Tech

The Billionaire Running for US Senate to Ban Elon Musk’s Self-Driving Cars

Dan O’Dowd won’t talk about housing, inflation, or COVID-19—just Tesla and “cyber Armageddon.”
Screen Shot 2022-04-29 at 2
(Photo courtesy of Dan O’Dowd's campaign)

California Senate candidate Dan O’Dowd will not talk about taxes. Or homelessness. Or climate change. Or inflation. Or housing. Or jobs. All he will talk about is how much he hates Tesla’s self-driving cars and the existential threat computers pose to humanity.

“My issue is more important than all of them, because it's basically about survival,” he told Motherboard. “When cyber Armageddon hits, and everything goes down, I don't think anybody’s gonna care about taxes.”

Advertisement

It’s not an exaggeration to say that O’Dowd has become obsessed with Tesla’s full self-driving cars. He owns three Teslas and claims to have driven nothing else for over a decade, but believes the full self-driving cars have become so dangerous that they need to be banned immediately. The front page of his website mentions Elon Musk’s car company an impressive 14 times. His multimillion-dollar national ad campaign is dedicated entirely to the topic. His non-profit, The Dawn Project, has calculated Tesla’s self-driving cars malfunction every eight minutes on city streets. And he’s circulating a petition calling on Congress to ban the vehicles immediately. 

O’Dowd, a self-proclaimed “self-made billionaire” from Santa Barbara, has spent the past 40 years running the obscure computer company Green Hills Software and writing software code that he claims “never fails and can't be hacked” for nuclear weapon systems, military encryption systems, the FBI, car companies, and others. Now, he’s the latest in a long line of eccentric and rich Californians to run for office, this time against Senate incumbent Alex Padilla and the typical hodgepodge of others.

To better understand O’Dowd’s two-pronged campaign promise to “make computers safe for humanity and ban Tesla Full Self-Driving,” Motherboard discussed his ideas and the race with O’Dowd. We’ve edited the below interview a bit for clarity and length.

Advertisement

So you're running a single-issue campaign around your desire for regulators to ban Tesla's full self-driving feature? Why are you doing that?
That isn't the whole campaign. That’s only a part of it. The big issue is making computers safe for humanity. What’s happened in the last five or 10 years is that we've been taking everything our life depends on and putting computers in control of it—the power grid, the self-driving cars, water treatment plants, dams, hospitals, voting machines—and much of it has been written with really bad software. Software that was never intended for safety critical systems. Software that crashes. Software that's unreliable. Software that's hackable. 

We're putting basically everybody's life at risk here to someone like Putin pushing a button, and all of a sudden, our power’s all shut down. And we can't charge our electric cars. But we also can't pump gas at a gas station without electricity. We have no cell phones. Everything's down. All the self-driving cars could be taken over and crashed or used as weapons. 

Your campaign website does focus predominantly, at least for now, around Tesla’s self-driving cars—why make that the primary focus to start?
Because it's really the poster child. It is the worst piece of safety-critical software I've ever seen. People say, “That's hyperbole. You're making that up.” I'm not making it up. We studied this. Now, this is not autopilot. Don't be confused by their product named autopilot, which is entirely different. The autopilot works on freeways and it seems to be reasonably good. But this full self-driving product is what you use for city driving, for stoplights and traffic lights and stop signs and all the sorts of complicated things that you get into doing in driving in a city and with a lot of sometimes crazy people too. 

Advertisement

Watch the videos. [Editor’s note: O’Dowd’s campaign ad splices together YouTube footage of people driving Tesla’s full-driving cars.] These are fanboys. These are people who love Tesla and love full self-driving. They will tell you this is not finished software. If you run it for just 10 minutes, it's likely to do something seriously wrong, like run a stop sign or run a stoplight or turn left out of a right-turn lane. It's pretty amazing how often this happens. We measured eight minutes. There's no other product, safety critical or not, I've ever seen that fails every eight minutes. Somebody said, “Well, what about Windows 3.0?” I said, “No. I used Windows 3.0. Yeah, it might crash twice a day. Not every eight minutes. Nothing is that bad.” And this is driving a car! If you don't watch that software every single second like a hawk, it will drive off the road or crash into something. It requires constant supervision. 

Now if these were professional test drivers, and there were 10 of them, or 50 of them, I'd say yeah, sure. It's 100,000 people that have been delivered this software. [Editor’s note: Musk recently said in an interview with TED’s Chris Anderson that the company has “over 100,000 people in our full self-driving beta program.”]

Advertisement

Do you remember when you started to realize that this was a big problem in your eyes and one worth building this campaign around?
Somewhere around seven years ago, it sort of finally dawned on me that all this software was vulnerable, and we were putting the entire society—literally the entire society—at risk. Going back to 1820. No trains, no airplanes, no cars, you know, nothing works. Nothing that requires electricity works. 

With respect to this self-driving thing and Tesla, I didn't know about this until the latter part of last year. Somebody said, “Do you know about this software and how it works? When they drive for 30 minutes, several times, it will do something amazingly dumb.” I looked at these things. I said, “Can this be true?” There are at least five or six companies doing self-driving software right now that is vastly superior to what Tesla's doing. It's comically bad. 

So to clarify, you are not against self-driving cars overall.
And I'm not against Tesla. I own three Teslas. I haven't driven anything but a Tesla for 11 years. And I don't have a problem with autopilot, per se. Those are fine cars, and as far as I know, autopilot’s okay. But this full self-driving—that's the $12,000 option you pay for—that software is terrible. 

Advertisement

I know you've been asked if you have a financial interest in a competing solution. I think those are CNBC words and others have been similarly skeptical.
We do not make cars. We do not make self-driving cars. We do not make self-driving software. But, we do sell software to almost all car companies. We sell people development technology. We show them how to find the bugs in their software. We show them how to make it secure. So people can't hack into it. We show them how to make their software faster. That's what we do for all kinds of software—all the automotive companies, all the aerospace companies, Hewlett Packard. We do inkjet printers and communications routers and military radios. 

I think it's all just a sham to say that somehow we're biased. The people who are really biased are all those people who are saying that. They own Tesla's stock, and they're just trying to boost up the stock price. Anybody who says anything negative about Tesla is threatening their investment. That's why they say what they say, when they go to the irrational lengths that they do to try to stop it until they get to these truly vicious, completely fabricated attacks on everybody who dares to say anything negative. That's the real conflict of interest that's going on.

You're running a state campaign to become California senator, but you've, as far as I understand it, launched a national TV ad campaign. Why make the choice to run the advertisements nationally?
We want to show everybody in the United States—and the world, but right now, let's stick to one country—these problems. These problems affect people in New Jersey as much as they do California. I'm trying to get the message to them that this particular product should be taken off the road. They're saying, “Why would you take it off the road? It's our best hope of saving a million lives a year.” That's their big argument. “You can't slow us down. If you delay Elon Musk from delivering this by making him fill out some useless paperwork, some useless government paperwork for some regulation, and he delays it by one day, shipping the product, then, well, 3,000 people will die and it's your fault.” 

Advertisement

It's totally ridiculous because they're not gonna save any of those million lives. Waymo might, because they have a self-driving car that drives 30,000 miles without crashing. [Editor’s note: This was true in 2020, though Waymo’s car had more issues per mile last year, at least according to data out of the California Department of Motor Vehicles.] Cruise has one. Pony.ai has one. AutoX, they do robo-taxis. No human driving that car. It doesn't need a human babysitter to prevent it from killing somebody. Those are the guys who are going to actually do it. 

Okay. Your ad was titled “Unsafe at Any Speed,” which I assume is an homage to Ralph Nader. Is he an inspiration of yours?
On this, yeah. He said, “There are problems. No one's addressing them. No one's dealing with it. Somebody's gotta deal with it.” And they did. So I am emulating his process here on a larger scale. He wrote a book. I write TV commercials. 

You didn't ask the question. But almost everybody asks why run for Senate. Because it's a political campaign, and there are real votes. I'm not talking about my opponents. oI'm not talking about any other position. I'm not talking about taxes or immigration or all the other things people like to argue about. My position is: This is more important. Securing the U.S. infrastructure against failure and attack is a more important issue. If they take that down, there won't be any more CO2 production. [Editor’s note: No CO2 production would presumably be a net positive for humanity, but you see his broader point.] There'll be no cars running, and there'll be no factories running, and there'll be no power plants. It's over. Okay, that's not an issue.

Advertisement

So you won't talk about any other issue except for this.
My issue is more important than all of them, because it's basically about survival. You won't be paying any taxes when everything goes down. There's gonna be no one to pay them to. All that stuff doesn't matter if somebody wrecks our entire system because we had bad software running it. That's my argument. I think a bunch of people are going to vote for me. I don't know how many, but I think they will. And then maybe, I can go to the politicians and say, “Look, this is what it's worth to you. If you will support my position, you're going to pick up those guys. So I think you should.” 

Surveys show that the most important issues are COVID, homelessness, housing costs, jobs, and inflation. You think that people will put the Tesla self-driving cars above all of those?
Well, I'm talking about the bigger issue. The self-driving cars are just the most egregious one right now. I can't take on everybody at the same time. Well, I don't know. It's an experiment. Can I convince people that this is the most important issue? The issue of betting millions of lives on bad software? We need to make the computers safe for humanity. 

If you were to be invited into a debate, how would you handle that out of interest?
Well, I'm sure they'd say, let's have a section on taxes. Let's all talk about taxes. They'll say we should raise them, and somebody will say we should lower them and change them to some other system. And they'll get to me and I'll say, “You know, when cyber Armageddon hits, and everything goes down, I don't think anybody’s gonna care about taxes. In fact, I don't think there's gonna be any way to pay your taxes. So this issue is just not relevant until we solve my issue.”

Advertisement

On every issue, I'll say the same thing. On national defense, I'll say, “This is the issue of national defense. It doesn't matter how many tanks we have. They’re gonna just bypass our tanks and just kill everybody by shutting down our water treatment plants or by opening the dams or shutting down all the hospitals. They'll do one of those things and then none of this matters. That's our issue of national defense." 

When the enemy can't kill us all, maybe we should make some tanks, some airplanes, some bombs, I don't know. See the point? I can take every issue and say: just doesn't matter until we solve my problem.

Is there particular legislation or regulations that you would kind of push for once you are in office? 
Oh, absolutely. So there are standards for how you test, design, develop software for airplanes. And it's been in place for over 20 years. There have been very, very few airplane crashes, commercial airliners. That process is very thorough in terms of how you do things.

There are more lives at risk with self-driving cars than in airplanes. Not today. But when we look up five or 10 years, there'll be more people in self-driving cars than there are in an airplane at any given time. And a failure could cause more disaster. We should use at least as stringent rules for developing software for self-driving cars. If we do, we will get self-driving cars that are as reliable as airplanes. And the computers will be able to drive very effectively.

Advertisement

But we need to write the software not like they do in Silicon Valley. Tesla writes software following the Silicon Valley model: Move fast and break things. Write software at furious rates, do minimal testing. And I think it was October of last year, October 24, as I recall, of last year [Editor’s note: true], they released a new version of the beta software and almost immediately there was a report from a bunch of those people calling in and saying, “Oh my God, I was driving down the road and [the car] just slammed the brakes on for no reason at all.” Within a short period of time, Tesla realized there was a problem. And they withdrew that release. Then they fixed the bug within hours. They rebuilt the software. They got into a couple of cars—I don't know how many cars, some number of cars, drove around Fremont, California at 70 degrees in the afternoon—and said, “Yeah, it seems to be working.” And shipped it out the next morning to their customers, to 10,000 people. 

Anybody who has any experience in any kind of safety design would be utterly horrified. Instead, they were lauded in the press for their quick reactions, solving the problem so fast and getting it out in the field and not requiring a recall. A safety engineer feels sick when hearing this. Everybody knows when you fix a bug in software, there's a fairly good chance you're gonna break something with your fix. They're fixed in the case, they failed, but now they're broken in something else.

Advertisement

When you change critical safety software, like the brakes in a car, you need to test it again in Nome, Alaska, the Saudi Arabian desert, Pikes Peak at 14,000 feet [in Colorado]. You need to test the new software in all reasonable conditions for a reasonable amount of time—not for a day, at least weeks—before you roll it out to 10,000 people. Oh my God! They keep doing this, one of these days, they're gonna cause a catastrophe. 

They're just slapping it together like they do in Silicon Valley, which works in Silicon Valley. When they screw up, maybe you got to reboot your computer, what's the big deal? Right? All your software crashes or locks up all the time. Silicon Valley methodology is to move fast and break things. Mark Zuckerberg had that written on the wall. It's get something out that works pretty good as fast as you can before anybody else. Grab market share. When the customers complain, fix their bugs and turn it around and turn it around, put more features in. More bugs and take some more bugs out. And it's always buggy, and it never works perfectly, but it works reasonably well. And you make a trillion dollars, right? That’s Silicon Valley. 

But the software that they produce with that methodology has lots of features, and lots of bugs. And we don't want lots of bugs in our cars, which is back to my original thing: making computers safe for humanity. When the computer is controlling 1000s of lives, we need to make sure it's done right and you can't change the software and ship it in one day. That can't ever be allowed to happen.

I know you’re self-funding. How much money are you willing to put into this campaign to get this issue out there?
What it takes. I most likely will not have accomplished all my goals by the end of this campaign. And I'll be doing it again. I'll be doing it again in other forms in the future. 

I read on your website that you’re the world's leading expert in creating software that never fails and can't be hacked. Is it actually possible to create software that never fails and can't be hacked?
I do it every day. And I'm the only one. No one else will even tell you they can do it. I did the operating system for the B-2, the B-52, the F-35, the F-22, the F-16. Military encryption devices, military radios, cell phones. I've been doing this for 25 years. My first project was the intercontinental nuclear bomber. I wrote the operating system in my library. And it controls everything—weapons navigation, flight systems. Basically all the other nuclear weapon systems out there. I have tons of systems out there that have never been hacked. The Russians, the Chinese military—what they want to hack the most is our military and the systems that we built. You can't hack them. 

We don't even think about a feature until we figure out how to make it secure. You say, well, the customer wants a feature. It's like, too bad. They can't have it until we figure out how to make it secure. Totally different mindset than the whole Silicon Valley mindset. And we've been doing that for more than 25 years.

There's lots of candidates, why are you the right person to represent the state of California in the U.S. Senate?
Well, if you believe, as I believe, that this issue of making computers safe for humanity and fixing the problems that we have is the most important problem, you should vote for me. You vote for me and—and everybody, not everybody, but a lot of other people believe that—then I'll fix the problem. And if I don't, I'm gonna keep working on it. 

Are you hoping to win the campaign or are you hoping to bring primarily attention to this issue?
If people really think that that is the most important problem and the majority of the people, then sure, I will use it to solve the problem. But I'm not going to take a convenient position because that's going to pick me up three extra percent

Do you have any thoughts on Elon Musk purchasing Twitter?
I think it’s a hell of a lot of money to pay to shut me up.