Photo: Tesla Motors
The thing with being able to issue updates to all vehicles built on a virtually daily basis is that it inadvertently makes the manufacturer use this feature. And if it can use it, it can also abuse it.
Those who play video games know just how frustrating it is when a developer decides to release a game early, even though it may not have gone through all the regular testing procedures in a satisfactory manner. But Christmas is coming, and that's when parents forget about how they'd like to see their children play less on the computer and buy them the games they want. Besides, what's the worst that could happen? The game will get some not so great initial reviews, but a few patches later, it'll all get better.
It looks like Tesla did something similar with its Autopilot feature. Unlike the video games developers, its incentive wasn't the Christmas sales craze, but the fact that it could be the first manufacturer to put a large fleet of vehicles on the road to gather data for it. Elon Musk insists this doesn't make the Autopilot a "beta test" product (because it did go through internal testing as well), but it kind of does. And by saying that "certainly moderate and maybe big advances" are possible with Tesla's current level of hardware (via Electrek), he's basically admitting to that.
He's saying that Tesla released a piece of software before it was sure there was no way it could be further improved. Perhaps he's right, and beta testing isn't the right name for it. It's more like a product rushed to the market. It certainly made the headlines when it was launched, and considering the hype Musk created around it, it's only natural that its first major malfunction was met with the same interest.
Now, Elon Musk revealed the latest developments the company is working on in two Twitter messages. He said Tesla is "working on using existing Tesla radar by itself (decoupled from [the] camera) w[ith] temporal smoothing to create a coarse point cloud, like LIDAR." Four hours later he came back to emphasize radar's advantages over the more sophisticated LIDAR: "[The] good thing about radar is that, unlike LIDAR (which is visible wavelength), it can see through rain, snow, fog, and dust."
So, this means that the Autopilot 1.0 has still got some life in it. It also suggests Tesla might not go down the LIDAR path at all, opting instead to improve its radar sensors over distance (since not working in bad weather would be a big inconvenient). As for Autopilot 2.0, it'll be very interesting to see what kind of hardware it uses.
It looks like Tesla did something similar with its Autopilot feature. Unlike the video games developers, its incentive wasn't the Christmas sales craze, but the fact that it could be the first manufacturer to put a large fleet of vehicles on the road to gather data for it. Elon Musk insists this doesn't make the Autopilot a "beta test" product (because it did go through internal testing as well), but it kind of does. And by saying that "certainly moderate and maybe big advances" are possible with Tesla's current level of hardware (via Electrek), he's basically admitting to that.
He's saying that Tesla released a piece of software before it was sure there was no way it could be further improved. Perhaps he's right, and beta testing isn't the right name for it. It's more like a product rushed to the market. It certainly made the headlines when it was launched, and considering the hype Musk created around it, it's only natural that its first major malfunction was met with the same interest.
Now, Elon Musk revealed the latest developments the company is working on in two Twitter messages. He said Tesla is "working on using existing Tesla radar by itself (decoupled from [the] camera) w[ith] temporal smoothing to create a coarse point cloud, like LIDAR." Four hours later he came back to emphasize radar's advantages over the more sophisticated LIDAR: "[The] good thing about radar is that, unlike LIDAR (which is visible wavelength), it can see through rain, snow, fog, and dust."
So, this means that the Autopilot 1.0 has still got some life in it. It also suggests Tesla might not go down the LIDAR path at all, opting instead to improve its radar sensors over distance (since not working in bad weather would be a big inconvenient). As for Autopilot 2.0, it'll be very interesting to see what kind of hardware it uses.
Google News