Click to Skip Ad
Closing in...

If you buy through a BGR link, we may earn an affiliate commission, helping support our expert product labs.

No, the hysteria over Tesla’s Autopilot is not blown out of proportion

Updated Jul 18th, 2016 11:39AM EDT
Tesla Autopilot Analysis: Why Tesla Needs to Worry
Image: Tesla

As someone who has written and negatively spoken at length of  Tesla, you might think it’s a bit odd that I went out and leased one. Except, even for every weakness I saw in Tesla, there were also undeniable strengths and achievements, and I wanted to experience that. With that said, we are now seeing something I have been saying for a while come true — Tesla is moving way too fast for a car company. This isn’t an iPhone, this is a car, and I believe that some of Tesla’s current and future problems are caused by a top down lack of focus and accountability.

I’m not going to make this into something incredibly high level, but instead offer up some specific examples of where Tesla’s lack of manufacturing, engineering, and safety experience puts its drivers at risk.

The first example is something as simple as a chime. Let me explain. I leased my 2015 P85D in November 2015, just as Autopilot was rolling out as a software update. These cars were on the ground with the hardware ready for Autopilot for a year, and then instantly, and practically overnight, they were outfitted with level 2 autonomy. It was one of the primary reasons I bought a Tesla. It completed wiped the floor with my Mercedes S-Class’ “Stop & Go” driver’s assistance feature.

The Mercedes would perform incredibly in traffic at up to speeds of 25MPH. Over that, though, it was bouncing like a ping pong ball all over the road. Tesla’s implementation was, at the time, the best on the market.

Anyway, since owning the Mercedes, I got used to something which I think is a smart move — I engage the driver’s assistance feature in the car if I ever have to glance at my phone, or fiddle with the infotainment system, so that there’s an extra safety net in case something happens in a split second. If a car in front of you slammed on the brakes, for instance.

In the Tesla, Autopilot was fantastic. And after months of using it, one day I am out driving around town and I instinctively double pull the stalk to engage Autopilot. I hear the success chime, but I don’t look at the display. I’m changing the radio station on the car’s 17-inch touchscreen display. All of the sudden I hear a riotous error sound, and as I look up, I see my car headed directly into a mailbox — I immediately swerve to not collide it. The vehicle was on Autopilot, I was going practically in a straight line around 35-40MPH. What happened?

What happened was, the chime for Autopilot engaged, and the error sound for Autopilot not being engaged, was, if not the exact same sound, the closest possible sound in the world to it. I figured Autopilot was on because of the success chime. That is just incompetence on Tesla’s part. I immediately fired off a tweet to Elon:

Lo and behold, the next software update pushed to the car fixed this problem and replaced the sound with a real noticeable error sound instead. You could say that it was my fault for not looking at the driver’s display to confirm that Autopilot was engaged, and that’s fine, I accept that unlike a lot of the narrative having to do with Autopilot in general. But the fact is, regardless of whether I should have noticed it, that is a design and safety flaw in the car that would simply not exist in most of the tier 1 car manufacturers who have been making cars for one hundred years.

Let’s talk about the unfortunate incident in Florida where someone tragically got killed in a Tesla which was operating on Autopilot. Why didn’t the car stop? Autopilot is supposed to stop in this situations. Automatic Emergency Braking is supposed to stop. Tesla blames this rare occurrence on the company tuning out road signs and glare from the sun at just the right angle, combined with this white truck that pulled across the road, that had a gap of space underneath.

Now, as far as my claim that Tesla is doing too much with too little, and too fast, let me tell you about the hardware in my Mercedes and then the Tesla to set the stage. In new model Mercedes cars equipped with driver’s assistance packages (some contain more features than others, but I believe what I mention below applies to all of them) they contain:

  • One forward-facing long range radar
  • One forward-facing short range radar
  • Stereoscopic (two) forward facing cameras
  • One rear-facing radar
  • Ultrasonic sensors around the vehicle

Tesla’s cars contain:

  • One forward facing radar
  • One forward facing camera
  • Ultrasonic sensors around the vehicle

Remember how I said that the Mercedes was terrible at the “autopilot” mode? It’s not because Mercedes is inept. It’s not because the Mercedes didn’t have the hardware to do this.

Tesla’s Autopilot is operating on one radar and one camera, and it doesn’t matter what you say about Tesla’s engineering and the company’s software advances, there is zero way you can argue that the Tesla’s driver’s assistance hardware suite is as capable as Mercedes. Is it possible the Mercedes would have stopped in front of the truck due to the aforementioned hardware? Yes.

Another problem I’ve become aware of with Tesla has to do with something as important as braking manually.

In both the Tesla and Mercedes, the driver’s assistance (Autopilot in Tesla) is disabled as soon as you manually put your foot on the brake. This disables traffic-aware cruise control, and disables the lane-keeping function that turns the wheel for you in both cars. Here’s where it gets interesting, however. In the Tesla, in Autopilot or not, when Automatic Emergency braking kicks in and you manually brake, it disables Automatic Emergency Braking, so you are now manually in control of the car’s braking power and on your own. How is that not optimal?

In the Mercedes, when it senses a collision, it will automatically pre-prime the brakes with the right amount of braking it calculates to help lessen or prevent the collision. This means when you brake, even if you don’t brake hard enough, the car is going to apply the braking force it senses necessary for you. In the Tesla, you brake and all safety features are off, and you’re left trying to prevent the collision using your split second judgement.

I don’t regret buying a Tesla. It was a fantastic way to experience a great car, the future of all cars, early. I switch cars pretty regularly so there wasn’t a specific reason I decided to go back to a Mercedes, though the fit and finish, shoddy manufacturing and service, and what seems like a general lack of knowledge and/or skill as far as safety is concerned certainly didn’t help. And I don’t want to anyone to think that I believe Tesla doesn’t care about safety, that is certainly not the case. Except the times have changed, and we’re no longer just making cars with reinforced steel pillars and curtain airbags and ABS brakes. We are making cars that are going to self drive. Cars that drive 85MPH with your family in them, Cars that you are at some point going to be able to enter an address into, go to sleep, and wake up at your destination. Hopefully not dead.

That is one main area where Tesla being such a new manufacturer matters a great heap, and I unfortunately think that as we see more and more Tesla’s on the road, and the sample size is greater, we’re going to see more of these types of issues. I also mentioned that this seems a top down issue, and something that really irked me a great amount is how on a recent earnings call, Elon Musk started to laugh in response to a question on if there had been any accidents with Autopilot. I’m still locating the clip, but it was the effect of, “there have been no accidents on Autopilot that we’re aware of” and “the only thing we’ve seen is people thinking it’s engaged when it’s really not” and laughter.

That’s not who I want in charge of the safety of my car.

Jonathan S. Geller
Jonathan Geller Founder, President & Editor-in-chief

Jonathan S. Geller founded Boy Genius Report, now known as BGR, in 2006. It became the biggest mobile news destination in the world by the end of 2009, and BGR was acquired by leading digital media company PMC in April 2010.

Jonathan is President of BGR Media, LLC., and Editor-in-chief of the BGR website.

What started as a side project at the age of 16, quickly transpired into 24-hour days and nights of sharing exclusive and breaking news about the mobile communications industry. BGR now reaches up to 100 million readers a month through the website, syndication partners, and additional channels.