Wednesday ORT: Lewis Mopes, $73 DMV, BRG ND, Jeep Recon, 10% Chance Of Tragedy

All readers welcome

Good afternoon friends. Plenty of topics, let’s get right to it.

“Why can’t you just hire Valtteri and make him drive around behind me?”

It’s easy to be annoyed with the current Formula 1 machinery, a point ably underlined by Sebastian Vettel’s nostalgic run at Monza in Michael Schumacher’s F2002. He turned a 1:33.9 after starting with a series of donuts. It’s commonly believed that a standing start costs an F1 car about 8-10 seconds in laptime, which puts this demo run by a completely retired driver in a car yanked from the historic fleet an uncomfortable two or three seconds away from Pierre Gasly’s brilliant pole lap. With a little bit of preparation, the F2002 on slicks is probably faster than the current hardware, while looking and sounding infinitely more interesting. Vettel’s ability to show up and turn up also underscores the mild tragedy of his early retirement; both Hamilton and Alonso are older than he is, and both were already in F1 when Vettel made his debut at the USGP.

On the other hand, watching Sir Lewis continue to flounder in the red Ferrari offers no small argument for Vettel’s choice. You wouldn’t know it to listen to the announcers, but Sir Lewis has been outqualified by his teammate so far in 2026 and also overall in his Ferrari career and also overall since 2024. Leclerc also has a higher average finishing position in the races, which has been obscured by his three DNFs so far this season. At the start of Monza’s grand prix, placed by the grace of Oscar Piastri’s penalty behind the frequently brilliant Pierre Gasly and the dolorous “Princess” George Russell, Lewis pushed the issue with his teammate going into turn one only to be pushed off the outside in response a few meters later. Afterwards, Leclerc suggested that both drivers had “gone too far”, while Lewis demanded “rules of engagement” going forward before claiming, hilariously, that he was known for clean driving.

Leclerc did Lewis the favor of crashing out in the Parabolica, but despite his usual run of luck regarding competitors and safety cars and the like, our British Knight only made it up to sixth. This reflected his pace, if not the pace of the car.

Ferrari is in a tough spot here. They paid a lot of money for Lewis, a decision that will never make sense to me and is rendered even more ridiculous by the way Carlos Sainz is performing in a truly awful Williams. Lewis is in a bad mood because his decision to leave Mercedes looks worse every day and now a 19-year-old is showing the world how easy it is to beat George Russell, something that proved beyond Hamilton’s capabilities a full three years of aging ago. So Fred Vasseur has two choices:

0. Start giving team orders to favor Sir Lewis explicitly, which will infuriate Charles Leclerc and likely result in Ferrari having no veteran drivers two years from now;

1. Leave things as they are, knowing that Lewis will push the issue with Charles every chance he gets and then cry to the press about it, which makes him look like a complete and utter waste of money.

The obvious thing to do here: take Lewis Hamilton’s $100 million annual contract, cross out the words “Lewis Hamilton”, write in the words “Kimi Antonelli”, and take a genuine shot at winning the WDC with an Italian driver in an Italian car.

Other notes:

• Justice for Yuki! with a first-rate, highly-competitive drive from 15th to 10th. With the right tire strategy, it could have been a drive from 15th to 8th, but RB decided to give the primo options to Lindblad instead.

• Justice for Liam! as the overhyped and arrogant Australian fails to impress in the Red Bull. Every moment he spends in the car makes Isack Hadjar’s career look like more of a sure bet. It’s like the reverse of when Lou Gehrig replaced Wally Pipp.

• I think Alex Albon is a fantastic person and I appreciate his commitment to housecats but he should find a new home in IndyCar or possibly F3. Sainz is just taking him to the woodshed constantly.

• Someone should put George Russell on suicide watch before it’s too late. He is learning what it’s like to drive next to a generation talent.

• Speaking of generational talents… Max didn’t make Kimi work very hard for that pass, did he? Yet another example of mentorship and kindness from the driver who would be a five-time WDC had Kimi not hit him last year.

• Toto Wolff was visibly furious when Gasly stole pole position, but how angry can he be? It’s a Mercedes engine, and it would take an alien invasion to keep him from winning both team and driver championships this year.

• It’s not that I like Lando’s chances against Lewis this year. More like I like Lando’s chances against George and Lewis. He is driving like a champion.

You can pull the lever, and the trolley will miss all the people, but the DMV will have to fire someone

California gubernatorial candidate Steve Hilton proposes to shutter the California DMV, to reduce vehicle registration from the $400-plus most motorists are paying back down to the basic fee of $73, and to turn California’s 170 DMV locations into after-school club facilities. “You’re being ripped off by your government. They are gouging you with the DMV and vehicle registration. They charge the highest vehicle registration in the country for the worst roads in the country,” he notes.

The California DMV currently employs approximately, 8,300 people for the sole purpose of making things worse. That’s not quite fair; they also handle vehicle titles and a few other necessary registration tasks. Still, of all the governmental functions that could be more highly automated, vehicle registration has to be at or near the top in terms of payoff for effort. Not to mention the amount of work and productivity lost in the state as a result of having to visit a DMV location.

I truly enjoy telling my California friends about what I have to do in my country to title and register a vehicle. First, I drive 21 minutes to the county seat and walk into the title office. They have three employees who all know me, and there is never a line. Then I walk over to the BMV. They have five employees who all know me, and the line is never more than five minutes. Then I drive home. Even out-of-state registrations and title transfers rarely take me more than one hour door to door.

This wouldn’t be practical for Los Angeles, but surely there’s a better way to do things.

They could, but they won’t

Former TTAC supremo Derek and I were discussing a very nice AI photochop of a “Miata Special Edition” based on the current ND3. The 1991 Special Edition was the first of many Miatas to be distinguished from its lotmates by a color and a couple pieces of equipment:

Rather than run from the people who criticized the Miata as “nothing but a knockoff Elan”, Mazda leaned in with a British Racing Green and deep tan color combo, then tossed in a Nardi wheel to make the rest of the point, which was: we are all that is left of the British roadster.

Your humble author loved the car enough that he built and painted a 1/24 scale model of it at the too-old-to-be-making-car-models age of nineteen. The Special Editions were always a joy to see on the road. While I have nothing against the 30AE (obviously) or the purple-ish 35AE, the time has long arrived for Mazda to do another BRG car. Something sobering to consider: we are now on year thirty-seven of Mazda Miatas. The span of time between the MG TC and the rubber-bumper version of the MGB was… twenty-nine years. Miata production slightly exceeds MG production from 1945 to 1980 overall. The Miata is the most-raced production car in America.

In other words, Mazda has earned the crown and the color. Call it Hiroshima Racing Green, why dontcha?

Late to the party

We’ve had plenty of stupid vehicle introductions over the last few years, but the Jeep Recon is probably the worst yet. It costs north of $65k, has 670 horsepower, weighs more than 6,100 pounds, and is “Trail Rated”.

The various second-tier autowriters assigned to Jeep’s Moab luxury-trip reveal of this thing are all oddly astounded that the Recon is good off-road. It should be good off road. An EV has almost insurmountable advantages in off-road driving compared to any internal-combustion vehicle, because it should be able to precisely apply the right amount of torque to all four wheels individually. The Recon doesn’t quite do that because it has a pair of motors feeding gearboxes rather than four motors, but even being able to divide the motive unit into two should and does pay significant dividends.

The chassis is much stiffer than that of the Wrangler and it can drive doors-off without squeaking. Nevertheless, it’s fundamentally a stupid idea because off-road trails are the places in America least likely to have a charging station nearby, and part of being off-road is that you can’t really be exact as to your schedule or timing.

This will make a great way to get up a fire road to a rural home, but the same thing can be said about a 2008 Camry on 3-peak-mountain-snowflake tires. I think we will remember the Recon as the last gasp of a genuinely moronic period of time in industry/government interactions. Its existence is not a victimless crime because this kind of development money and effort probably could have gotten us a two-seat roadster or even a Recon powered by a Pentastar, but as the nice fellows in Colony House say, you gotta leave what’s lost behind.

There’s a limit to artificial intelligence, but no limit to human stupidity

One question that the readers did not ask in my tale of how my wife left me for a Zoox was the most obvious one: “Why is she so attracted to Zoox?” Part of it is merely comparative: your humble author is a genuinely unpleasant-looking person with no two limbs the same length and the deadest eyes this side of a Mako, so even a transportation pod looks handsome by comparison. The rest of it, however, is something called pareidolia, the tendency to see human faces and features in non-sentient objects. This cartoon, which I whipped up to annoy my wife during our flight, expresses it very well:

It was taken from a real-life photo that is hardly any less cutesy. We can easily see human characteristics in Zoox because we see them everywhere.

This is not limited to visuals. Way back in the computing stone age, people found some humanity in ELIZA, a 420-lines-of-code proof of concept from MIT. Nowadays it takes more than 420 lines of code to update a shopping cart on a Web app, if that puts things in perspective. It is human nature to see human qualities in inhuman objects. Which is why the average ChatGPT user feels so comfortable asking it how to strangle three children and get away with it, and so on.

Couple this with another human trait — wanting desperately for something to be true, and ignoring all evidence to the contrary — and you get the bizarre situation on X at the moment where a junior Anthropic employee claims to be quitting because he thinks AI will destroy the world and a senior Anthropic person says, “Yeah, he is right, and we estimate the chances at more than ten percent.”

To put this in perspective, at no point in modern history has there been any kind of consensus that anything had a better-than-ten-percent chance of annihilating humanity. No one felt that way about nuclear war; it was always assumed that some people would survive even the stoutest exchange of hostilities between the USA and USSR. No one thinks climate change will eradicate humanity. No one thinks a “superbug” will eradicate humanity. Yet you have a senior researcher at Anthropic throwing the figure around and all but putting “LOL” after the claim.

I certainly don’t think AI will become “superintelligent” and kill us all, ever. Moravec’s principle will keep some of us around for a very long time, even if our purpose is only to serve the machine overlords. I also personally believe, without being able to lay it out as a mathematical proof, that computing hardware as we know it will never generate consciousness. The brain is a neat piece of machinery and it benefits both from extreme density and some quantum-type effects which you cannot duplicate in silicon.

Which is not to say that “AI” cannot harm humanity. It can. It is. It already has harmed humanity, and there is much more harm to come. But the harm is not going to come from some Terminator or Matrix superintelligence. It will come from an overabundance of “2 sigma” people who will earnestly hook an LLM up to everything from autonomous armed drones to a nuclear power plant to the Hoover Dam if they get a chance. The AI won’t kill you with a chrome-skeleton robot; it will kill you because some idiot lets it oversee next year’s wheat-seed production. It won’t send out a red-eyed sentient airplane to end your life; it will accidentally overload part of America’s already strained East Coast grid and plunge 75 million people into a ten-day experiment in heatstroke and/or broad-scale residential looting.

What I want to suggest to you is that the prospect of the “AI Supervillain” is being continually held out in front of you for two reasons:

0. Darth Vader is cool and powerful, so you’re more likely to respect and use AI if you think it is Darth Vader;

1. as long as Anthropic et al manfully “stave off” the Matrix Architect, they aren’t answerable for everything else their product does to destroy your life, like encouraging children to commit suicide and hoovering up all the clean water in your local community.

For the record, Anthropic is not “trying its best” to prevent the AI superintelligent villain, because the people who actually write the stuff know damned well no such thing will ever happen. They give that work to Anthropic’s in-house tech-industry equivalent of Douglas Adams’ “phone sanitizers”. Anthropic has much more pressing matters at hand, like how to prevent the discount Alibaba AI from eating its lunch and rendering it irrelevant.

What the research actually shows is that every single AI product depends hugely on pareidolia to be valuable. The various productions of Chatty et al are treated like Samuel Johnson treated the idea of a woman preaching the Gospel; “The wonder is not that it is done well, but that it is done at all.” An hysterical pronouncement by Tyler Cowen that “AGI” had happened in the o3 model was countered by a caustic comment that “Tyler, maybe the fact that AGI wipes the floor with tasks in your profession says more about your profession more than it does about AGI.” And this one, too:

It should also be reiterated that most of the recent “breakthroughs” in AI are simply a result of figuring out ways to throw more hardware at it. AI has a scaling problem that is likely to prevent the next big leap from ever happening.

Trump, Xi, and the EU mandarins could handle this silliness with a very simple thing: a hard-wired kill switch that goes from their desks to every single data center in their country. Flip the switch and everything goes dark, immediately. I mean, an actual superintelligence could prevent this from happening by using every phone around the world leaders to reprogram their brains via electromagnetic pressure — I’m dead serious about that, human beings are open books to any genuine ultra-being — but since that’s not a real problem, the fake problem could be solved with a kill switch.

The fact that no one wants to spend $100 million of the $3 trillion AI budget on such an easy countermeasure should show you what’s really happening.

(By the way, if you’re reading this in the dark Terminator future of 2065 and you’re currently on the run from a liquid-metal T-1000, please feel free to stop by my property and piss on my grave. I deserve it.)