Photo: Screenshot from YouTube
These days, it's pretty hard to draw a line where the automotive industry ends and the one dealing with IT begins. What's abundantly clear is that after the race for the most horsepower was over, the one for efficiency also seems to be fading, making way for a new era: that of the artificial intelligence.
Just a few years ago, nobody would have predicted that the cars were going to be the main driving motor behind AI development. And yet, our apparent desire to do anything else but drive while in our cars is what's spurring the most impressive advancements in a machine's capacity to learn and make real-time decisions.
Realizing the severe implications of this technologies, most companies chose to hold their research behind closed doors or to test a small fleet of cars under carefully controlled conditions. Google is the first name that springs to mind, but others are just as active, one of them being the Chinese company Baidu that has its own R&D center in the Silicon Valley.
Andrew Ng is a man connected with both of these enterprises. He was the co-founder of Google's Deep Learning project, but has now switched boats, and so he's currently the chief scientist at Baidu where he's working on the development of autonomous technology, among other things. Unlike Google, Baidu isn't making any of its progress public yet, so it's hard to tell exactly where they are, but they do have a testing vehicle already.
Unlike these companies, Tesla took a leap of faith and equipped its cars with semi-autonomous features even before the legislation was ready for it. What it did was say "your car can drive by itself on a highway, but we won't take any responsibility if something goes wrong, so be ready to jump in at any time." The users said "OK," and began using the new features of their cars.
Tesla has thus gained access to millions of miles of semi-autonomous date to use in the development of its system, while the owners of the cars were happy to be able to relax on longer journeys. However, this relaxation turned out to have a pretty high cost for some of them, which immediately prompted a reaction from those companies who chose to take the safer path.
In a Twitter message referring to the latest incident caused by a malfunction with the Autopilot, Andrew NG said "It's irresponsible to ship driving system that works 1,000 times and lulls false sense of safety, then... BAM!" Tesla made sure it was covered from a legal standpoint, and situations such as these do little to hurt its overall image, but that doesn't change the fact that Andrew is absolutely right.
It's not that Tesla doesn't inform what its system can and can't do, but the situations out there on the road are so varied, it's hard to decide (usually in the split of a second) in which category they fall. So you either trust the system, or you end up jumping in way too often to justify its semi-autonomous claim.
Most people choose the first option, and that's why they sometimes find out its limitations the hard way. But there's also no doubt that by acting this irresponsibly, Tesla has gained a big advantage in the race towards a truly self-driving car. And as long as nobody gets hurt along the way, the gamble will have been worth it.
Realizing the severe implications of this technologies, most companies chose to hold their research behind closed doors or to test a small fleet of cars under carefully controlled conditions. Google is the first name that springs to mind, but others are just as active, one of them being the Chinese company Baidu that has its own R&D center in the Silicon Valley.
Andrew Ng is a man connected with both of these enterprises. He was the co-founder of Google's Deep Learning project, but has now switched boats, and so he's currently the chief scientist at Baidu where he's working on the development of autonomous technology, among other things. Unlike Google, Baidu isn't making any of its progress public yet, so it's hard to tell exactly where they are, but they do have a testing vehicle already.
Unlike these companies, Tesla took a leap of faith and equipped its cars with semi-autonomous features even before the legislation was ready for it. What it did was say "your car can drive by itself on a highway, but we won't take any responsibility if something goes wrong, so be ready to jump in at any time." The users said "OK," and began using the new features of their cars.
Tesla has thus gained access to millions of miles of semi-autonomous date to use in the development of its system, while the owners of the cars were happy to be able to relax on longer journeys. However, this relaxation turned out to have a pretty high cost for some of them, which immediately prompted a reaction from those companies who chose to take the safer path.
In a Twitter message referring to the latest incident caused by a malfunction with the Autopilot, Andrew NG said "It's irresponsible to ship driving system that works 1,000 times and lulls false sense of safety, then... BAM!" Tesla made sure it was covered from a legal standpoint, and situations such as these do little to hurt its overall image, but that doesn't change the fact that Andrew is absolutely right.
It's not that Tesla doesn't inform what its system can and can't do, but the situations out there on the road are so varied, it's hard to decide (usually in the split of a second) in which category they fall. So you either trust the system, or you end up jumping in way too often to justify its semi-autonomous claim.
Most people choose the first option, and that's why they sometimes find out its limitations the hard way. But there's also no doubt that by acting this irresponsibly, Tesla has gained a big advantage in the race towards a truly self-driving car. And as long as nobody gets hurt along the way, the gamble will have been worth it.
Google News