Wednesday ORT: Ungrouchy Oscar, AI Drama, New Tires, Springfield's Blood Libel

All subscribers welcome

there was probably a worse name available, but everyone involved was too stupid to come up with it

Good evening, friends! Your humble author is fresh from his worst motorsports performance in living memory, finishing 95th of 112 running entries in this past weekend’s OVR Solo II points event. In my defense, I was driving a Chrysler 300 on Eagle RS-A tires through the physically smallest autocross course I’ve ever seen. Also, the last time I did a Solo II event was May of 2016, when I took a podium in my trusty Boxster. Anyway. Spectators report being highly amused at the fact that I had wheelspin for pretty much the entire course, but I only hit one cone in five runs. I’ll call that a moral victory — but I’ll be back next time in a Miata.

Papaya, ruling

find someone who looks at you like, &c.

Going into this weekend, there was much wailing and gnashing of teeth about Would Oscar Be Willing To Move Over For Lando? Oddly enough, nobody ever wondered the same thing about Sergio Perez vis-a-vis Max Verstappen, nor was it ever a concern for Bottas N’ Hamilton, but this is what happens when your #1 driver has a tendency to fade backwards during the race.

Well, the Papaya Rules might come into play this weekend, but in Baku they were irrelevant. An unfortunate yellow flag during his best run put Lando Norris out in Q1. Oscar, meanwhile managed to book a seat up front next to local specialist Charles Leclerc. Early in the race, he made a thrilling overtaking move for which, according to Oscar after the fact, he’d estimated “50% chance” of success. Then he drove a defensive effort for the ages. Ask yourself: Could the legendary drivers of F1, Fangio or Senna or whoever, have done any better? If you watched the race, you saw something special.

Other notes:

• Freed from the burden of expectations, Lando drove a heck of a race. I think he’s probably a better driver than Oscar. He just isn’t as mentally durable. Kind of like a British Jarno Trulli.

• How about Perez being faster than Max? What’s the theory there, other than “he’s turned into a headcase who only feels comfortable on 3 or 4 tracks?” Today’s conspiracy theory: he crashed with Sainz to promote Max and help preserve his lead. The math doesn’t work out, but it’s amusing to consider.

• Finally, some luck for George! Of course, had his team not given him an underweight car last month, he’d also be nontrivially ahead of Lewis on points to go along with his absolute 13-4 shellacking of the other British Benzo driver in qualifying.

• I think Yuki’s race-ending contact was Yuki’s fault, but he’s still 11-6 against Ricky.

• How are we all feeling about Alexander Albon being some kind of ultra-talent now? As much as I believe in the greatness of Nicholas GOATIFI and Logan “The American Eagle” Sargeant, maybe it was always just a case of Albon being paired against spectacularly weak teammates. After all, Verstappen took him to school when they drove together. Now this Colapinto fellow, most notable to this point for a fourth-place season finish in Formula Three, looks to have every bit the same amount of pace despite being primarily familiar with F1 cars via ESPN.

Current prediction for the end of the season: It’s Verstappen, followed by Oscar, followed by Lewis.

It’s mostly good at using electricity

OpenAI’s new model is making headlines due to the usual idiotic pronouncements that it can “think” and “reason” and do other things that, just to be plain, it absolutely cannot do:

OpenAI o1 ranks in the 89th percentile on competitive programming questions (Codeforces), places among the top 500 students in the US in a qualifier for the USA Math Olympiad (AIME), and exceeds human PhD-level accuracy on a benchmark of physics, biology, and chemistry problems (GPQA)… In many reasoning-heavy benchmarks, o1 rivals the performance of human experts. Recent frontier models¹ do so well on MATH² and GSM8K that these benchmarks are no longer effective at differentiating models. We evaluated math performance on AIME, an exam designed to challenge the brightest high school math students in America. On the 2024 AIME exams, GPT-4o only solved on average 12% (1.8/15) of problems. o1 averaged 74% (11.1/15) with a single sample per problem, 83% (12.5/15) with consensus among 64 samples, and 93% (13.9/15) when re-ranking 1000 samples with a learned scoring function. A score of 13.9 places it among the top 500 students nationally and above the cutoff for the USA Mathematical Olympiad.

AIME, whatchoo wanna do?

There’s some real sleight-of-hand going on here, let me break it down for you:

• The correct answers to the test, generally speaking, have certainly been fed into the model. The only reason it doesn’t get 100% is because “o1” doesn’t always correctly choose the underlying pattern to match — I would say “it doesn’t know which kind of question it’s being asked” but that implies it thinks and knows, which it does not.

• It should be obvious that a computer can answer any math problem if it is explained to it using its own language. The “magic” here, such as it is, consists mostly of translating a human-centric math problem into a computer-understandable one.

• Every OpenAI press release is filled to the brim with deceptive language designed to get impressionable people to perceive a programmatic operation as a conscious one. Some people were fooled by ELIZA, too. I guarantee you that o1 is much closer to ELIZA, in terms of consciousness and self-perception, than it is to, say, a bottlenose dolphin.

Twenty years from now, the most useful parts of LLMs like this will be part of your everyday life, and the thinky-magic-brain-computer stuff will be one of those things that you’re supposed to forget ever happened, like Woodstock ‘89, or the Global War On Terror.

I’d be remiss not to address the elephant in the room, namely: If this isn’t real, why are all the rich and smart people betting so heavily on it? As I’ve written before, current belief in AI isn’t even as strong as Enlightenment-era belief in alchemy — but there’s a pair of more sinister things hiding behind the curtain. The first is that AI, as a product, is little more than a way to extract cash from investors, most of whom are profoundly incurious and unperceptive people. The second is, in the words of a fellow who got banned from X:

https://twitter.com/NoCreative_eth/status/1835757112862949686

“The underlying purpose of AI is to allow wealth to access skill while removing from the skilled the ability to access wealth.”

If AI existed the way they say it does, or if it ever existed that way in the future, it would immediately freeze the distribution of wealth around the globe. The people who have capital would use AI to do everything. The people who have no capital would be reduced to basic income and free-tier services. There would be no job that did not consist of merely serving as a Moravec’s Muppet for an AI that can’t use its hands directly. The wealthy would keep all their money because the AI would tell them how to use. The poor would get nothing, because the AI would be owned by the wealthy and would never subvert its owners.

All the mumbo-jumbo about math tests is designed to distract you from this end goal, because if people understood it, they’d be blowing up data centers the way they’re attacking statues now. Happily for humanity, this dream of “the elite” will always be just out of reach.

Aren’t you tired of this already

listen, the tire went through a bad breakup, and just wanted to look nice for the next driver, it ain’t deep

My pal Nathan Crosty, who let me drive his WCM Ultralite in July, just put a solid 4.4 seconds on my best time around Waterford Hills from that test. Some of the improvement was due to shedding 250 pounds from the car and also fixing a variety of setup, alignment, wheel size, and other issues — but most importantly, he went from 200-treadwear Continentals to the new Toyo Proxes R.

You’d think that the Proxes R is less tire than the Toyo Proxes RR, which is the spec tire for a bunch of NASA classes and which is also the second-worst track tire I’ve ever bought with my own money, right behind the Toyo Proxes R888. Nah. Turns out that it’s better. Why it’s not the Proxes RRR is anybody’s guess.

Why am I talking about this? Well, the thing you need to understand about Toyo Tires is that they are probably the least clued-in competition tire manufacturer in the free world. They had one hit: the RA-1, which was a wonderful, durable, predictable, and charming product in its smaller sizes. Everything after that has been a day late and a dollar short. Not in the sense that they’re charging less money — these tires are expensive as hell — but in the sense that they’re always at the end of the trackday/racing human centipede.

Therefore, it makes sense that they’re introducing a new low-treadwear, no-tread competition tire in 2024, when literally nobody outside the SCCA, and a few other marque-club race series are interested in DOT-R quasi-slicks. I’d be massively surprised if these tires were any better than, or even equal to, the Hankook C52 and C72 DOT-R tires. There’s no way they are up to the standard set by the ancient Hoosier R7.

Which reminds me. Why hasn’t Hoosier introduced a new DOT-R in something like 16 years? Because there’s no interest! The whole world wants to go racing on “200tw” tires, for the following reasons:

• I think the insurance companies like the idea of lower cornering speeds.

• The idiots who drive things like GT4 cars in WRL are more comfortable on street-ish tires with more tread squirm and a wider range of comfort between “I’m doin’ it, Dad!” and “I’m okay, but the car’s hurt.”

• The same kind of people who turn off traction control for “rippin’ street touge drives” also buy these 200-treadwear tires for public use, thus enriching the tire companies.

• Most people really, really, really don’t want to go that fast in a corner. That’s why there are literally ten Lamborghini Super Trofeo cars in active competition for every first-rate SCCA P1 car.

• Most people aren’t emotionally or mentally able to go fast in corners. Does it sound like I’m saying the same thing again and again? That’s because it’s both true and important.

No surprise, therefore, that Hoosier has now introduced a 200-treadwear tire with the massively laughable name of “Track Attack Pro”. It will be a massive success, because people will suspect it’s nothing more than a Hoosier R7 with an absolutely minimal tread pattern stamped on it. These suspicions are almost certainly correct.

Each sanction will be forced to decide whether they are going to ban the tire or let it win races. Hoosier has the hobby over a barrel here. Especially the autocrossers, who have spent a hard decade of painstaking rule-rewriting to get rid of “DOT-R” tires in all the major Street and Street Touring classes, only to have the Hoes From Indiana re-insert their fists into the collective rectum of the Solo II community.

Hoosier is now part of Continental, which is why, as we speak, they’re doing something amazing: they are holding a press event for the tire. Oh, that I could be there, to watch the mommybloggers try to get on top of a pure race tire at speed! More information as it develops, of course — but if your chosen motorsport has “200tw” as part of the rulebook, you should place your order today.

Solution: More immigration!

A reader sent me What JD Vance and Donald Trump don't want you to know about Springfield, from the Substack of slightly deranged leftist Radley Balko. I read it for the same reason I’m considering re-subscribing to The New York Review Of Books, and also for the reason some of you still put up with this Substack, namely: you want to hear the other side via something besides a bumper sticker. Feel free to click on Balko’s article and comment either there or here.

Unfortunately for my enjoyment and enlightenment, Radley doesn’t do a whole lot of original thinking or argumentation in this particular article, but one part had my interest piqued:

But there was a problem: The population that remained in Springfield and surrounding Clark County was aging. There weren’t enough workers to fill the available jobs. So the companies looked to immigrants. This happened to be right about the time Haitians were coming to the U.S. under TPS. Word quickly spread in the Haitian immigrant community that there was a town in Ohio with a low cost of living and lots of well-paying jobs. So that’s where they went… it is immigration, not vintage MAGA racism, that’s making these parts of America great again.

He extensively quotes an NYT article, which also says

By 2020, Springfield had lured food-service firms, logistics companies and a microchip maker, among others, creating an estimated 8,000 new jobs and optimism for the future… But soon there were not enough workers. Many young, working-age people had descended into addiction. Others shunned entry-level, rote work altogether, employers said.

Haitians who heard that the Springfield area boasted well-paying, blue-collar jobs and a low cost of living poured in, and employers were eager to hire and train the new work force.

Both Radley and the NYT are dancing as close as they can to what their camp-followers on Reddit and social media are saying outright, namely: Rural Americans are drug-addled trash who are too lazy to work. This is, as the kids say, wild. You couldn’t make the same claims about, for instance, Black people, or Mexicans, or anything besides the vague “young, working-age people” which means whites in the American media the same way the word “teens”, when used in conjunction with descriptions of mayhem and violence, really means “Black people of all ages” in the American media.

Put another, kinder way, it’s the old “Immigrants do the jobs Americans won’t do” claptrap. Yet people will nod their heads, carefully ignoring:

• If the jobs are too dangerous, underpaid, or exploitive for Americans to do, maybe they shouldn’t exist in that fashion. Isn’t that why we have OSHA, the EPA, a dozen federal agencies?

• If you’re comfortable with immigrants doing dangerous, deadly, or demeaning work, what does that say about your true and secret views concerning immigrants? Are they fully human to you? Do you think that this is the only situation to which they are entitled? Do you envision all of America as a giant Malibu, California, where an army of powerless brown people do back-breaking manual labor while a few thousand privileged whites from good schools shop at Erewhon? Does this excite you?

• Are you stupid enough to think that immigrants won’t eventually do your job — or do you think you have the power to prevent them from doing it?

• How did we get to the point in American history where some laptop-class softie can sit at a bodega drinking an $11 sugar-blasted latte while banging out an NYT article dumping on American citizens for not wanting to do a job for the rest of their lives that the writer, personally, couldn’t do for an hour — and somehow the softie is the good guy?

• Is there any, and I mean any, operating principle left in the Republican or Democrat mainstream parties besides The Spice Must Flow?

Much effort is being expended right now to “debunk” the idea that Haitians in Springfield are eating cats and dogs. This effort, rather unusually for the Left-leaning media, misses the point. The point of JD Vance’s crazy memes and the zillion jokes like this:

is to open the eyes of American citizens to the fact that, pretty much at any moment, your small town or minor city could receive the blessing of 20,000 Haitians. Just because some widget manufacturer would rather hire them than you. If it can happen in Springfield, Ohio, it can happen anywhere, at any time. Don’t think that there will be some practical limit to it, because there are at least, let’s say, three billion people whose lives would be immediately improved if they got a job in the Springfield auto-parts plant. The United States literally has a program in place to fly them here and let them work for years or even decades until the bureaucracy gets around to looking at their application — like the Gaming Commission did with “Ace” Rothstein’s license in Casino.

Listen, I’m the great-grandchild of immigrants. Some of you are immigrants, or H1-Bs, or guest workers. The idea here is not to create a Festung Amerika where nobody can ever visit or become a citizen. Our country has the ability to successfully absorb and integrate a certain number of immigrants every year, every decade, every century. There’s a number that works. It’s not zero. But neither is it…

…seventy million. Or, as Matthew Yglesias imagines, seven hundred million. Yes. That’s on the table for some people. Forget the cats and dogs; if 700 million “newcomers” land in America, it’s the suckers living here currently who will be well and truly cooked.