The Debate Over Tesla’s Full Self-Driving
![Tesla Self Driving NASDAQ:TSLA Tesla Self-Driving Scam](https://www.valuewalk.com/wp-content/uploads/2021/03/tesla_1615531404-300x200.jpg)
Whitney Tilson’s email to investors discussing Volkswagen isn’t rebranding itself Voltswagen; is Elon Musk a con artist?; debate over Tesla’s full self-driving.
Q4 2020 hedge fund letters, conferences and more
Volkswagen Isn't Rebranding Itself
1) In last Tuesday's e-mail, I wrote:
We recommended Volkswagen (VOW3.DE) in our Empire Stock Investor newsletter earlier this month, and it's already jumped 26% as investors are starting to wake up to the absurd valuation disparity between Tesla and other automakers (you can learn more about Empire Stock Investor right here).
Volkswagen was in the headlines yesterday after it posted a press release on its website on Monday, apparently inadvertently, as it was dated April 29. In it, the company announced a new name for its U.S. operations, Voltswagen of America, calling the change a "public declaration of the company's future-forward investment in e-mobility."
As a result, the stock rose 4.7% yesterday to an all-time high.
It turns out, however, that it was a clever marketing stunt, according to this Wall Street Journal article: No, Volkswagen Isn't Rebranding Itself Voltswagen. Excerpt:
Volkswagen is staying Volkswagen.
Volkswagen AG's U.S. subsidiary said Tuesday the company would rebrand itself as Voltswagen of America to promote its electric car strategy, but a spokesman for the parent company in Germany later said the move was a joke.
The name change, which immediately lit up social media and online news sites, was originally intended as an early April Fools' Day stunt to get people talking about VW's ambitious electric car strategy as the company rolls out its first all-electric sport-utility vehicle, the ID.4, in U.S. dealerships, the spokesman said.
Elon Musk Is A Con Artist: Tesla's Full Self-Driving Scam Revisited
2) Speaking of electric cars, I stirred up a hornet's nest recently when I sent the nearly 3,000 folks on my Tesla (TSLA) e-mail list (to join it, simply send a blank email to: [email protected]) a link to a 12-minute video entitled: "Elon Musk Is A Con Artist: Tesla's Full Self-Driving Scam Revisited" (it's been removed from YouTube and Vimeo, but as of this morning, it was still posted here and here).
In my comments, I wrote:
It's over the top – I don't think Musk is a con artist. He's built multiple GREAT companies making REVOLUTIONARY products (most notably cars and rockets).
But he's clearly been overpromising autonomous driving to a MASSIVE degree, which is not only a fraud on investors, but much worse, incredibly DANGEROUS as many Tesla drivers are lulled into complacency.
Tesla's "Full Self Driving," which Musk claims is Level 5, is no more than good Level 2 – driver assistance.
Hopefully regulators will finally rein Musk and Tesla in – though I'm not holding my breath...
P.S. I'm getting tired of people e-mailing me every time I send out something negative on Tesla, accusing me of being short the stock. (Ditto for ARKK.) I no longer run a hedge fund. I am in the investment newsletter business. I am not short ANY stock (and haven't been since I closed my hedge funds in late 2017).
This led to many interesting responses and conversations, which I wanted to share here because the debate about autonomous vehicles ("AVs") isn't just about one man/company/stock. I think AVs are going to change the world, as my colleague Enrique Abeyta and I discussed in our recent webinar, "TaaS 2.0" – you can watch the replay here.
Here are four e-mails I received:
- "There's nothing in this video that hadn't been pointed out before. It's been so obvious all along. The only material contributing piece of evidence the video missed was the October 19, 2016 press conference – almost five years ago! – in which Musk claimed Level 5 for the first time in association with saying that's why you should buy a Tesla now. Here's an excerpt from the transcript:
"'Elon Musk: Basic news is that all cars exiting the factory have hardware necessary for Level 5 Autonomy so that's in terms of cameras, compute power, it's in every car we make on the order 2,000 cars a week are shipping now with Level 5, literally meaning hardware capable of full self-driving for driver-less capability.'"
- "I watched the entire video. I am not seeing a scam at all. Clearly this is a massive undertaking by Tesla. I watched several hours of video with James Douma explaining Tesla FSD. Astounding the potential it will serve transportation and it is getting closer. I know you are familiar with Tony Seba your TAAS ideas would have roots to his writings I believe. Maybe the regulators are not reigning him in is because he is on the right track and the chances of implementation are close. You should try to interview him, I think both of you have shared goals. Not likely to see him on a mountain top. But more likely to see you on Mars."
My reply:
There's a judgement call about whether the word "scam" is appropriate, but saying you have Level 5 (or will have it within months) when you have Level 2 qualifies in my book.
- "Consider the advice my dad gave me – can't take the heat get out of the kitchen [I left out an expletive]. Maybe you should try writing about Ford and Chevy's EVs."
- "You all have it wrong. Tesla auto-pilot cars in Q4 of 2020 ranged 3.45 million miles per accident. Compare that to "stupid" cars that only got 484 thousand miles before an accident. You are the one committing fraud. If the public was educated on the truth, non-auto-pilot cars would be outlawed and robo-taxis would be everywhere. Auto-pilot will only get better over time, whereas human drivers will remain just as incompetent as ever."
My reply:
Tesla has never revealed its underlying data so I'm skeptical of its claims – even more so because it disingenuously compares its high-end luxury cars to all cars on the road. When my friend Chris Brown of Aristides Capital (aka Midwestern Hedgie) did an apples-to-apples comparison in 2018 (using the Models S and X; the Model 3 wasn't out yet), he found that Teslas were far more dangerous than comparable cars: Tesla's Driver Fatality Rate is more than Triple that of Luxury Cars (and likely even higher). Excerpt:
Tesla CEO Elon Musk has worked hard to convince shareholders, media, and, most-importantly, car buyers that Tesla is the "safest vehicle" on the road. He said in May that the brand's fatality rate is "approx 4X better than average," expanding upon his claim by Tweeting, "According to NHTSA, there was an automotive fatality every 86M miles in 2017 (~40,000 deaths). Tesla was every 320M miles. It's not possible to be zero, but probability of fatality is much lower in a Tesla." What is the actual record of Tesla cars when compared apples-to-apples with their luxury peers? In this article, we answer that question with statistical rigor...
Even without making any adjustment whatsoever for missing fatality data, Tesla drivers are much more likely to die than their peers driving other luxury cars. Eleven deaths in 265,290 vehicle-years is a stunningly high driver fatality rate of 41.46. That's quadruple the rate of Audi and BMW, and more than triple the rate of all luxury cars combined.
My take: I don't see any reason why Teslas would be more dangerous than comparable cars, but Tesla's self-selected, self-reported data doesn't persuade me that their cars are safer either...
Its Level 2 Autopilot system is excellent and no doubt enhances safety WHEN USED CORRECTLY. But when Musk oversells it as Level 5: 1) it's fraud; and 2) it's dangerous because many drivers stop paying attention.
Autopilot: Musk's Wishful Thinking or Tesla's Greatest Advantage?
3) Here's my friend Vitaliy Katsenelson's take: Autopilot: Musk's Wishful Thinking or Tesla's Greatest Advantage? Excerpt:
I have mixed feelings about Autopilot. It can be liberating. On our trip from Denver to Santa Fe, we used Autopilot 80% of the time, as most of the road was a clearly marked highway.
However, when you drive on a road that doesn't have a clearly marked median, Autopilot is not to be relied upon and, in fact, can be very dangerous. On two occasions when the road curved and the lane markings were interrupted by an intersection, Autopilot almost took me into incoming traffic; I had to intervene.
Autopilot is almost ready for prime time. The problem is that the word "almost" should never be used in the same sentence with "Autopilot." An almost-working Autopilot is like an almost-working airplane – it might lead to deaths. This is why self-driving reliability standards are measured four or five digits to the right of the decimal point. Autopilot has to be reliable 99.99999% of the time. Achieving the 9's to the right of the decimal point is much more difficult than to the left of it.
Musk's Approach To Autonomy
4) My analyst Kevin "100-bagger" DeCamp shares his thoughts:
As with most other debates around Tesla (especially when they get heated), people tend to take sides and reside in their filter bubbles – Elon is a fraud or Elon is a god.
As usual, it turns out it is much more nuanced, especially when it comes to autonomy.
First of all, Musk deserves criticism for the way he's handled expectations for autonomy and the promises he made for many years. He is always late on timelines – hence the phrase "Elon time." However, the common retort is he always eventually delivers and he's accomplishing impossible things – so does it really matter if he's "late"?
With autonomy, however, it has been next-level late relative to his promises, leading some to call it a scam, or "Elongate" if you prefer.
But, what matters NOW is do they finally have the pieces in place to be on a clear path to solve autonomy in the near future? If this is true, then this debate will quickly become irrelevant.
I am still not convinced 100%, but I am gaining conviction by the day. If people are answering this question with a resounding NO simply by looking at Musk's empty promises in the past and the flaws of Tesla's current Full Self-Driving beta without doing a deep dive on the fundamental differences between Waymo's versus Tesla's approach, pseudo-lidar, neural nets, machine learning, and all of the end-to-end solutions (especially the fleet data collection and curation capabilities) that Tesla has put in place to solve artificial intelligence ("AI") vision only autonomy, then there is a decent chance that they will be shocked by how wrong they end up being. Only time will tell, but this is not the time for bears to declare they were right all along on FSD – although I know they will and it might be their last chance!
By the way, Tesla is clearly the leader in "TaaS 1.0" (Transportation as a Service), but if they show the market within a few months that they have a clear path to solve autonomy, then you better believe that the stock will react violently to the upside (yes, even from its $600 billion market cap) as the world realizes that they will be the leader in "Taas 2.0" as well...
I replied: "Agreed – but from what I've read (I'm no expert), Musk's approach to autonomy will never get Tesla to Level 5. You gotta have Lidar."
Our mutual friend added:
I agree 100% that the line between "god" and "fraud" can be thin as well as fluid over time. I remember vividly around 20 years ago when the last "new energy god company," Enron, was hailed as the savior of all things Earth and mankind, but fell from this impressive position in a matter of months.
One day Ken Lay and Jeff Skilling were "gods" – and then suddenly they were frauds.
Here today, gone tomorrow – some might say "reversion to the mean" or "the law of averages." What are the odds that an unprofitable (sans subsidies) company that has 0.7% of the world's auto market but close to 50% of the market cap will at some point revert to the mean? It does strike me as a reasonable bet.
Kevin replied:
To Whitney's point about Level 5 not being possible without Lidar – this appears to be the expert consensus and I am certainly no expert either so it may ultimately end up true. But how many times have experts been wrong in the past? Here are some experts discussing Musk's view on Lidar: Is Lidar Doomed? Autonomous Driving Experts Answer (Compilation Video) (10:18).
Although what Waymo has accomplished in Phoenix is impressive, it is nowhere near true Level 5 and you couldn't pick an easier place on Earth to drive. I don't understand how Waymo will break out of that geo-fenced box without having a much larger fleet to solve vision. I feel like there is a chicken-and-egg problem. How do you scale it? With Tesla's approach, I can at least envision how it is theoretically possible to solve self-driving.
When you add to the enormous complexities of autonomous driving to the fact that interference between multiple Lidars is still not a solved problem (see this article) and could obviously get worse as more self-driving cars are on the streets, I don't think we are close.
In the hypothetical scenario where vision is solved with sufficient depth perception through pseudo-Lidar using cameras and deep learning, then Lidar may actually be unnecessary and a crutch that simply gets you to geofenced line-following robots quickly, but can never get you all the way to level 5 without a large enough fleet and sufficient data. If you don't need these very expensive, energy and resource hungry training wheels in the end, why ever use them in the first place? Skate to where the puck is going, right? That was Musk's big bet and we'll see if he's right in the very near future.
My guess is that Tesla gets there first, but Waymo and comparable autonomy efforts continue to slowly solve bigger and bigger geofenced areas as the market is enormous. It will be fascinating to watch this unfold!
Best regards,
Whitney
The post The Debate Over Tesla’s Full Self-Driving appeared first on ValueWalk.
Source valuewalk