say, “This is what we think the standard should be based on.” The voting happens in 3GPP from a cellular standpoint, with the first round later this year. In December, the New Radio, or NR specification, will be set, and the full release for 5G happens in Q3 of 2018.
Once the NR spec is set, that’s when you’ll start to see development around the modems, around the networking equipment to support that. Full release 15 is at greenlight. That’s when you see the Capex orders come in from network operators to the infrastructure companies. Intel puts its chipset designs into production. You start to see some operators roll out networks as early as 2019. You’ll probably see most do them around 2020. Typically, once the full release is done, it’s about 18 months until you start to see the networks deployed in a broader way.
VB: How do you help people get an appreciation for how important this is? What’s at stake? What will people get out of it?
Topol: We focus on the things that 5G is about that 4G was not, if that makes sense. 4G was the era of the smartphone — data proliferation, access to media, mobility in general, with something in your pocket. 5G is an era beyond the smartphone. Over the next five to 10 years, we have billions of connected things coming up all around us. As we make everything from a refrigerator to a car to a home to an enterprise network smarter, the compute that’s happening—more data is sent through a network, whether to help improve the service model or the user experience of those things, or to harvest that data for machine learning and data analytics, any sort of behavior or artificial intelligence work.
We look at 5G as the platform that helps all of those other verticals grow. We see some of the early use case research. We see a lot of promise for 5G in automotive. We see a lot of promise for smart home and enterprise, if you move networks more to a fixed wireless capability. Not just relying on fiber and other LAN connections. We’re also looking at industrial automation. We have a few projects in that space, helping get a lot of that incubation going. What does a connected factory mean? How does it operate? How would a factory benefit in productivity, in the way machines are set up and run and optimized, when the factory is connected? What capability does that bring?
How do we help people with that vision? We show as many of those use cases as early as we can. We go out and showcase with automotive companies, showing them 5G in a car, so they can see the way data comes into the car for the way it functions, for safety, and more important for bandwidth, the way our experience inside the car is going to change when it’s autonomous. When you’re sitting in the back seat of that car, your experience changes. You’re not required to be fully focused on where it’s going and what it’s doing. The bandwidth requirements in the car are going to grow exponentially as our time is freed up inside the vehicle.
We do the same thing in industrial automation. We set up a partnership with GE and Honeywell and Ericsson called the 5G Innovators Initiative. We’re blueprinting what we think the factory of the future could and should look like, how it would operate. We try to give people that vision. We’ll do that for how drones are used. We’ll do it for fixed wireless in home and enterprise. We’ll do it for media and viewing experience. We’re working with media companies to look at the way media is captured and how it’s transmitted and how it’s consumed. That’s another major focus area of that initiative.
Image Credit: Intel
VB: I’ve seen different kinds of reports about how big 5G is going to be as a contributor to the global economy. What do you think about that? Is this going to transform society?
Topol: It opens new business models. As you talk about a smarter city, or a more efficient factory, or a vehicle that can run autonomously, it does change our behaviors, our habits, and what we can do with our time. I’d agree that 5G will open up many new business opportunities, primarily because of the data that’s coming off these smart and connected machines all around us. The ability to analyze, compute, and harvest that data is going to make us smarter in the way we lay out cities, set up factories, drive cars, and do other things around us.
There’s a tremendous opportunity. I don’t have a specific study we’ve done yet. Our focus is to deliver the technology baseline and the use cases around it. We’re confident that our partners are going to build great business models around it.
Our strategy from the start, as we said at CES, is end-to-end capability. Intel is the only company that sells hardware solutions from the device all the…
SAN FRANCISCO — Uber has signaled that it might break up with the star engineer who led its driverless car efforts, as the company seeks to disentangle itself from a high-stakes lawsuit that could affect its future in self-driving vehicles.
To comply with a court order, Uber’s top lawyer told the engineer, Anthony Levandowski, in a letter that if he did not comply with the lawsuit, he could face “adverse employment action,” which may include the loss of his job, according to a court filing on Thursday.
“While we have respected your personal liberties, it is our view that the court’s order requires us to make these demands of you,” Salle Yoo, Uber’s general counsel, said in the letter. “We insist that you do everything in your power to assist us in complying with the order.”
The letter was the latest turn in a legal battle between two technology giants. Mr. Levandowski has been at the center of a trade secrets case involving Uber and Waymo, the self-driving car business that operates under Google’s parent company. Waymo has accused Uber of conspiring with Mr. Levandowki, a former Google engineer, to steal trade secrets from Google for Uber’s autonomous vehicle designs.
The letter is the first public sign of a split between Mr. Levandowski and Uber, which to date has not tried to force him to cooperate with the investigation in the case. Mr. Levandowski has been close to Uber’s chief executive, Travis Kalanick, who wooed the engineer to join the ride-hailing company and purchased Otto,…
(Reuters) — Qualcomm Inc said on Thursday it had demonstrated how electric vehicles could be charged wirelessly while driving, a technology some believe will help accelerate the adoption of self-driving cars.
The smartphone chipmaker said a so-called “dynamic charging” test took place on a test track in Versailles, France. It used two Renault Kangoo vehicles driving over embedded pads in the road that transferred a charge to the cars’ batteries at up to 20 kilowatts at highway speeds.
Experts believe that self-driving cars of tomorrow will be electric and require a…
Mr. Brin’s presentation in Rancho Palos Verdes, Calif. — including a video of a compact two-seater autonomously doing laps around a parking lot — jolted Mr. Kalanick, according to two people who spoke with him. Google, the search giant — long considered an Uber ally — seemed to be turning on him. And even as Uber was a growing force to be reckoned with, it was lacking in self-driving car technology, an important field of study that might affect the future of transportation.
So Mr. Kalanick spent much of 2015 raiding Google’s engineering corps. To learn about the technology, he struck up a friendship with Anthony Levandowski, a top autonomous vehicle engineer at “G-co,” Mr. Kalanick’s pet name for Google.
The two men often spoke for hours about the future of driving, meeting at the Ferry Building in San Francisco and walking five miles to the Golden Gate Bridge, according to two people familiar with the executives, who asked for anonymity because they were not authorized to speak publicly.
The friendship developed into a partnership. Mr. Levandowski left Google last year to form Otto, a self-driving trucking start-up. Uber acquired it months later for nearly $700 million. Mr. Kalanick subsequently appointed Mr. Levandowski to run Uber’s autonomous vehicle research.
That relationship has since set off a legal morass, with Google’s self-driving vehicle business — now called Waymo — accusing Mr. Levandowski of creating Otto as a front to steal trade secrets from Google, then using the findings with Uber’s driverless cars. On Monday, afederal judge in San Francisco barred Mr. Levandowski from working on a crucial component of Uber’s self-driving car technology for the duration of the case.
The implications are set to reverberate far beyond the courtroom. Any setback for Uber will shake up the driverless car industry, which is locked in a bitter race to introduce and commercialize autonomous cars. Silicon Valley tech titans and Detroit automakers are making huge investments — bets that autonomous vehicle technology will usher in a new age of how people get around. For some companies, especially traditional carmakers, their very survival is at stake.
While Google has been developing autonomous vehicle technology for more than a decade, others have raced to catch up. General Motors, Ford, Apple, Tesla, Volkswagen, BMW and Mercedes-Benz are among those that have jumped in. All are competing — and sometimes cooperating — for a slice of a new market expected to top $77 billion over the next two decades, according to a study from Boston Consulting Group.
Uber has been ahead of many others in publicly testing autonomous vehicles. Last year, the company began a pilot program of autonomous cars in Pittsburgh; it has also done testing in San Francisco and Tempe, Ariz.
That aggressiveness has spurred an intense rivalry with Waymo. Waymo’s legal pursuit of Uber and Mr. Levandowski is out of corporate character; Google has tended to refrain from suing former employees who move to competitors. Many at Google and Waymo are incensed at Mr. Levandowski and how he may have betrayed them for a rich payday, according to current and former employees.
That has pushed Waymo to strike back. Beyond suing Uber, Waymo said on Sunday it had teamed up with Lyft, a ride-hailing rival, on driverless car initiatives.
“This is a race where every single minute seems to…