Author: Kevin Drum

Did you know that Ukraine has destroyed a third of Russia's Black Sea fleet, nearly half of that just since September? I didn't:

The victories have been all the more surprising because Ukraine does not have a traditional navy or a fleet of warships. Instead, Ukraine has used sea drones and missiles to attack Russian ships.

U.S. officials believe Ukraine has sunk 15 Russian ships in the past six months.... As a result of the attacks, Russia has moved its fleet back from Ukraine’s coast and out of the western Black Sea.

....“Ukrainian grain is now again flowing through the Bosporus to Africa and China, which are Ukraine’s traditional markets,” [Radoslaw] Sikorski told reporters in Washington at a breakfast hosted by The Christian Science Monitor.

Apparently the Russians have virtually no defense against the irregular Ukrainian attacks. That's something to think about when you ponder the future of conventional navies.

You're aware of the practice of judge shopping, right? Most often it's used by conservatives, who file suits in specific locations where they're guaranteed to get a wingnut judge who will issue a nationwide injunction based on even the flimsiest evidence. Favored judges include Matthew Kacsmaryk (mifepristone, "Remain in Mexico"); Drew Tipton (immigration); Mark Pittman (student loans, minority businesses); Reed O'Connor (Obamacare, gun control); and Terry Doughty (COVID vaccines, Biden social media ban). All are in the 5th Circuit, where their rulings are likely to be upheld on appeal; all but Tipton are Trump appointees; and all but Doughty are in Texas.

The reason judge shopping works is that there are some jurisdictions with only one, or maybe two judges. File in the right place and you're nearly guaranteed to get the Trump judge of your dreams. Until today:

Judges themselves have moved to crack down on that tactic by adopting a new policy that mandates that all federal suits aimed at invalidating a national policy or statute or a state law or executive order be randomly assigned among judges throughout the judicial district where the case is filed...rather than being retained in the particular geographic division where they are filed and assigned only to the judge or judges in that division.

CNN provides an example of how this works:

For instance, the abortion drug case [mifepristone] was filed in the Amarillo Division of US District Court of the Northern District of Texas, where US District Judge Matthew Kacsmaryk...is assigned to hear all cases under local rules.

Under the new policy, in the future, such a case would go through the random selection process for the entire district court — which has 11 active judges and five senior judges — thus significantly lowering the odds that it would be assigned to Kacsmaryk.

Filing your right-wing lawsuit in Texas will still give you a better chance of winning, but it no longer guarantees that you'll get one of the rubber stamp judges. Progress.

POSTSCRIPT: It's worth a quick note that the ban wasn't the brainchild of wild-eyed liberals. It was spearheaded by Jeffrey Sutton, chief judge of the 6th Circuit, who was appointed by George W. Bush and has been responsible for several very conservative rulings. Even the right-wingers finally got embarrassed by the antics of the Texas judges.

Hum de hum. A couple of reasonable folks have recently weighed in on the question of whether TikTok should be banned in the US. The first is from Noah Smith, who says it should be and points to two pieces of concrete evidence: (1) a study showed that TikTok suppresses videos China doesn't like, and (2) TikTok's official moderation policy is to censor content China doesn't like.

This is pretty weak tea. The suppression study is badly flawed because it doesn't compare TikTok to other social networks in the same years. The moderation story acknowledges that TikTok's policies are not directly aimed at anything specific to China and, in any case, are no longer in use. What's more, TikTok has a longtime goal of minimizing political content entirely because they think it's bad for business.

So there's very little there. Noah also addresses the issue of TikTok surveillance, but again, there's virtually no evidence of anything significant. Nor could there be. TikTok currently keeps all data on American users in Texas, where it's unavailable to Chinese authorities.

Matt Yglesias takes a different approach. He says China is bad and we shouldn't really require tons of evidence about this:

Here’s the analogy I like to use. It’s 1975 and a state-owned Soviet firm wants to buy CBS. What happens? Well, what happens is they wouldn’t be allowed to.... There would be no detailed factual analysis or demand for gold standard evidence that a Soviet-owned television statement might do Moscow’s bidding or that television is capable of influencing public opinion. We’d reject the idea out of hand. And rightly so, because the downsides would be very clear, and the upside minimal.

....We don’t need to wait for open and shut evidence that the platform is being censored or used to deliberately promote propaganda. We should just have a very strong presumption that it will be used in the way and act accordingly.

This gets dicey very fast. If no evidence is required, it's a little hard to argue the other side, isn't it? Just for a bit, then, let's look at some evidence.

First, about 20% of adults watched Walter Cronkite back in the day. Today, about 5% of adults get news from TikTok—most of it nonpolitical. There's really no comparison.¹

Nor is anyone suggesting that we allow China to buy an American news platform. The question is only whether Chinese platforms should be allowed to distribute content in the US. In broad terms this is an issue unique to the internet era, but it's still worth pointing out that the Soviet Union was allowed to distribute news during the Cold War. TASS had several bureaus in the US and Soviet propagandists famously visited the US frequently to provide "their side of the story" to American teens.

Don't get me wrong: there's obviously not much comparison between a huge Chinese social network today and a rarely-used Soviet wire service fifty years ago. At the same time, how much comparison is there between the Soviet Union of the Cold War era and the China of today? China is plainly an authoritarian country that brutally censors content internally and plays hardball against its critics outside the country. But we trade with China; we allow lots of Chinese students in our universities; there's extensive two-way tourism with China; and China is not generally an expansionary military threat.²

As near as I can tell, the threat of China using TikTok to surveil American teens is close to zero. User data on Americans is held on American territory and there's not really much that China could do with it anyway.

On the propaganda side the evidence of a threat is nearly as meager. TikTok probably does try to avoid pissing off the CCP, but its actual level of suppression or censorship seems to be pretty limited and pretty low. What's more, TikTok is just one small part of a massive modern infotainment industry, not one of three monopolistic TV networks like we had in the '60s.

So I'm still stuck in the same place. I really, really try not to get caught up in moral panics, which have a pretty appalling track record. TikTok strikes me as an almost dictionary example of a moral panic that we'll look back on in ten years and wonder what the hell we were thinking. On the threat of China more generally, I don't like the Chinese government much but I'm not panicked either. They're a geopolitical bad actor but not an existential threat. For now, anyway, I don't feel the need to ban a social network just because it's Chinese owned.

¹The number is obviously higher for teens, probably around 20% or so. But again, it's mostly nonpolitical news consumption.

²It's a military threat, but mostly in places on its border that it's claimed for a long time: Taiwan, Tibet, the South China Sea, and a couple of smallish territories that it disputes with India.

It's time for some zoo pictures. The LA Zoo, for some reason, requires you to walk several hundred yards up a "promenade" after you get your ticket but before you get to the zoo itself. This is boring because there's very little along the way except for a few little pop-up shops.

But. There is one thing: flamingos. So today you get the entrance to the zoo and a flamingo captured along the way up the promenade. Next week we'll start in on the zoo proper.

The Justice Department has released a transcript of Joe Biden's interview with special counsel Robert Hur, and the Wall Street Journal has read it:

Other news outlets basically agree, mostly opting to present their findings as "nuanced" or providing "context." But they fundamentally say the same thing: Biden was being Biden. He joked, he veered off into stories, and he got two or three things wrong during a five-hour interview. In no way did he seem like a forgetful grandpa, as Hur gratuitously implied in his report.

This whole thing has been a travesty. Someday maybe we'll get a special counsel—just one!—who isn't a partisan Republican with an axe to grind. Someday.

Yikes! The BLS reported CPI inflation of 5.4% in February. Core CPI was lower at 4.4%:

This is a big increase, driven by inflation in services and an increase in gasoline prices over the past month. Other commodities were either flat or down, including food, which showed no price increase from last month.

On a more conventional year-over-year basis, headline CPI inflation was 3.2% and core inflation was 3.8%.

The Stormy Daniels hush money trial is scheduled to begin in two weeks. Naturally this means that Donald Trump's legal team is just now getting around to asking for a delay while we wait for the Supreme Court to rule on whether Trump is automatically immune to prosecution for anything he did while president.

The timing of this is obviously suspect, but you may also be wondering what official presidential acts Trump is trying to protect. After all, this case is about alleged hush money paid to keep Stormy Daniels quiet during the 2016 campaign, months before Trump was inaugurated.

One possibility is that although Michael Cohen made the payments in 2016 on Trump's behalf, he was reimbursed in 2017, while Trump was in office. But it turns out that's not part of Trump's argument. Instead, he's arguing that part of the prosecution's case includes evidence of statements Trump made in 2018. That's it. The actual acts predated his presidency, but some of his incriminating public statements were made on Twitter two years later. And nothing he said while in office can be used in the case.

This seems mighty thin to me, but IANAL. Stay tuned.

The whole Kate Middleton Photoshopping scandal is bizarrely inexplicable. Journalists have now identified nearly two dozen edits to the Mother's Day picture, and the inexplicable part is that they're both clumsy and pointless. Kate says she did the edits herself, which is kind of believable. Certainly no expert did this stuff.

And why? The edits are all tiny and accomplish nothing. The Guardian's imaging manager thinks that several images were composited together, which would explain the large number of anomalies but not the reason for doing it. Photoshop itself wouldn't have been so clumsy constructing a composite, which means Kate manually cut and pasted pieces from different frames into the primary shot.

But that doesn't make sense either. I don't really care much if Kate decided to spruce up an image a little bit, but on a purely technical level I'm burning with curiosity about what really happened.

This is from an article Derek Thompson wrote for the Atlantic last month:

This is a remarkable finding, but I was unable to replicate it until Thompson was kind enough to show me exactly where it came from. As I thought, it was from the American Time Use Survey, and compares "Animal and pet care" with "Socializing and communicating." But there's a problem here.

ATUS always reports two results. The first is the average time spent among all people, even those who don't participate in a particular activity. The second is the average time spent among those who participate. Thompson used the first for socializing and the second for pets.

That's apples to oranges. Here are the numbers for participants for both metrics:

There's another issue with the ATUS data: it reports that only a fifth of women spend any time at all on pet care. However, since anyone who owns a pet has to spend at least a few minutes caring for their pets, if only to feed them, this suggests that only a fifth of women own pets. That defies every other survey in existence, which report that 50-60% of women own pets.¹

So I'm not sure what to make of the ATUS data. However, my guess is that "participants" ends up being a proxy for the most dedicated pet owners, who are naturally likely to spend a fair amount of time on their critters.

There's not much question that socializing has diminished and pet care has gone up. The data all points that way. But I think the situation is not quite as dire as Thompson suggests.

¹It's less for men, but only by a little.

Sometime in the next few months a star called T Corona Borealis is going to explode. How do we know this? Because T CrB isn't going to be a supernova—which would be unpredictable and destroy the star for good. In fact, T CrB isn't even a single star. It's a binary star: a white dwarf orbiting around a red giant. The white dwarf sucks in material from the red giant, and every 80 years it gets dense enough to explode in a nova. But only the outer layer of the dwarf explodes. The rest is left intact to begin the cycle anew.

I was in the desert last night doing a bit of astrophotography, and while I was there I decided to shoot a picture of T CrB. This is the before picture. If the weather cooperates, I'll shoot the same picture when it explodes, and you'll see the same white dot except bigger. Exciting!