Elon Musk shares painfully obvious idea about how difficult self-driving cars are



[ad_1]

Illustration from article titled Elon Musk shares painfully obvious idea about how difficult self-driving cars are

Photo: Ringo HW Chiu (AP)

Elon Musk, one of the richest people on the planet, tweeted Thursday night about the real problem with self-driving cars. And as you can probably guess by now, it’s one of those things that feels deep until you stop to think about it for three seconds. That’s more or less Musk’s mark at this point.

“Much of the real-world AI needs to be solved for autonomous driving to be widespread and unsupervised, because the entire road system is designed for biological neural networks with optical imagers,” Musk wrote.

What does it mean that the system was designed for “biological neural networks with optical imagers?” It’s a fancy way of saying that our current road network was designed for the human brain and human eyes.

Musk’s sentiment is correct but perhaps painfully obvious in a way that makes the billionaire seem like playing the game as a human being. And the underlying assumption of the tweet is that 1) AI cars would be so much easier without those pesky human-centric designs. and that 2) the problem with AI in cars is not necessarily the technology of cars. The problem, as Musk wants to suggest, is the technology that surrounds cars, or the infrastructure that has enabled all of the way we drive over the past century.

Why is Musk tweeting something like this in the world? Perhaps because he has to blame elsewhere following so many damning reports about his automaker Tesla. Musk has been widely criticized for Tesla’s promotion of a feature called AutoPilot, which is supposed to allow semi-autonomous driving.

AutoPilot isn’t intended to allow fully autonomous driving, if you believe Tesla’s advocates, but you don’t have to look too far to find people testing AutoPilot’s limits. Videos have gone viral of people driving Teslas with no one in the driver’s seat, as you have almost certainly seen by now. But people have died with AutoPilot engaged, the most recent example being two people in Texas where firefighters had to call Tesla for advice because the blaze burned for more than four hours. Firefighters used more than 30,000 gallons of water to extinguish the blaze, which kept turning on again because of the car batteries. Tesla denies that no one was driving, despite what local authorities say, but an investigation is underway.

Ironically, what Musk more or less discusses in his latest tweet is contrary to his real talents. Musk is widely regarded in popular culture as a visionary with good ideas and little following. His “public transport” ideas like the Hyperloop are a prime example. In 2013, Musk unveiled his idea for a vacuum tube vehicle system that could travel at 600 miles per hour. And he shared this idea with the world hoping someone would make it happen. But it turns out that the difficult thing in creating a high-speed mass transit system isn’t the technology, it’s the politics and land rights, something I pointed out at the time, even before Musk officially announced his idea.

Musk would love to tear down the road network and build something easier to identify for his Tesla cars and to drive safely. But that’s not the problem he faces. It would be like robot makers requiring all new buildings to have only one floor and no stairs, as most human-style robots struggle to climb stairs. The old Darpa Robotics Challenge had real world obstacles For a reason. We want robots to adapt to our world, we don’t want to change our world to make robots more comfortable. As reporter Kelsey Atherton Put the, “Tech interprets humans as flaws and seeks to work around them.”

But none of this is new. Darpa tried to make a completely autonomous car in the 1980s, but it went on to become confused by snow on the road at their test track outside of Denver. Darpa didn’t say, “Okay, well, we’ve got to keep it from snowing Colorado and then that’ll be fine.” They admitted that they had not yet solved the puzzle.

The problem, if Musk is to continue on this path of autonomous driving in the real world, is to do something for our current roads on which his cars can be safely driven. If he wants to build a safer transportation system from scratch, he should invest in monorails. Or, better yet, he can dig a tunnel and build something that would really blow us away.

Musk promised us the latter in places like Chicago and Las Vegas, but all we got was a Tesla with human pilots moving. slowly in a tunnel. When given endless possibilities with new ground, he simply falls back on his old ways. But it all makes sense. Musk is fundamentally a businessman, not a visionary.



[ad_2]

Source link