Thursday, May 9, 2024

Elon Musk: Threat Or Menace? Part 5

Source
Much of this series has been based on the outstanding reporting of the Washington Post, and the team's Trisha Thadani is back with Lawsuits test Tesla claim that drivers are solely responsible for crashes. My main concern all along has been that Musk's irresponsible hyping of his flawed technology is not just killing his credulous customers, but much more seriously innocent bystanders who had no say in the matter. The article includes video of:
  • A driver who believed Autopilot could drive him home despite his being drunk. The car drove the wrong way on the highway and killed another innocent victim of Musk's hype.
  • Autopilot rear-ending a merging vehicle and killing another innocent victim, a 15-year-old.
  • Autopilot slamming into a broken down vehicle on the highway. When the Tesla driver left the wreck she was hit and killed by another car.
  • Autopilot speeding through a T-junction and crashing into a parked truck.
Below the fold I look into Tesla's results, Musk's response, the details revealed by the various lawsuits. and this excellent advice from Elon Musk:
"If somebody doesn’t believe Tesla is going to solve autonomy, I think they should not be an investor in the company."
Elon Musk, 24th April 2024

The Results

In Tesla’s biggest problem: cars, Drew Dickson looks at Tesla's first quarter results:
Of Tesla’s total quarterly sales of $21.3bn, 82 per cent were indeed “automotive revenues” while the rest were energy and services.
...
Tesla burned through $2.5bn of cash in the quarter. Inventories grew by over 10 per cent to $16bn.
82% of $21.3B is $17.5B, so Tesla has almost an entire quarter of unsold cars on hand.

One problem is that Musk's persona as an extreme right-wing troll has been putting off the key Tesla customer demographic, well-off liberals who care about climate change. Another problem is that Tesla lineup of models is old and expensive. Tesla used to recognize that they needed a cheaper product but:
A cut-price Model 2 was first teased at the 2023 Tesla AGM, with Musk saying in January that it would be in production towards the end of next year, but the expected spring product announcement never came.

Tesla now states it is “accelerating” plans, though as with the Cybertruck it’s easy to mistake the accelerator for the brakes. The notion that a Model 2 might be built in new factories in Mexico or elsewhere have been replaced with vague commitments to retool existing infrastructure and production lines.
The resources that could have developed a Model 2 or refreshed the existing models instead went to develop the "Incel Camino", the Cybertruck. This isn't just a $82K laughing-stock, but a manufacturing nightmare that will be lucky to sell 20% of Musk's 250K/year projection, especially since it cannot be road-legal in either of Tesla's #2 and #3 markets (China and EU). It will definitely be a drag on the results for some time. So the Models S (2012), X (2015), 3 (2017) and Y (2020) will have to soldier on for a while.

This aging product line isn't attracting customers:
  • Sequential growth in units sold was down 13 per cent sequentially.
  • Tesla’s price per vehicle, excluding regulatory credits and leasing or finance income, was $38,924.
  • This was down 13 per cent from $44,642 last year, which itself was down 11 per cent from $50,037 the previous year.
Extreme pricing pressure is forcing affordable vehicles on Tesla, irrespective of whether it chooses to launch one. Amid a lack of demand for EVs in general, and Teslas in particular, its quarterly automotive revenues were down nearly 13 per cent over the past year and by over 19 per cent sequentially.

“Clean” automotive margins (which exclude regulatory credits and leasing income) were down from 29.7 per cent in the first quarter of 2022, to 18.3 per cent in the first quarter of 2023, and again to 15.6 per cent in the first quarter of 2024. If you back out the new IRA US tax credits (which Tesla doesn’t seem to disclose) then automotive gross margin looks to have fallen even further, to around 14.1 per cent.
Shrinking margins on shrinking sales hit earnings per share:
GAAP EPS was down 53 per cent year-on-year, accelerating from the 23 per cent drop in the first quarter of 2023. Even using non-GAAP EPS it’s a 47 per cent decrease over the past year.

In the summer of 2022, when the stock was above $300 share, analysts were expecting Q1’24 EPS of $1.80. Instead, they got $0.45. That is a 75 per cent downgrade to expectations.
Source
And the upsell of Fake Self Driving isn't helping, as Craig Trudell reports in Tesla’s Self-Driving Software Is a Perpetual Revenue Letdown:
Tesla released its 10-Q, a quarterly report that provides a more detailed view into the company’s financial position. For several years running, Tesla has provided regular updates in these statements on how much revenue it’s taken in from customers and not yet fully recognized. Some of this deferred revenue relates to a work-in-progress product: Full Self-Driving, or FSD, for short.

Tesla’s deferred automotive revenue amounted to $3.5 billion as of March 31, little changed from the end of last year. Of that amount, Tesla expects to recognize $848 million in the next 12 months — meaning much of the performance obligations tied to what it’s been charging customers for FSD will remain unsatisfied a year from now.
...
In these filings, Tesla also reports how much deferred revenue it’s actually recognized — and the Austin-based company has consistently undershot its own forecasts. It has recognized $494 million of deferred revenue in the last 12 months, short of the $679 million that it projected a year ago.
Source
Tesla's CFO quit last August:
The carmaker reported this week that its operating margin shrank to 5.5% in the first quarter, the lowest since the last three months of 2020. The measure of profitability was at 16% when Zachary Kirkhorn, Tesla’s then-chief financial officer, said during an earnings call that it was key to the company.

“As a management team here, we’re most focused on what our operating margin is,” he said in January 2023, in response to an investor question on a different earnings metric. “That is what we’re primarily managing to now.”
General Motors operating margin is 7.35%. But not to worry, Tesla isn't a car company, its an AI and robotics company:
If the auto business is worth 3 or 4 times the multiple of a Stellantis or Volkswagen, then it would get a forward PE of, say, 20x. That’s more than generous for a business the CEO talks about as a legacy sideline.

Street numbers for Tesla are consistently far too high but even using the 2024 consensus EPS of $2.64, Tesla would be worth just over $50 per share. Using today’s diluted shares (and assuming that they don’t issue more, which they will) that works out to a market cap of $181bn.

Tesla’s fully diluted market cap at pixel time is still $580bn. Simplistically, that means shareholders are already paying around $400bn for corporate experiments in “robotics and AI”, along with anything else Musk has or tries to conjure up.
Tesla will definitely issue more shares, for example after the 13th June shareholder vote when they will reward Musk's corporate experiments in robotics and AI by reinstating the $56B incentive package cancelled by the Delaware court.

Pumping The Stock

About 70% of the stock price is based on Musk hyping the technology. Thus for Musk it is more than twice as important to pump the stock as it is to sell more cars. He has to follow two strategies:
  • Make the results look better in the short term by increasing margins. The obvious way to do this is to cut costs, even though this will reduce profits in the longer term. After all, in the longer term Tesla isn't about selling cars, it is about AI and robotics.
  • Distract people from looking at the results by unleashing the hype cannon.

Cutting Costs

The knee-jerk reaction of US companies to bad quarterly results is to lay off staff, but they generally target the less successful parts. Elon Musk not so much:
Even Tesla's harshest critics must concede that the company's Supercharger network is its star asset. Tesla has more fast chargers in operation than anyone else, and this year opened them up to other automakers, which are adopting the J3400 plug standard.

All of which makes the decision to get rid of senior director of EV charging Rebecca Tinucci—along with her entire team—a bit of a head-scratcher. If I were the driver of a non-Tesla EV expecting to get access to Superchargers this year, I'd probably expect this to result in some friction. Musk told workers that Tesla "will continue to build out some new Supercharger locations, where critical, and finish those currently under construction."
Like most of the recent desperation moves, this was Musk's decision:
The decision to cut the nearly 500-person group, including its senior director, Rebecca Tinucci, was made by Chief Executive Officer Elon Musk in the last week, according to a person familiar with the matter.
In return for gorvenment subsidies, Tesla had been turning Superchargers into a separate business:
Access to high-speed charging is critical to EV adoption, and Tesla invested billions of dollars into developing a global network of Superchargers that became the envy of other automakers. It’s also a critical driver of Tesla sales, and the carmaker pointed to the division’s growth during its first-quarter results just last week.

“Starting at the end of February, we began opening our North American Supercharger Network to more non-Tesla EV owners,” Tesla said in its shareholder deck.

The Musk-led company has also signed charging partnerships with carmakers including Stellantis NV, Volvo, Polestar, Kia, Honda, Mercedes-Benz and BMW. It’s not clear who will now oversee Tesla’s partnerships with those companies. GM, Volvo and Polestar were all due to open NACS chargers to their customers in the immediate future, according to Tesla’s website.
But maybe Musk couldn't resist a chance to mess with the competition:
The job eliminations mean Rivian, Ford and others have lost their main points of contact in Tesla’s charging unit shortly before the kickoff of the busy summer driving season. Tinucci was one of the main executives building and managing outside partnerships and was thought of highly, two people who had worked with her inside and outside of Tesla said.
Musk Undercuts Tesla Chargers That Biden Lauded as ‘a Big Deal’ by Craig Trudell suggests a political motive:
In addition to potentially compromising budding partnerships with other carmakers looking to tap Tesla’s chargers, another consequence of Musk’s move may be undercutting Biden’s EV push in the midst of his reelection campaign. Presumptive Republican nominee Donald Trump has repeatedly attacked electric cars on the campaign trail and predicted a “bloodbath” for the auto industry if he isn’t elected.
Faced with a huge short-term threat to his wealth Musk isn't concerned with the longer term, when unlike robotaxis, Superchargers could have been a nice little earner:
Tesla had been building a tidy charging business over more than a decade. BloombergNEF estimates that the company delivered 8% of the public charging electricity demanded globally last year. Before Musk’s surprise decision, the researcher was projecting that Tesla’s annual profit from Supercharging could rise to around $740 million in 2030.

That level of earnings is now likely out of reach, as BNEF’s estimates assumed Tesla would accelerate the pace of installations through the end of the decade. Musk had given indications this was the plan.
Musk may already be having second thoughts:
The move will slow the network’s growth, according to a person familiar with the division, who asked not to be identified discussing private matters. There already are discussions about rehiring some of the people affected in order to operate the existing network and grow it at a much slower rate, the person said.
Way to motivate the team, Elon!

Musk believes the future depends upon robotaxis but:
Many Tesla fans had been holding out hope that Musk would debut a cheap Model 2 EV in recent weeks. Instead, the tycoon promised that robotaxis would save the business, even as both of its partially automated driver assistance systems face recalls and investigations here in the US and in China.

Delivering on that goal is more than just a technical challenge, and it will require the cooperation and approval of state and federal authorities. However, Musk is also dissolving the company's public policy team in this latest cull.
Cutting off communication with the regulators who will have to approve robotaxi service isn't likely to help. And if there was another technology critical to Tesla's success it would be batteries:
Earlier this month, Tesla engaged in another round of layoffs that decimated the company and parted ways with longtime executive Drew Baglino, who was responsible for Tesla's battery development.
Jonathan M. Gitlin rounds up reactions in What’s happening at Tesla? Here’s what experts think. He quotes Ed Niedermeyer:
Car companies "go bankrupt because A, they overinvest in factories, and then demand falls off. Which... that fits the profile," said Niedermeyer. "And B, they don't invest in products. Not investing in products is sort of a longer-term cause, and the proximal cause is [that] demand falls, and you've been investing in too many factories, and you get crushed by those fixed costs. So those cases that are common across most auto industry bankruptcies are certainly there."

But with almost $27 billion of cash on hand, that shouldn't happen any time soon. "The thing that is really hard to understand is that if you have tens of billions of dollars in cash but you're losing market share and you're losing margin, losing pricing power, and all the other things that are happening with the business—you don't cut your way out of that problem," Niedermeyer continued. "That's the confusing part about all this. What would you use that cash for if not to solve those problems? And yet, instead, they're cutting.

"One of the things I've said for a really long time, and I think this is what's happening, is that an automaker is not really real until they survived a serious downturn," Niedermeyer said. And while the broader economy looks fine, EV sales are battling a strong negative headwind. "The car game is a survival business. You can capture more upside than the other guy in the good times. And that can be really good for your stock. But if you do that by not investing in the things that protect you in the downturn, it doesn't matter. And you're just another one on the list of defunct automakers,"
Musk isn't listening, because he is still firing people:
On Sunday night, even more Tesla workers learned they were no longer employed by the company as it engaged in yet another round of layoffs. ... The latest round of layoffs has affected service advisers, engineers, and HR.

Hyping The Technology

The Washington Post team's Faiz Siddiqui and Trisha Thadani report that Tesla profit plunges on price cuts, but company unveils plans for affordable models:
CEO Elon Musk, who has a unique penchant for redirecting the conversation, used Tuesday’s earnings call to deflect from the poor numbers, focusing instead on the company’s commitment to artificial intelligence and a fully autonomous car. Details on Tesla’s apparent new offerings — which include the “more affordable models” and the “cybercab” — were scant and did not address how the company would overcome the technological and regulatory hurdles ahead.`
Musk has form when it comes to hyping his technologies and companies. His tweeting that funding had been secured to take Tesla private at $420/share led to a settlement with the SEC that is still in place:
The supreme court on Monday rejected an appeal from Elon Musk over a settlement with securities regulators that requires him to get approval in advance of some tweets that relate to Tesla, the electric vehicle company he leads.
The hype is starting to wear thin but not yet with the markets, as Brandon Vigliarolo points out in Musk moves Tesla's goalposts, investors happily move shares higher:
Elon Musk has a strategy and you may have seen it before: When things aren't going well, he'll say something wild to take everyone's eyes off the trouble, and raise share prices with dreams.
...
The first quarter of 2024 didn't go well for Tesla, either economically or reputationally. As we reported earlier, sales fell, net profit tumbled off the same cliff Tesla's stock price earlier careened over, and production and deliveries decreased as well.

But give Musk a chance to toss out a flash grenade and he'll do just that: This time around with some wild predictions about his automaker producing a "purpose-built robotaxi" dubbed the "Cybercab," and Tesla's latest vision for the future as one in which it is focused on "solving autonomy."
...
"It's like some combination of Airbnb and Uber, meaning that there will be some number of cars that Tesla owns itself and operates in the fleet … and then there'll be a bunch of cars where they're owned by the end user," Musk said. He added the fleet will likely grow to include "several tens of millions" of vehicles by the end of the decade.
Last year Tesla shipped 1.8M vehicles. There are 6 years left to the "end of the decade". Musk is promising to ship an average of at least 3M vehicles/year, all of which would be enrolled in the robotaxi fleet. Even if this were plausible, one has to question where all the riders would come from for a fleet 2.5 times bigger than Uber's global driver list. Note that in the US 36% of adults have used Uber or Lyft, so the market is already close to saturated. I'm sure we all remember that:
Musk spent plenty of time in the 2010s claiming he'd have one million robotaxis on the road by 2020.
Pumping the stock full of hype is a Musk habit:
Getting in trouble over "Full-Self Driving" claims? Stick a guy in a robot suit and call it Optimus to distract shareholders. Fail to get FSD realized this year - again? Just kick it down the road. Journalists calling him out on his nonsense? Rant about the "woke mind virus" and the media on Twitter.

Of course, Optimus has been nowhere to be seen and was barely mentioned during the call. Likewise, Tesla's dreams of tens of millions of robotaxis on the road in the next six years rests on the need for serious technological breakthroughs the automaker has failed to make despite years of trying. Oh, and a ton of permits if this is to operate in the States, at least.
Vigliarolo isn't alone. In Musk Sells the Tesla Dream, But Don't Ask for Details Liam Denning notices a detail from the earnings call:
There was an odd tweak to the low-cost vehicle strategy Tesla laid out in March 2023, when management talked about cutting costs in half with revolutionary manufacturing methods. Now, Tesla talks about melding aspects of next-generation platforms with its existing ones in the new models, enabling the company to build them on existing manufacturing lines. To be clear, that is an intriguing possibility, offering efficiencies to reduce stubborn costs.

But also to be clear: It won’t deliver a $25,000 Model 2 anytime soon — “this update may result in achieving less cost reduction than previously expected” — and also isn’t what Tesla talked about only a year or so ago. It is a major overhaul of strategy requiring details.
Tesla is starting to have serious competition:
So consumers — some of whom are turned off by Musk’s incessant posting on X, the social platform he owns, and by his controversial political comments — have a lot of choices when it comes to buying an electric car. Tesla’s share of the EV market in the US was roughly 51% in the first quarter, Cox says, down from almost 62% a year earlier.

The competition is even fiercer outside the US, where Chinese carmakers dominate. About half of all EVs sold globally are Chinese brands — BYD, the top brand within China, sold more cars than Tesla did in the last quarter of 2023, though Tesla regained the lead in the following quarter.
To respond to this competition, Tesla has understood for a long time that they needed a $25K Model 2:
Musk first teased about such a car in September 2020, saying a series of innovations Tesla was working on would enable it to make an EV at that price within about three years. As recently as January, Musk said Tesla was “very far along” with work on its lower-cost vehicle.
But as always, Musk's schedule was just a fantasy, and then the need to pump the stock took over:
Then, in early April, Reuters reported that Tesla had shelved plans for the cheaper vehicle to prioritize its robotaxi, creating bedlam among investors. The tension within Tesla over Musk’s desire to focus on the robotaxi is nothing new. It was chronicled by Walter Isaacson, who wrote in his book published in September that the billionaire had “repeatedly vetoed” plans to make a less-expensive model. Musk refused to give any details about a new, more-affordable model when asked about them by analysts on the first-quarter call.
My guess is that it has dawned on Tesla that, without the resources sunk into the Cybertruck, they simply can't build a $25K car and make money, unlike the competition:
China’s EV advantage is in batteries — the most expensive part of an EV. They’re much cheaper in China because of the country’s control of the mining and processing of component materials such as lithium, cobalt, manganese and rare earth metals. UBS analysts say BYD had a 25% cost advantage over North American and European brands in 2023. Its cheapest model goes for $10,000. Tesla’s cheapest Model Y — the world’s best-selling car of any kind last year — is about $35,000 in the US after accounting for federal tax credits.
China's other advantage is in driver assistance technology:
“Chinese EVs are simply evolving at a far faster pace than Tesla,” agrees Shanghai-based automotive journalist and WIRED contributor Mark Andrews, who tested the driver assistance tech available on the roads in China. The US-listed trio of Xpeng, Nio, and Li Auto offer better-than-Tesla “driving assistance features” that rely heavily on lidar sensors, a technology that Musk previously dismissed, but which Tesla is now said to be testing.

The Robotaxi Rescue

According to Musk the thing that will transform Tesla's profitability is a robotaxi. Lets assume for the moment that, despite being dependent only upon cameras, Tesla's Fake Self Driving actually worked. In Robotaxi Economics I analyzed the New York Times' reporting on Waymo and Cruise robotaxis in San Francisco and concluded:
These numbers look even worse for Tesla. Last year Matthew Loh reported that Elon Musk says the difference between Tesla being 'worth a lot of money or worth basically zero' all comes down to solving self-driving technology, and the reason was that owners would rent out their Teslas as robotaxis when they weren't using them. This was always obviously a stupid idea; who wants drunkards home-bound from the pub throwing up on their Tesla's seats? But the fact that the numbers don't add up for robotaxis in general, and the fact that Hertz is scaling back its EV ambitions because its Teslas keep getting damaged because half of them are being used by Uber drivers as taxis, make the idea even more laughable.
Even for Waymo, it turns out that replacing a low-wage human with a lot of very expensive technology (Waymo's robotaxis "are worth as much as $200,000"), and higher-paid support staff isn't a path to profitability.

It is true that Tesla's robotaxis would be cheaper than Waymo's, since they won't have the lidar and radar and so on. But these things are what make the difference between Waymo's safety record, which is good enough that regulators allow them to carry passengers, and Tesla's safety record, which is unlikely to impress the regulators.

The regulators have a lot of reasons to be skeptical. Back in 2021 they started investigating Autopilot:
The U.S. government has opened a formal investigation into Tesla’s Autopilot partially automated driving system after a series of collisions with parked emergency vehicles.
...
NHTSA says it has identified 11 crashes since 2018 in which Teslas on Autopilot or Traffic Aware Cruise Control have hit vehicles at scenes where first responders have used flashing lights, flares, an illuminated arrow board or cones warning of hazards.
Since then the evidence has piled up, as the Washington Post team report:
At least eight lawsuits headed to trial in the coming year — including two that haven’t been previously reported — involve fatal or otherwise serious crashes that occurred while the driver was allegedly relying on Autopilot. The complaints argue that Tesla exaggerated the capabilities of the feature, which controls steering, speed and other actions typically left to the driver. As a result, the lawsuits claim, the company created a false sense of complacency that led the drivers to tragedy.
Musk claimed they would never settle these cases, but:
Tesla this month settled a high-profile case in Northern California that claimed Autopilot played a role in the fatal crash of an Apple engineer, Walter Huang. The company’s decision to settle with Huang’s family — along with a ruling from a Florida judge concluding that Tesla had “knowledge” that its technology was “flawed” under certain conditions — is giving fresh momentum to cases once seen as long shots, legal experts said.
The regulators move slowly but they keep moving:
Meanwhile, federal regulators appear increasingly sympathetic to claims that Tesla oversells its technology and misleads drivers. Even the decision to call the software Autopilot “elicits the idea of drivers not being in control” and invites “drivers to overly trust the automation,” NHTSA said Thursday, revealing that a two-year investigation into Autopilot had identified 467 crashes linked to the technology, 13 of them fatal.
Last December, the NHTSA forced Tesla to recall more than 2M vehicles because Autopilot:
has inadequate driver monitoring and that the system could lead to "foreseeable misuse,"
The agency suspects the recall wasn't adequate:
The National Highway Traffic Safety Administration disclosed Friday that it’s opened a query into the Autopilot recall Tesla conducted in December. The agency is concerned as to whether the company’s remedy was sufficient, in part due to 20 crashes that have occurred involving vehicles that received Tesla’s over-the-air software update.
The recall involved an over-the-air update, but Tesla's attitude to regulation showed through:
the agency writes that "Tesla has stated that a portion of the remedy both requires the owner to opt in and allows a driver to readily reverse it" and wants to know why subsequent updates have addressed problems that should have been fixed with the December recall.
What is the point of a safety recall that is opt-in and reversible? Clearly, it is to avoid denting the credibility of the hype. The NHTSA is not happy:
In a separate filing, NHTSA detailed findings from its investigation that preceded the December recall. The agency found that Autopilot didn’t sufficiently ensure drivers stayed engaged in the task of driving, and that Autopilot invited drivers to be overconfident in the system’s capabilities. Those factors led to foreseeable misuse and avoidable crashes, at least 13 of which involved one or more fatalities, according to the report.

“Tesla’s weak driver-engagement system was not appropriate for Autopilot’s permissive operating capabilities,” NHTSA said. This resulted in a “critical safety gap” between drivers’ expectations and the system’s actual capabilities, according to the agency.
The NHTSA is skeptical that the recall was effective:
But NHTSA says it knows of at least 20 crashes involving Tesla Autopilot that fall into three different categories. It says there have been nine cases of a Tesla having a frontal collision with another vehicle, object, or person, for which there was time for an alert driver to have avoided the crash. Another six crashes occurred when Teslas operating under Autopilot lost control and spun out or understeered into something in a low-grip environment. And five more crashes occurred when the driver inadvertently canceled the steering component of Autopilot without disengaging the adaptive cruise control.

NHTSA also says it tested the post-recall system at its Vehicle Research and Test Center in Ohio and that it "was unable to identify a difference in the initiation of the driver warning cascade between pre-remedy and post-remedy (camera obscured) conditions," referring to the supposedly stronger driver monitoring.
The agency is giving Tesla until July 1st:
to send NHTSA a lot of data, including a database with information for every car it has sold or leased in the US, with information on the number and dates of all Autopilot driver warnings, disengagements, and suspensions for each of those vehicles. (There are currently more than 2 million Teslas on the road in the US.)

Tesla must also provide the cumulative mileage covered by Autopilot, both before and after the recall. NHTSA wants Tesla to explain why it filed an official Part 573 Safety Recall Notice, "including all supporting engineering and safety assessment evidence." NHTSA also wants to know why any non-recall update was not part of the recall in the first place.
Finally, Mike Spector and Chris Prentice report that In Tesla Autopilot probe, US prosecutors focus on securities, wire fraud:
U.S. prosecutors are examining whether Tesla committed securities or wire fraud by misleading investors and consumers about its electric vehicles’ self-driving capabilities, three people familiar with the matter told Reuters.
...
Reuters exclusively reported the U.S. criminal investigation into Tesla in October 2022, and is now the first to report the specific criminal liability federal prosecutors are examining.

Investigators are exploring whether Tesla committed wire fraud, which involves deception in interstate communications, by misleading consumers about its driver-assistance systems, the sources said. They are also examining whether Tesla committed securities fraud by deceiving investors, two of the sources said.

The Securities and Exchange Commission is also investigating Tesla’s representations about driver-assistance systems to investors, one of the people said.
This is all about Autopilot, but Fake Self Driving has problems too, as the Washington Post team reported in Tesla worker killed in fiery crash may be first ‘Full Self-Driving’ fatality:
Two years ago, a Tesla shareholder tweeted that there “has not been one accident or injury” involving Full Self-Driving, to which Musk responded: “Correct.” But if that was accurate at the time, it no longer appears to be so. A Tesla driver who caused an eight-car pileup with multiple injuries on the San Francisco-Oakland Bay Bridge in 2022 told police he was using Full Self-Driving. And The Post has linked the technology to at least two serious crashes, including the one that killed von Ohain.
The regulators still approve Waymo's cautious and well-engineered robotaxi effort. Uber's and Cruise's robotaxi efforts flamed out. Given the lack of sensors, the history of crashes, the fact that their "autonomy" technology is still at level 2, and the resistance to regulation, why would any regulator approve even the testing, let alone the revenue service of a Tesla robotaxi?

After Robotaxis, What?

Now that the effectiveness of the robotaxi hype is starting to fade, it is time for Musk to roll out the next shiny object. Dan Robinson reports on it in Elon Musk's latest brainfart is to turn Tesla cars into AWS on wheels:
EV carmaker Tesla is considering a wonderful money-making wheeze – use all of that compute power in its vehicles to process workloads for cash, like a kind of AWS on wheels.

The Elon Musk-led outfit said in its recent earnings conference call for calendar Q1 that it had noticed its vehicles spend a considerable amount of their time just sitting there not moving. Many pack in a decent amount of processing power, so why not get them to do something useful and earn some cash for the company as well?

Speaking on the conference call, Musk said that he thought most Teslas were probably used for about a third of the hours in a week.
Seriously? Unless you're a gig worker for Uber or Lyft, who clocks 56 hours/week sitting behind the wheel? I can't believe that Musk is under-estimating the potential here:
"And now that we have already paid for this compute in these cars, it might be wise to use them and not let them be, like, buying a lot of expensive machinery and leaving to them idle. We don't want that. We want to use the computer as much as possible and close to like basically 100 percent of the time to make full use of it," Elluswamy said.

"It takes a lot of intelligence to drive the car anyway. And when it's not driving the car, you just put this intelligence to other uses, solving scientific problems like a human or answering dumb questions for someone else," he added.
...
"If you get, like, to the 100 million vehicle level, which I think we will at some point get to, and you've got a kilowatt of usable compute – I think you could have on the order of 100 gigawatts of useful compute, which might be more than anyone, more than any company, probably more than any company," he mused.
Tesla is currently selling around 2M vehicles/year, so "at some point" will be sometime in the 2070s, by which time the vast majority of the vehicles Tesla has shipped will have been scrapped, and even if they still work 50 years of Moore's law will have made all but the last few obsolete.

Robinson starts thinking about the details:
Of course, all this compute capacity isn't sitting conveniently clustered together in a datacenter. It is distributed here and there, reached via a cellular connection in each Tesla, or possibly via Wi-Fi if the car is on the owner's driveway.

So the model Tesla would be looking at is perhaps more akin to edge computing, such as Heata in the UK, which uses heat from servers in homes to provide domestic hot water and rents out the compute capacity via cloud company Civo.

Among the issues we can see is that Tesla would be effectively using electricity that the car owner has paid for to run any workloads while it is idle, so would they get a cut of the money generated?

Yes, it seems. CFO Vaibhav Taneja, said "the capex is shared by the entire world. Sort of everyone owns a small chunk, and they get a small profit out of it maybe."
...
IDC Senior Research Director for Digital Infrastructure Andrew Buss said the idea sounds technically feasible, but the potential downsides are perhaps too big to justify it being actually implemented.

"They'd not even be edge processing nodes as the code and data would have to be centrally managed and stored and then packaged and sent for processing before being returned once complete," he told The Register.

Other downsides include third-party code and data running on a private asset, Buss said, and if taking power from the battery, this would accelerate the degradation of these, which are the single most expensive and crucial part of a Tesla and need to be kept in as optimal a shape as possible for longevity and consistency of range.

In other words, Tesla might well find that implementing this idea may prove more trouble than it is actually worth for the returns it generates.

And as The Register noted after the earnings conference, Elon has a habit of throwing out wild ideas when things aren't going well to distract the punters and energize investors. This could well be one of them.

20 comments:

David. said...

Ananya Bhattacharya reports that Indians who pre-ordered Teslas in 2016 are giving up and chasing refunds:

"Within hours after the company opened its pre-booking portal for India in April 2016, Gondal, a fan of Elon Musk and Tesla, paid $1,000 (then 66,237 rupees) to pre-book the Tesla Model 3. ...
For almost seven years after that, Gondal, founder and CEO of health-tech startup GoQii, patiently waited for his Model 3. By 2023, when there was still no sign of the car, he decided to cancel his booking — and that was the start of another ordeal.

Gondal had to chase Tesla’s India executive over emails for six months before he received his refund in June 2023. “There was no communication, no emails. And even years later, there was no apology [from the company],” he told Rest of World.

Tesla still does not sell its cars in India — the third-largest auto market in the world."

David. said...

Bloomberg reports that Tesla Ramps Up Job Cuts in China as Sales Slowdown Bites:

"Additional layoffs began earlier this week, extending cuts in mid-April that were part of the electric vehicle maker’s pledge to slash global headcount by more than 10%, said the people, who asked not to be identified as they’re not authorized to disclose the information publicly.

The latest move affects a range of departments including customer service staff, engineers, production line workers, and the logistics team at Tesla’s Shanghai plant — home to more than half the company’s global production, according to the people. The layoffs last month more directly impacted sales representatives, they said."

Will the last person leaving TESLA - turn out the lights.

David. said...

The conclusion of Kevin Williams' long and detailed I Went To China And Drove A Dozen Electric Cars. Western Automakers Are Cooked is:

"If the U.S. and Europe get what they want—a crackdown on Chinese imports—it doesn’t feel like it would result in better cars. It feels like it would keep buyers of those markets locked to cars that aren’t executed as well. It’s nakedly protectionist because deep down, all of the Western auto executives and some hawkish China pundits understand that Chinese EV and PHEV models are more compelling than what European, other Asian, and American brands have come up with.

I’ve seen it with my own two eyes. We’re cooked."

David. said...

Zeyi Yang's How China is regulating robotaxis isn't good news for Tesla:

"American and Chinese robotaxi companies are both under pressure to start making money. Considering the advanced hardware and software in a robotaxi and the human intervention still needed to operate and maintain them, robotaxi rides have much higher costs than taxi rides today, which makes the business hard to scale up.
...
In December 2023, China’s first regulation on commercial operation of autonomous vehicles went into effect. It sets some ground rules for different kinds of vehicles: roboshuttles or robotrucks still need to have in-car safety operators, while robotaxis can use remote operators. The ratio of robotaxis to remote operators cannot exceed 3:1, and operators need to pass certain skill tests. There are also rules specifying what data the companies need to report when accidents happen."

David. said...

Wait, what? Craig Trudell reports that Elon Musk Pledges to Grow Supercharger Business He Just Decimated:

"Elon Musk touted plans to expand Tesla Inc.’s Supercharger network just over a week after firing almost all of the roughly 500 people who ran the business.

Tesla will spend “well over” $500 million on growing its network this year, Musk said on X, the social media network he owns. Ten days ago, the automaker’s chief executive officer wrote that Tesla planned to add chargers at a slower pace and focus more on uptime and existing locations."

Anyone who believes Musk's statements about the future after all this time is incapable of learning from experience. And, "to decimate" means to kill off one in ten. He fired the entire organization.

David. said...

Reuters reports that Eight hundred protesters attempt to storm German Tesla factory:

"Protesters opposed to expansion of the US electric vehicle maker Tesla’s plant in Grünheide near Berlin clashed with police as some of them attempted to storm the facility on Friday.

Some 800 people took part in the protest, according to the organising group Disrupt Tesla, which claims the expansion would damage the environment."

Tardigrade said...

"And, "to decimate" means to kill off one in ten. He fired the entire organization."

To be fair didn't Musk say he was going to layoff 10% of Tesla? Then they're just part of the 1-in-10 being eliminated. I'd imagine back in the day that anyone talking specifically about the dead Roman soldiers after a decimation would refer to them as having been decimated, and not just reserve this term for the entire cohort.

Re: Robotaxis. It reminds me of the scifi tropes where people are used as cheaper general-purpose computers for tasks. Do autonomous taxi companies have reason to believe that the computational and equipment demands for autonomous taxiing will ever be cheaper in sum on an ongoing basis than human drivers? I guess at the very least there's a market for people who don't feel comfortable driving with another person.

David. said...

1) My comment about decimation referred to the headline, which specifically says that Musk decimated the Supercharger organization, which he completely destroyed.

2) Musk appears to have already done a double-decimation on the whole of Tesla global.

3) Evidence from Cruise and Waymo is that their robotaxis are not making money.

Tardigrade said...

"1) My comment about decimation referred to the headline, which specifically says that Musk decimated the Supercharger organization, which he completely destroyed."

Yes, this was my point. You'd say of those specifically murdered in a decimation that they were decimated. You'd not just say that the cohort was decimated. E.g. when pointing to a mass grave following a decimation the Roman would say to the other Roman that those people had been decimated, instead of some more convoluted circumlocution saying that the legion had been decimated and that those bodies in the mass grave were the 1-in-10 chosen to be killed.

So here, saying that the Supercharger organization within Tesla was decimated is okay in my book as the decimation itself is of the larger organization. Maybe not technically accurate in the Roman way, as Tesla employees were not split into units of 10 to determine who would be eliminated. But it's been a few thousand years since anyone has properly done a Roman decimation, and yet we still use the word.

I used to be hypercritical of this use of decimate as well, but have started to feel more generous given the caveats of the term.

"2) Musk appears to have already done a double-decimation on the whole of Tesla global."

Has Musk finally under promised and over delivered?! Not to make light of this, as it sucks to be terminated. I have been, and lost 15 pounds in 14 days. Immediate terminations deliver a shock to the system that is not healthy. Chronic stress from a gradual phaseout is also bad, but it can be done in such a way that it's healthier than a termination (even with pay). I wish the government would calculate out these costs to the economy like they do other economic shocks. Maybe we'd get some legislation favoring employee-rights, temp agency regulations, and employee-controlled workplaces over oligarchy-controlled workplaces.

3) Not surprising for the reason of R&D testing. Will this remain the same as it leaves the R&D stage? As you indicate this may depend on regulations. Regardless, either costs have to come way down, and I don't know enough about computation and sensors and insurance and regulatory requirements to know if this is possible, or they need to start charging a premium and only sell services to those willing to pay that premium.

David. said...

The "right-sizing" continues. Fred Lambert reports that Tesla’s head of Cybertruck manufacturing is out:

"Tesla’s head of Cybertruck manufacturing has left the company. It’s unclear if he was involved in yet another round of layoffs or if he left on his own accord.
...
At least 10% of the workforce has been let go, but Electrek has heard that as much as 20% of the entire headcount could be gone by the time everything is said and done.

Tesla’s automotive business, including charging and manufacturing, as well as new product launches, took the biggest hit as Elon Musk appears to be transitioning Tesla away from its EV manufacturing roots to focus on autonomous driving products.

Now, the latest Tesla executive to leave is Renjie Zhu, director of manufacturing in charge of Cybertruck production."

David. said...

Jonathan M. Gitlin confirms the brilliance of Musk's long-term strategy in Elon Musk laid off the Tesla Supercharger team; now he’s rehiring them:

"But last week, Musk announced that Tesla would spend more than $500 million building out more chargers, just days after saying the focus would instead be on uptime at existing locations. And to do that, Tesla will need to rehire a whole bunch of people.

That started with Max de Zegher, who was an executive under the previous head of Supercharging, Rebecca Tinucci. (At the time of the layoffs, Electrek reported that Musk got rid of the entire team because its Tinucci did not lay off enough workers on her own.)"

David. said...

How much of a manufacturing nightmare is the Cybertruck? The line is supposed to produce 2500 vehicles/week, but AP reports:

"Tesla is recalling 3,878 of its 2024 Cybertrucks after it discovered that the accelerator pedal can become stuck, potentially causing the vehicle to accelerate unintentionally and increase the risk of a crash.
...
The recall involves model year 2024 Cybertrucks made between Nov. 13, 2023 and April 4, 2024, according to the NHTSA.

That is 3878 vehicles in 20 weeks, or 194/week, or 8% of capacity. Tesla claims they are now producing 1000/week, but even if you believe that the line is running at 40% of capacity and thus losing a whole lot of money.

David. said...

It isn't just Tesla's autonomy that the Feds are skeptical of. Ashley Belanger reports that Feds probe Waymo driverless cars hitting parked cars, drifting into traffic:

"Crashing into parked cars, drifting over into oncoming traffic, intruding into construction zones—all this "unexpected behavior" from Waymo's self-driving vehicles may be violating traffic laws, the US National Highway Traffic Safety Administration (NHTSA) said Monday.

To better understand Waymo's potential safety risks, NHTSA's Office of Defects Investigation (ODI) is now looking into 22 incident reports involving cars equipped with Waymo’s fifth-generation automated driving system. Seventeen incidents involved collisions, but none involved injuries."

And Keith Laing reports that Spate of Self-Driving Probes Points to US Setting Higher Safety Bar:

"The National Highway Traffic Safety Administration initiated two probes just this week into Waymo and Zoox, the driverless technology subsidiaries of Alphabet Inc. and Amazon.com Inc. While the circumstances are different for Tesla Inc. and Ford Motor Co. — their vehicles offer driver-support systems that require constant supervision — what the companies share in common are car crashes that drew the attention of an agency tasked with rooting out safety defects.

The investigations point to a more hands-on approach by a regulator that’s been relatively forbearing of automated-driving systems until recently. While executives and bureaucrats alike have touted the technology’s potential to make roadways safer, NHTSA has steadily accumulated information on how the systems are actually faring on streets and highways. It’s now setting a high bar — in Zoox’s case, for example, the agency initiated a defect probe after just two crashes resulted in minor injuries."

David. said...

Richard Speed reports that Tesla self-driving claims parked in court:

"Tesla is facing a lawsuit over claims made about its self-driving technology after a US judge rejected the company's motion to dismiss the case.

Although Judge Rita Lin dismissed some of the claims, the order [PDF] clears the way for disgruntled Tesla owners to pursue action based on the company's increasingly specific boasts about what its cars are capable of and the cross-country driving abilities that are forever just around the corner."

David. said...

Mike Masnick's Musk’s ExTwitter Fumbles Copyright Law, Loses Data Scraping Lawsuit explains that:

"Internet companies have been pushing to argue that their terms of service can block all kinds of scraping, perhaps relying on the eventual injunction blocking HiQ. Both Meta and ExTwitter sued a scraping company, Bright Data, arguing that its scraping violated their terms of service.

In January, Meta’s case against Bright Data was dismissed at the summary judgment stage. The judge in that case, Edward Chen, found that Meta’s terms of service clearly do not prohibit logged-off scraping of public data.

Now, ExTwitter’s lawsuit against the same company has reached a similar conclusion.

This time, it’s Judge William Alsup, who has dismissed the case for failure to state a claim. Alsup’s decision is a bit more thorough. It highlights that there are two separate issues here: did it violate ExTwitter’s terms of service to access its systems for scraping, and then, separately, to scrape and sell the data."

Judge Alsup writes that Twitter wasn't injured:

"Critically, the instant complaint alleges no such impairment or deprivation. X Corp. parrots elements, reciting that Bright Data’s “acts have caused injury to X Corp. and . . . will cause damage in the form of impaired condition, quality, and value of its servers, technology infrastructure, services, and reputation” (Amd. Compl. ¶ 102). Its lone deviation from that parroting — a conclusory statement that Bright Data’s “acts have diminished the server capacity that X Corp. can devote to its legitimate users” — fails to move the needle (Amd. Compl. ¶ 98). To say nothing of the fact that, as alleged, Bright Data and its customers are legitimate X users (subject to the Terms), the scraping tools and services they use are reliant on X Corp.’s servers functioning exactly as intended."

David. said...

Kevin T. Dugan adds to the downward price pressure with Who Wants 30,000 Used Teslas?:

"At the start of the year, after Hertz announced it was selling off its fleet of Teslas — backtracking on a plan to buy up 100,000 of the electric vehicles — the news sounded good for Bijay Pandey, a 34-year-old self-employed data worker in Irving, Texas. “I have another vehicle, and I was trying to add one for my wife because gas prices were too high,” he said. When he found out that it came with a $4,000 tax credit — even better. “That’s what attracted me,” he added. So, the day after Valentine’s Day, he bought a red 2022 Long Range Model 3 with 70,000 miles on it. It ended up costing just about $25,000, not a bad deal for a car that can sell for about $47,000 new."

There were a whole set of problems:

"Hertz said that it would swap the car for Pandey, but for about two months he waited — making $500 payments on his auto loan — before getting a replacement. “I realized why they were trying to get rid of those Teslas,” he said. “If anything happens to a Tesla, then the bill is too high.”

David. said...

Edward Ludlow got a free 30-day trial of Fake Self Driving and reports on the experience in Tesla’s ‘Full Self-Driving’ Struggled Soon After Leaving My Driveway. He notes:

"I drove several hundred miles with this system engaged, and while I lived to tell about it and was impressed at times, I can confirm that it doesn’t live up to the name. In fact, FSD tended to have trouble assisting me almost immediately after leaving my driveway. ... The system doesn’t make vehicles autonomous, and you the driver are required to have your hands on the wheel and eyes on the road at all times.

My test drives ranged from trips to the grocery store, jaunts to Napa and commutes back and forth between my home outside San Francisco and Bloomberg’s bureau on the Embarcadero. I was unable to make the more than 25-mile journey without at least a few disengagements per trip.
...
After trying it out, I have a hard time seeing how Tesla will make the jump from a consumer car you own and must supervise at all times, to a robotaxi that entirely drives itself. Being part of the experiment was interesting, but not all that comforting."

Tardigrade said...

I don't think I'm happy about those scrapping decisions.

Everything posted to social media is, at minimum, copyright both the original poster and licensed to the social media company by the poster. This copyright and license should be inviolate under copyright laws. Fair use exemptions obviously exist for things such as search engines. But it has to be demonstrated that Bright Data's copying (aka scraping) is for a fair use purpose in order for it to be legal.

Google was rightfully sued for similar practices, though I guess the dismissal of Authors Guild, Inc. v. Google, Inc. shows that copyright protection isn't all it's cracked up to be in the US.

I'm curious why they Meta/Musk didn't make a claim that this copying directly impacts their ability to resell the same copyrighted content that they have duly licensed.

David. said...

Awareness of the gap between Tesla and TSLA is spreading even to Bloomberg, see Liam Denning's Musk Loses Autonomy in Race for Tesla Robotaxis:

"No one values autonomy more than Elon Musk. Except, perhaps, other Tesla Inc. bulls. For example, RBC Capital Markets’ base case valuation for Tesla is just over $1 trillion — roughly double the current valuation — of which less than 6% relates to the manufacture and sale of electric vehicles. Some $627 billion of it relates to “robotaxis” — a figure bigger than the entire current market cap.
...
Investors recently added $82 billion to Tesla’s market cap in a single day on news that Chinese officials had granted tentative approval for deploying its driver-assistance systems there. Much as that lifted spirits, what a perfect example of the power that bureaucrats have to make, break or delay robotaxi ambitions. In pivoting to autonomy, Musk has placed more of Tesla’s fortunes in the hands of public officials and processes beyond the routinely accommodating environs of the boardroom and the stock market."

David. said...

It is no wonder that Hertz is dumping its Teslas. Not merely are they extraordinarily expensive to repair, and their resale value is collapsing, but customaers who rent them are in for a nasty surprise, as James Gilboy reports in Hertz Charging a Tesla Renter for Gas Was Not an Isolated Incident:

"Last week, we reported on a customer who was charged $277 for gasoline his rented Tesla couldn't have possibly used—and now, we've heard from other Hertz customers who say they've been charged even more.
...
Fellow Hertz customer Toan Le reported an even worse experience with their Tesla rental earlier this month. They told us they prepaid $329.83 for a week with a Model 3, and returned the car expecting to pay only $25 for Hertz to charge it. Le was then apparently billed $690.32, some of which was redundant billing for the rental they say they'd already paid for.

To add insult to injury, their invoice (shown here) indicates more than two thirds of that, $475.19, was a fuel charge, which was applied in addition to the $25 charging fee. They also faced a $125.01 "rebill" for using the Supercharger network during their rental, which other Hertz customers have expressed surprise and frustration with. Charging costs can vary, but a 75-percent charge from a Supercharger will often cost in the region of just $15."

The article has other similar stories.