Elon Musk is driving me nuts.
He keeps promising that a breakthrough is just around the corner for Teslas that will not only make fully autonomous driving possible but that will let owners mint money by making their vehicles available as part of a robotaxi network that Tesla will manage. And – this is the really crazy part – many investors seem to believe him even though he’s been making such empty promises for more than eight years.
He’s also now back to talking up the idea of colonizing Mars. He’s said that a million people will be living on Mars in 20 years – 30 years, tops. Even though I realize he’s in his carnival barker guise and is promoting interest in Space X, the sheer stupidity of his claims is breathtaking. We won’t have a single person living on Mars in 20 years, let alone 1 million.
While the connection between Musk’s fantasies and insurance is less direct than I generally like to be in my notes to you, I do think there are a couple of lessons here about how to test claims of coming breakthroughs, including in insurance. I’ll try not to froth at the mouth too much as I lay them out.
The first and more obvious test concerns the credibility of the visionary -- and Musk has none on the autonomous capabilities of Tesla vehicles, in my not-so-humble opinion.
It's easy to see why Musk had loads of credibility in what I think of as his Iron Man days a decade ago. He had co-founded an online city guide in the early days of the first Internet boom and sold it for more than $300 million in 1999. He then got involved early with PayPal and made a fortune. He founded Space X in 2002, with the then-audacious idea that private companies could get into space flight -- and with reusable rockets, no less. That company thrived, too. Musk put up most of a round of financing for Tesla seven months after its founding, became chairman and later won the right in a lawsuit to call himself a co-founder as the company became the leading force in electric vehicles. For good measure, Musk came up with the idea for the "no money down" model for solar installations, which was extremely popular with customers, and helped two cousins launch Solar City.
While he has stumbled more recently with his purchase of Twitter, seemingly on a whim, his Starlink network of satellites for internet access has become a major player in communication.
That record is extraordinary. It makes Musk one of the business titans of at least the last half-century.
But somewhere along the way he decided that selling electric vehicles as a way of saving the planet from climate change wasn't enough of a hero story. Tesla was also going to lead the way in making cars autonomous.
With each passing year, it's become more obvious that Musk's claims have been wildly inflated or even false, but he can't seem to back away, whether because of ego or because so much of Tesla's stock value depends on the promised autonomy capabilities.
In April, I wrote that we should ignore Musk's promise of a major robotaxi announcement in early August, and, sure enough, he's now postponed the announcement to October. But two months of delay isn't anywhere near enough time to solve all the problems with Musk's autonomy strategy. (I went into detail in that piece both on Musk's long history of falsely promising that autonomy was within sight and on the bad technology choices he's made, which I believe leave him permanently behind his competition, especially Google's Waymo, so I won't repeat myself here.)
Even if Musk somehow unveils a breakthrough on his self-driving technology in October (and he won't), he's still years away from being able to deliver on his claims that owners of Teslas will be able to make them available as part of a robotaxi fleet.
How will a car get charged once it leaves the owner's control? How will Tesla know it needs to be cleaned once some sandy kid gets in or some adults have a party in the back? Who will clean it? Who's liable when the car is involved in a drug deal? When it's in an accident? What happens when people stop a car from moving by standing around it (or simply by placing orange traffic cones on the hood, as some anti-AV residents in San Francisco learned they could do)? How do you position cars around a city while they wait to be summoned, so the summoner only waits a few minutes? How do you handle all the payment logistics with cars designed for private, individual usage?
There are more questions, too, all of them much more easily handled by a company owning the fleet, a la Waymo or General Motors' Cruise unit, than by a company managing a fleet of hundreds of thousands of privately owned vehicles.
Musk's overblown claims are catching up to him, according to a Washington Post article. It says:
"The U.S. Justice Department is probing the company’s marketing of both Full Self-Driving and Autopilot, Tesla’s advanced driver-assistance systems. California’s Department of Motor Vehicles is also reviewing those features in light of provisions including a 2022 law prohibiting companies from using marketing and language that would “lead a reasonable person to believe that the feature allows the vehicle to function as an autonomous vehicle.” Tesla has received inquiries from the Securities and Exchange Commission related to its claims to investors, according to news reports and public filings. And a civil lawsuit in California represents drivers who say they were defrauded by the company’s claims [that their cars would earn money as robotaxis] and are seeking refunds and damages over their purchases."
Tesla's lawyers are left to argue that Full Self-Driving doesn't really mean full self-driving and that Autopilot doesn't really mean the car can operate autonomously -- in other words, that drivers have to keep their attention on the road at all times, have to be prepared to take back control at any moment and are responsible for all accidents.
Saying, as Tesla does, that its cars are "self-driving but not autonomous" strikes me as an argument that only a lawyer could love. In the meantime, as the various investigations and lawsuits make their way through our cumbersome legal system, we as consumers and investors need to make our own assessments of whether Musk has any credibility left on his autonomy claims.
The second test is probably more relevant to the insurance industry, given that we tend not to have the sort of splashy icons like Musk, Larry Ellison or Steve Jobs and his "reality distortion field" who can sell an idea just based on personal charisma. That test is based on plausibility and should be based on a "devil's advocate" process designed to find holes in assumptions. (My thinking is based on the process that Chunka Mui and I laid out in "Billion Dollar Lessons," a book we published in 2008 on the lessons to be learned from major business failures, but others have outlined similar processes.)
In Musk's case, it's trivial to find holes in the assumptions he's making about colonizing Mars, as reported in a lengthy recent piece in the New York Times on how he has recently begun concrete planning for his long-held dream.
It took a massive effort to land men on the moon, 240,000 miles away, and no one has done it since, but Musk is planning on transporting 1 million people to Mars, which is 140 MILLION miles away on average. He expects to pull off the most complicated logistical feat in the history of the world without any major problems, not just transporting those people but having each wave keep building out infrastructure to welcome the new arrivals -- and that infrastructure had better be ready, because each flight to Mars would take roughly nine months, and once a group is launched there's no turning back. Musk expects, based on no particular testing to date, that people will thrive in an environment where the atmosphere is 95% carbon dioxide and gravity is only 38% of what it is here on Earth.
Even if the logistics work perfectly, think about the expense. Reports say the cost of putting a pound of something into orbit via Space X can be as low as $1,200, so what would the cost of sending a pound of something to Mars be? Given the kind of rocket that would be needed to get to Mars, the nine months of travel time in each direction and the need for the rocket to be able to land on Mars and return to Earth for the next shipment, I'll spitball an estimate of $250,000 to $500,000 a pound. And we'd need to send hundreds of millions or billions of pounds of material to Mars to construct a Musk city that would house 1 million people.
Finally, ask yourself why we'd go to all this effort. Sure, it'd be great for Space X, but why would the rest of us want to spend all the money and other resources? If we thought Earth would be uninhabitable in the next half-century or so, then, absolutely, but no one is predicting a future that extreme. And even if the forecasts on climate change were that dire, why wouldn't we spend hundreds of trillions of dollars fixing Earth rather than trying to colonize a planet that we've never even been to and that is far less hospitable than Earth has been?
In the insurance world, we've had something of a lesson over the past decade of insurtech about the need to check for plausibility. While I hardly got every call right, I doubted the peer-to-peer model from the get-go because I didn't think there'd be enough capital there to cover anything beyond modest losses. I dabbled a bit in the possibility that a Big Tech player would disrupt insurance but was cautious because of all the regulation in insurance and all the risk capital that's required -- Big Tech avoids regulation at all costs and likes asset-light models, not asset-heavy ones like insurance.
While I think a lot of the irrational exuberance has disappeared from the expectations for insurtech, I'd still hold up Musk's claims about autonomous Teslas and Mars as object lessons. The track record of the person making a claim needs to be inspected carefully -- even if, like Musk, the person has a spectacular record in other areas -- and someone needs to lay out all the ways a plan can go wrong, lest they get swept aside in the enthusiasm that a grand vision can create.
Cheers,
Paul