Tariffs, Tech, and the View from China
Plus: US-Russia cooperation, thirsty data centers, tariff fears, AI fears, US-Iran conflict fears, and more!
Note: If you want to listen to the livestream I did on YouTube with Nikita Petrov and Boris Shoshitaishvili last Friday, it’s now on the NonZero podcast feed. And if you want to join Nikita and Boris for their next livestream—the official launch of their kosmic new show Kosmopolitika—just click this link this Monday, April 14, at 1 pm US Eastern Time. They’ll be fielding questions—and, judging by the array of topics last Friday (politics, AI, noosphere, psychedelics, mind-body problem, etc.), no subject is off limits. Also note: No Earthling next week, owing to Easter-related travels and tariff fatigue. Other content will continue to flow.
—Bob
—The Trump administration is considering using drones to bomb Mexican cartels, possibly without Mexico’s consent, as part of its efforts to fight fentanyl trafficking, NBC News reported. US drones are already conducting surveillance flights over Mexico in order to “build a target deck” for future airstrikes, a former official said.
—A Russian rocket ferried an American astronaut and two Russian cosmonauts to the International Space Station, where the trio will live for eight months as they conduct scientific experiments. Space research is a rare area of continued cooperation between the US and Russia, but that collaboration is expected to end around 2028, when Moscow says it will pull out of the ISS program and start work on its own space station.
—Forty-four percent of Republicans and Republican-leaning independents say tariffs on China will have a positive effect on the US economy, but only 17 percent say tariffs will be good for their own bottom line, according to a new poll from the Pew Research Center. The poll was conducted after Trump placed a 20 percent tariff on Chinese goods but before he increased those levies to 145 percent. (See below for more on Trump’s trade war with China.)
—The long-extinct dire wolf has been “de-extincted,” resurrected via the magic of biotechnology, according to Time Magazine’s cover story—but not according to Scientific American, which interviewed scientists who find such language over the top. While the biotech firm Colossal did edit the DNA of embryonic gray wolves to instill some dire-wolf traits (including white fur), a mere 14 gene edits, these scientists say, isn’t enough to bridge the gap between two species whose last common ancestor lived 5.7 million years ago.
—Federal agents at Detroit Metro Airport detained the lawyer of a University of Michigan student who was arrested last year for her involvement in a pro-Palestine rally. Agents separated the attorney, Amir Makled, from his family upon their return from a vacation abroad and then, during a 90-minute interrogation, demanded access to his cell phone contact list—a demand he successfully resisted.
—Immigration officials said they will start combing the social media posts of applicants for visas and green cards, hoping to weed out people who “support antisemitic terrorism, violent antisemitic ideologies and antisemitic terrorist organizations.” J-Street, a progressive pro-Israel group, condemned the move, saying that the “fight against antisemitism won't be advanced by attacks on 250-year-old cherished American rights like free speech.”
Humanity has just over two years to save itself from annihilation, according to Daniel Kokotajlo. Kokotajlo, who left OpenAI last year to protest the company’s “recklessly racing” toward artificial general intelligence, has gotten a lot of attention lately for a study he co-authored about the future of AI. The study’s projections, presented as a fictional futuristic narrative, depict American AI companies as facing a stark choice by the end of 2027: either slow down development and ensure that their technology aligns with human values, or storm ahead and risk creating a superintelligence that will within a few years kill us all.
Why, you might ask, would these companies choose door number two? In Kokotajlo’s analysis—conducted with a team that included a prominent forecaster—a lot of it comes down to China. If it looks like Chinese competitors could win the race to superintelligence, US companies will have a strong motivation, and government encouragement, to speed ahead.
In a conversation with Kokotajlo, podcaster Dwarkesh Patel spelled out what would seem to be the moral of the story: “It seems like you’re saying it’s a mistake to try and race against China, or race intensely against China.” Wrong! “Not saying that,” Kokotajlo interjected sharply, emphasizing that he doesn’t “want China to get to superintelligence before the US.” American companies will just have to find a way to “thread the needle,” he said—race fast enough to ensure American AI dominance but not so fast as to hasten the apocalypse.
Silicon Valley is said to be full of brilliant people who see farther than the rest of us, but, when the subject turns to China, some of them seem to suffer from tunnel vision. They believe that a headlong push for AI progress could be globally catastrophic (whether via sci-fi-ish AI “takeover” or more mundane forms of conflict and chaos), and they recognize that racing against China for AI supremacy increases the chances of catastrophe. But they can’t seem to envision an alternative to that race—like, say, pursuing a detente with China that might permit the two countries to agree on a responsible approach to AI development.
The view of China that animates Silicon Valley China hawkism has a lot in common with the view of China that prevails within the US foreign policy establishment, which in turn seems to overlap with the view of China in Donald Trump’s brain—the brain that brought us a world in which, at last count, American and Chinese tariffs on each other’s products were 145 percent and 125 percent, respectively. It’s a view of China that fails to take into account China’s view of the world.
The failure here isn’t a failure to grasp nuances of Chinese culture or important epochs in Chinese history (not even the century that many Chinese consider the “century of humiliation” by western powers, though it couldn’t hurt to keep that in mind). The central failure, rather, is a failure to grasp that Chinese people, like so many other people, are people.
Consider one of the several Trump bargaining masterstrokes that have lately made convulsion the stock market’s default condition. Early this week, after China responded to an American tariff increase of 34 percent with its own tariff hike of 34 percent, President Trump wrote on Truth Social, “[I]f China does not withdraw its 34% increase… by tomorrow, April 8th, 2025, the United States will impose ADDITIONAL Tariffs on China of 50%.” Now, there may be a chance—a small chance, but a chance—that a threat like that could have the desired effect. But it definitely can’t have that effect if you make the threat publicly.
To explain why, you don’t have to resort to the explanation deployed by the New York Times, which quoted a China expert saying that “Xi has built up an image of himself as a defiant strongman”. You just have to remember that leaders of nations don’t like their people to see them caving in to threats made by strutting bullies who lead adversarial nations—a fact that is a corollary of the larger truth that people in general don’t like to be seen by other people as caving in to threats made by pretty much anyone.
Sure, context matters—and the Times gets points for trying to explain how the current political and geopolitical context makes the perception of abject submission even less appealing to Xi Jinping than it might otherwise be. Still, various important failures of elites in Washington and Silicon Valley to understand China’s perspective are grounded in simple failures to understand human nature.
Consider a human tendency that’s been amply documented by political scientists: Things nations do for defensive purposes are often construed as offensive by their adversaries (a perception that is part of the sometimes disastrous dynamic known as the “security dilemma”). This is a pretty straightforward emanation of human nature—of our tendency to spend much more anxiety on threats we could face than threats other people could face (combined, perhaps, with our tendency to think of our own intentions as benign). Yet if you suggest to supporters of chip export restrictions that Beijing may see this policy as an attempt to slow China’s economic growth (the very growth that justifies the Chinese Communist Party’s continued rule), the reaction is often one of surprise. After all, the chip restrictions are defensive—they’re done out of fear that a superintelligence-powered China might extinguish democracy globally. Doesn’t Beijing understand that?
Meanwhile, if China responds to this perceived threat by strengthening ties with other autocracies, these same China hawks will likely see that as offensive, a threat to democracy worldwide (even though, if you pay attention, you’ll notice that China seems willing to strengthen ties with any democracies that are so inclined). And so it goes: A positive feedback loop of threat amplification—the kind of loop that famously contributed to the outbreak of World War I.
This kind of logic can play out in all kinds of realms. Just this week, a bipartisan congressional commission warned that the US must “take swift action” to avoid falling behind China in biotechnology. “There will be a ChatGPT moment for biotechnology, and if China gets there first, no matter how fast we run, we will never catch up,” the National Security Commission on Emerging Biotechnology wrote in its final report. The commission recommends, among other things, $15 billion in biotech subsidies and restrictions on biotech exports to China.
The commissioners shared some predictions. If the US beats China, we could use our edge to facilitate the mining of rare earth minerals, develop pest- and drought-resistant crops, and cure cancer. But if our adversaries win, they could “engineer ‘super soldiers’ with genetically enhanced physical capabilities,” create and release microbes that quietly eat away at our bridges and roads, and cure cancer—but, of course, withhold the cure from US patients. The commission’s perception of the situation is clear: America seeks better living through biotech; China seeks “global economic and military supremacy.”
—Major tech firms are operating or building at least 62 data centers in arid regions around the world, where facilities face lower risk of metal corrosion and other humidity-related problems. But some observers—including a former Amazon sustainability expert—worry about the effect of the data centers’ water consumption on drought-prone regions, MaterialSource and The Guardian report. The computing facilities use vast amounts of water for cooling, which can lead to waste. Some use “open loop” systems, which dissipate much of the water via evaporation, and many facilities use water that is “treated with chemicals to prevent corrosion and bacterial growth, rendering it unsuitable for human consumption or agricultural use,” according to researchers at the University of Tulsa.