The Never Trumpers Discover Their Ass With Both Hands (And Still Get It Wrong)

Ken Paxton won his Senate primary last night by nearly doubling up John Cornyn, a 24-year incumbent senator who is, by any honest measure, just as much a Trump rubber stamp as Paxton. The only difference between them is that Cornyn kept his nose clean and Paxton is a walking rap sheet with a side of adultery. Texas Republicans looked at that choice and said: more of the criminal one, please. Culberson County went 79-21 for the guy whose own wife "biblical grounds."

The hopium crowd is, predictably, losing their shit about this. The argument goes that a crook this cartoonish is finally the one Democrats can beat. Great. Sure. The same people said that about Beto in 2018 and he ran a near-perfect campaign against a genuinely weak Ted Cruz during a Democratic wave year and still lost by 215,000 votes. Texas in 2026 is moreRepublican than it was then, not less. The purple Texas narrative has been wish-casting for going on a decade now. At some point you have to look at the fucking data.

But here's where it gets really fun. Jonathan V. Last over at The Bulwark, who is sharp enough that I read him even when he's wrong, has published his Triad today arguing that Democrats need to go black-hat to peel off some slice of Paxton's voters. The specific suggestions include vice signaling, running against the Democratic Party itself from the right, flirting with conspiracy theories, and attacking Israel from an America First position. JVL himself admits he doesn't like any of this. He thinks it's bad and potentially dangerous. He then publishes it anyway because, I suppose, the newsletter must go on.

This is the Never Trump intellectual tradition in full flower: identify a real problem with genuine analytical clarity, then propose a solution that would make everything measurably worse while congratulating yourself for your unsentimental toughness.

Let's walk through exactly why this is stupid.

The sitting-out mechanism that JVL is vaguely gesturing at, the idea that some slice of Cornyn voters could be convinced to stay home if Democrats are sufficiently transgressive, is theoretically plausible yet practically useless in Texas. Republican turnout infrastructure in that state is not built on enthusiasm. It is built on sorted identity. Those golden-retriever-with-an-R voters are not contingent. They do not need to be inspired. They show up because that is what they do, the same way they go to church and fly their flags and do not, under any circumstances, vote for Democrats who are flirting with goddamn chemtrail theories.

The Mamdani comparison that's floating around as proof of concept is even worse. Zohran won a Democratic primary in New York City, a jurisdiction where the median voter is genuinely to the left of the national party on rent, healthcare, and yes, Palestine. His "transgressive" positioning was transgressive relative to the Democratic establishment, not relative to his actual voters. That's the opposite of what JVL is proposing. He's asking Texas Democrats to be transgressive relative to their own voters in hopes of attracting voters whose entire identity is built around cultural contempt for them. These are not analogous situations. They are not even in the same galaxy.

And then there's the Fetterman problem, which is maybe the most important thing nobody in the Never Trump ecosystem wants to talk about. Even if you buy the theory, even if some black-hat Democrat somehow threads the needle and wins, what the hell do you think you've got? You've got a politician whose entire brand is built on not being a reliable Democrat. That brand has to be continuously fed or it dies. You don't get a 50th vote. You get Joe Manchin. You get Kyrsten Sinema, who optimized so completely for maverick positioning that she ended up with no constituency at all, too weird for Democrats, not Republican enough for Republicans, functionally irrelevant to everyone.

The pattern here is not subtle:

The black-hat strategy selects for candidates whose entire political survival depends on being a pain in the ass to the coalition they're theoretically part of. You are recruiting saboteurs and calling it strategy.

James Talarico is probably going to lose to Ken Paxton in November[1]. That outcome is going to be correct. Texas is not ready to be won at the statewide level, the structural hole is too deep, and the answer is not to set your own candidate on fire in hopes that some Paxton voters find the smell interesting. The answer is that Talarico runs hard, builds infrastructure, tests messaging, and Democrats live to fight another cycle with their coalition intact rather than demoralized by watching their candidate audition for a MAGA-lite role nobody asked for.

The Never Trump tribe remains, bless their hearts, convinced that there is a clever strategic path through this that does not require confronting how thoroughly fucked the underlying math is. JVL knows better than most of them. Which is exactly what makes this particular piece such a spectacular faceplant.

1 - sure, there's a chance, but I am not buying it at the moment. Perhaps in September the polls will tell a different story, but I ain't gonna hold my breath...