[ad_1]
John McNeil — the former Tesla and Lyft CEO and current CEO of DVx Ventures — spread the word last week at a global business forum on what he calls the “Tesla Innovation Method” for radical simplification.
After his presentation, we sat down to chat about Tesla’s automation goals, its new robotaxi, and why building electric cars is a tough business.
On a Tesla robotic car
Tesla introduced Autopilot, an advanced driver assistance system (ADAS) that can perform limited automated driving tasks, in 2015. Automation has been a key Tesla initiative for nearly a decade, McNeil said.
“It was a topic of conversation every week, if not every day,” he told TechCrunch.
That’s partly why McNeil, along with many others, was surprised by the much-anticipated burger Tesla unveiled as a robo-taxi earlier this month.
“A lot of times, Elon will have something up his sleeve that he won’t reveal until one of these product events. And so I was really wondering, what does he have up his sleeve? What does he have? “Because Waymo and Cruise are already at Level 4. And Tesla is at Level 2. He must achieve some major achievements that he will talk about in order to move from Level 2 to Level 4.”
(the Sai Level 2 automation is defined as ADAS that can do things like brake assist and lane centering, but requires a human to constantly drive and supervise. Level 4 means the car can drive itself without a human under certain conditions.)
But in the end, Musk only unveiled 20 robotaxi prototypes. “That’s not to say they don’t have something up their sleeve yet, but I was a little surprised that we just got a prototype.”
McNeil said he expects Tesla to reveal a roadmap to explain what will happen and when in order to truly build trust in the brand. Confidence comes through more detail, he says.
(After our interview during Tesla’s third-quarter earnings call, Musk said he hopes to launch autonomous transportation service in California and Texas in 2025.)
On automation through vision only
McNeil’s presentation was about how Tesla can succeed by streamlining operations. I asked him whether Tesla’s choice to pursue autonomous driving through cameras alone, rather than relying on lidar and radar, was a bit of a false simplification. Yes, using only cameras may simplify the sensor array, but it creates a lot of work on the back end.
On top of being an automaker, an electric vehicle charging company, and a solar energy company, Tesla is now building Dojo, a supercomputer that it hopes will help it train the neural networks needed to achieve full self-driving through vision alone. The Dojo is a bet that Musk admitted might not work out.
McNeil expressed concern about Tesla’s vision-only approach. “A human can drive a car, and we only have two eyes. So, if you give a car eight eyes, it should be able to solve the problem, right? But I think that misses part of the challenge, which is that there are things that make us unable to see.” Such as snow, fog, sun glare, darkness and other things.
“Having eyes may not be enough. Lidar can see all these things… And I think when people’s lives are at stake, you can’t cut costs or cut the amount because there’s a safety issue. So some people say you can only solve the problem.” Through vision, but I wonder if that’s really true because I’m not sure human vision is enough and that’s why we have 40,000 traffic deaths a year.
Warning: McNeil sits on the board of GM’s Cruise, which marketed Level 4 autonomous driving using lidar technology before grounding its fleet last year after a safety incident.
“If the eye can’t see through the dark, through the rain, around a corner, etc., I don’t know how a supercomputer can solve that.”
About the idea of Tesla adopting lidar technology
Tesla has supported its vision-only approach for years. It’s not resolved yet, and it probably never will get there. I asked McNeil if it was crazy for Tesla to reintroduce lidar back into the system. Or did the automaker sink its costs into cameras?
McNeil said lidar costs have come down a lot, and financially, it wouldn’t be a big problem for Tesla to adopt the sensor. But from a reputation standpoint? Tesla has a track record of not being afraid of the sunk cost fallacy or even customer frustration, McNeil said.
He pointed to Musk’s statements since at least 2019 that any Tesla purchased next would have the computing and sensor suite needed for full self-driving when the feature becomes available. Tesla was forced to back away from that promise, and during the automaker’s third-quarter earnings call, Musk said vehicles with older hardware would receive a free update to accommodate unsupervised FSD.
On why electric vehicle startups are failing today
Fisker is going through a messy bankruptcy. Rivian loses approximately $30,000 for every car it produces. Canoo has changed its business model so many times, that we have lost track. In short, building cars from the ground up is hard. I wondered: Do you need a wild person like Elon Musk at the helm to be successful?
“I think you need a crazy, effective leader to get through the challenge of everything you have to prove as a car company,” McNeil said. “It’s hard to describe how difficult it is to take 10,000 parts from 72 countries and make a car out of them, and then put the software together and get the car going. Like manufacturing this stuff is really the really hard part.
Reflecting on his disappointment with Tesla’s robotaxi event, McNeil said it’s really easy to produce a prototype, but it’s really hard to mass-produce it.
“This is where the auto industry is incredibly brutal, even for incumbents,” he said. “You’re seeing Stellantis really struggling now. You’re seeing Volkswagen really struggling. That’s because the fixed costs are so high. If you don’t get the product right and you don’t sell it, it’s unforgiving.”
[ad_2]