My debate with a friend about Tesla and Waymo's autonomous driving

My longtime friend Marcelo Lima runs a hedge fund called Heller House. He shares my admiration for Warren Buffett and has been to many Berkshire Hathaway (BRK-B) annual meetings – here's a picture of us from the 2017 meeting:

We do not, however, agree on another of the world's largest stocks – Tesla (TSLA). (Fun fact: Tesla has the 10th highest market cap in the world, followed by Berkshire at 11th.)

For reasons I outlined in Monday's e-mail, I'm skeptical whether Tesla will, in the near future, be able to safely remove the safety drivers from its robotaxis, as Waymo has already done in a number of cities. (See my September 22 e-mail for a video of my first ride in a safety-driverless Waymo in San Francisco.)

Marcelo is much more optimistic and agrees with this post yesterday on social platform X from Tesla CEO Elon Musk:

I've included Tesla on my "Stinky Six" stocks to avoid list, whereas Marcelo is long the stock.

Both of us always want to hear thoughtful critiques of our viewpoints, so we engaged in a spirited debate via text message.

With Marcelo's permission, I'm sharing it below because a) I learned a lot, b) Tesla and Waymo's parent company Alphabet (GOOGL) are two of the most valuable and important companies in the world, and c) I think autonomous, full self-driving ("FSD") is a critically important technology that will benefit humanity greatly (fewer injuries and deaths from accidents, less traffic, and immense time savings and higher quality of life if people can read, relax, watch a video, nap, etc. while their car drives them to their destination).

The conversation started on Saturday when I poked the bear by texting Marcelo this video I shared in Monday's e-mail – a glowing review of Chinese automaker Xiaomi's new electric vehicle. And I included this comment: "Tesla is screwed."

Here's a transcript of the debate that ensued:

Marcelo: I just watched it. Why would you ever believe that?

Me: Generally speaking, when the Chinese make something twice as good for half the cost, the other companies are screwed... That's why...

Marcelo: Wrong. Notice it doesn't even drive itself. And it's not half the cost, it's 20% cheaper. You don't realize this, but non-self-driving cars will be completely obsolete very soon. And it's not twice as good. It's worse than Tesla's Model Y. It doesn't even drive itself. For 20% less. Haha. It's dead on arrival.

Me: I don't think Tesla will be able to achieve full autonomy without light detection and ranging ("lidar") – meaning this is all a Musk con (though he no doubt believes it).

Marcelo: It's already here, Whitney. Look at these videos of my Tesla, with FSD 14.2.1, driving me around Miami, and I never touch the wheel. You are wrong.

Me: No it's not. Tesla is where Waymo was five years ago – it still needs a safety driver. That's the difference between 99% or even 99.9% and the required 99.999999%.

Marcelo: You're pulling random numbers out of thin air. Prove that your "five years" comment is anywhere near the truth.

Me: I rode a Waymo five years ago – see my June 29, 2020 e-mail. It was amazing – but needed a safety driver. Just like Tesla today. When I took a Waymo in September in San Francisco, it didn't have a safety driver.

The difference is that Waymo correctly realized from the beginning that lidar and many other sensors were needed, while Musk said, "My team and I are amazing enough to go without all that expensive stuff."

He has pulled off miracles in the past, so I'd never bet against him (by shorting the stock), but I think he's going to fall flat on his face and have to go back to the drawing board (in terms of hardware on the cars). But I hope I'm wrong – it would be amazing if he can pull this off.

My fear is that, to prop up the stock, he'll pull the safety driver – and one of his cars will kill someone.

You won't want to own the stock that day because suddenly everyone will realize that it's a con and all the cybercab/robotaxi promises are an illusion. And then you're left with a declining automaker trading at 270 times earnings.

Marcelo: You are obviously wrong.

Me: I've been debating this for years (since Musk first promised that every Tesla was equipped to be a robotaxi and would become one within a year... in 2016) with various Tesla bulls.

I assure you it's not obvious. I may be wrong, but a lot of smart people agree with me – just as a lot of smart people agree with you.

Marcelo: It doesn't matter who agrees with you. What matters is if your analysis makes sense or not. And it doesn't.

Me: I have no position in the stock and have told my readers it's a bad short at any price. But you may be falling into one of three traps: 1) "I've made a lot of money with this guy/stock," 2) "I love the product," or 3) "Too long to be wrong." I hope it works out for you and Tesla – but if I'm right, this stock has 80% to 90% downside. As a friend, I urge you to size the position accordingly.

Marcelo: No, Whitney, it has nothing to do with that. Like I said, your analysis is incorrect.

Me: Let's archive this and revisit in a year!

We picked up our conversation yesterday after Marcelo posted this insightful analysis on X (I recommend reading the entire post). It begins:

After I read it, he texted me: "You see, Waymo is clearly screwed, not Tesla."

This led to a continuation of our debate (transcript below)...

Me: You may be right, but don't be so confident. There is nothing clear about this. Musk, as usual, is pursuing a high-risk, high-reward approach that pretty much everyone else in the autonomous-driving field thinks is doomed (relying on camera only). I hope he succeeds, but I would not bet on it (nor would I bet against him).

Marcelo: I have a strong opinion, loosely held, and I change my mind often. Do you ever change yours? You've been wrong on Musk forever – it might be worth asking why...

Note, I am not discussing stock prices here, I'm discussing fundamental technology and fundamental business, that's all.

Don't rely on what others say, rely on your own analysis. It doesn't matter what "pretty much everyone else" thinks. What matters is the reality where the rubber meets the road. I can see the product for myself, I use it daily, and have been testing it for over a year. I've seen the evolution – it's extraordinary and astonishing!

Me: I have the humility to recognize that no matter how hard I study (or how many tweets I read), I will never be an expert. So, 1) I put this in my "too hard" basket, and 2) yes, I rely on other experts.

As for the rubber meeting the road... Tesla, with a safety driver, is where Waymo was when I took an amazing ride in one, with a safety driver, five years ago. Whether Tesla can ever achieve true autonomy with Waymo-level safety (which is far beyond FSD 14.2.1, which I agree is wonderful) with only eight cameras is unknowable at this point. You seem to think it's 99% likely – I think it's 33% likely. I cannot prove I'm right, nor can you.

Marcelo: You keep anchoring on random dates. Tesla is way beyond where Waymo was five years ago, obviously, as you can see from the numbers. Tesla will remove safety drivers in Austin in a few weeks. Cybercab manufacturing is scaling in 2026. Elon wouldn't do this to have billions of dollars sitting idle. Just connect the dots.

Me: I agree that Musk will do this. But I think there's a better-than-even chance (67%?) that it will end in tragedy – there will be some edge case that cameras-only don't pick up that will result in a robotaxi killing someone (see the General Motors Cruise case study), which will result in:

a) A public outcry and media firestorm,

b) Regulators requiring Tesla to put safety drivers back in the robotaxis,

c) Musk once again displaying his overwhelming arrogance and tone-deafness, making things much worse by attacking the regulators, media, and the stupid person who walked out between two cars and was hit and killed, and trying to "win the argument" by citing safety statistics (which nobody will believe because Tesla, unlike Waymo, hasn't released all its data), and

d) Tesla's stock crashing by 30%-plus.

Just connect the dots.

Marcelo: Again, I don't give a damn about the stock – forget the stock! Let's talk about the end state. Even if what you describe happens, which is inevitable (FSD is 8 times safer than the average human driver, but it still crashes). The point is it's safer than a human driver.

The end state is clear: FSD will be mandated just like seatbelts are mandated. It's so obvious.

Me: Stop saying things are "so obvious" – when you say this to someone who disagrees with you, you're saying they're stupid. And it reeks of overconfidence, which is deadly when it comes to investing.

The threshold isn't whether something is "safer than a human driver."

Marcelo: What is the threshold?

Me: Do you think anyone – drivers, regulators, etc. – would tolerate an autonomous system that killed 37,000 Americans a year? I don't know what the threshold is, but it's far higher. 10 times? 100 times? 1,000 times? This is yet another unknowable that makes predicting the future here anything but obvious.

Here's what I do know: GM's Cruise was basically put out of business because of one accident in which it wasn't really at fault and the victim didn't even die!

Waymo hasn't had even one accident where someone was injured, much less someone dying. Now that Waymo has set the bar at this level, how much success do you think Musk is going to have, after one more accident with injuries or deaths, arguing that, "Well, maybe our robotaxis aren't as safe as Waymo's, but they cost less and, hey, they're still much safer than humans!"?

Marcelo: Right now, 37,000 Americans are killed in car accidents per year. If FSD lowered that to 3,700, it would be amazing. It's already there.

Me: 1) If every car were Waymo, it would be – to pick a number out of the air – 370, and 2) you're being rational. Since when are humans (and the politicians and regulators who serve them) rational when it comes to dying in a car accident or having their child run over by a car with no driver?

I've recommended Joby Aviation (JOBY), the leading company for new electric aircraft. A good friend is a senior executive there, I've spent a full day at the company, and I closely follow the mind-bogglingly lengthy, arduous process the Federal Aviation Administration ("FAA") is putting Joby through before it's allowed to fly a single commercial passenger. This is the same process for every new commercial aircraft.

The FAA doesn't have a specific standard, but according to Gemini, it's something like one catastrophic failure per billion flight hours for major transport systems, translating to roughly one accident per millions of flights, with industry averages showing modern commercial safety at about 1 fatal accident for every 5 million to 16 million flights.

Joby can't come in and say, "Hey, our air taxi is 10 times (or 100 times) safer than a regular taxi, so approve it." The standard for anything new and different is exponentially higher.

Marcelo: Total red herring argument. There's nothing like Joby out there to compare against.

Me: I think human behavior (and resulting behavior by politicians and regulators) regarding the safety levels that will be demanded of new technologies will prove to be similar. Waymo understands this. I doubt Musk does. His IQ is off the charts, but his EQ is very low. It's a major risk factor I don't think you're appreciating.

Marcelo: You are very wrong that his EQ is low. It is, in fact, the complete opposite – he will be remembered as humanity's greatest philanthropist (literally: love of humanity). Everything he does is for the love of humanity, every project, every company, literally.

Did you listen to what Elon said about safety in the latest third-quarter earnings call? Here's the key part:

Within a few months, we expect to have no safety drivers at all, at least in parts of Austin. We're obviously being very cautious about the deployment. Our goal is to be actually paranoid about deployment because obviously even one accident will be front-page headline news worldwide. It is better for us to take a cautious approach here.

Ashok Elluswamy, vice president of autopilot and AI software, added:

We continue to operate our fleet in Austin without anyone in the driver's seat, and we have covered more than a quarter million miles with that. In the Bay Area, where we still have a person in the driver's seat because of the regulations, we cross more than a million miles. We continue to see that the robotaxi fleet works really well. Customers are really happy, and there's no notable issues. On the customer side, customers have used full self-driving supervised for a total of 6 billion miles as of [October 21, 2025]. That's a big milestone. Overall, the safety continues to be very good.

That's where Marcelo and I left off our conversation for now. I hope you enjoyed reading it as much as we enjoyed engaging in it!

Best regards,

Whitney

P.S. I welcome your feedback – send me an e-mail by clicking here.