🚨 Announcement: We’re migrating to Substack
Well, the day has finally come for this lovely newsletter to migrate to Substack. For the curious, the primary reason is the engineering maintenance required to keep our publishing system functional against constant underlying platform changes. Substack (we hope!) offers better long-term stability so that I can mostly write words instead of code.
We will be splitting up this newsletter into separate editions. That means you, dear reader, can subscribe to what you want. We will have:
- Lux Recommends — the same weekly roundup of sharp, strategic reads on the world to come we’ve run for almost ten years now. This will include links to our other content.
- Dispatches — my writing from the frontiers of science, technology, finance and the human condition
- Interviews — our edited Q&As with leading experts
- Event Announcements — upcoming community meetups and Riskgaming runthroughs near you
Your subscription will be automatically moved to Substack next week, and you will be able to manage your preferences by going to https://www.riskgaming.com/ once complete.
For those keeping track: this newsletter started on TinyLetter, migrated to Twitter’s Revue, then MailChimp and now Substack, all in about four years. Welcome to the chaos of digital media.
In other words: thank you for sticking with us. If you have any questions, I send this newsletter from my email, so just hit reply.
How Trump killed the American manufacturing renaissance before it started
Laurence and I headed down to Austin this week to host Riskgaming sessions at the brand-new Endless Frontiers conference, which convened a group of luminary foreign policy and defense officials alongside tech founders and venture capitalists for two days of discussion about the future of America.
While broad in scope, the discussions tended to center on a few recurring themes, including: how America can reindustrialize in order to have sovereignty over critical supply chains; how America can set itself up well to compete with China; and how America can foster a service ethic that encourages everyone to seek agency in building up this country for another generation.
These are popular topics on this circuit, and the subjects and many of the answers were consistent with the Reindustrialize conference in Detroit I attended last year. I wrote up the central dilemma that participants faced there:
That’s where I had to break with the narrative of the conference. For deindustrialization wasn’t a cultural shift, but rather a premeditated strategy organized through economic incentives that pushed American workers to higher-margin industries and away from low-margin work. It was a naive attempt to boost America’s anemic productivity growth in the 1980s and 1990s by taking advantage of the imminent IT revolution to raise living standards.
Here’s the thing: it worked. America’s per capital income exceeded $76,000 in 2022, and has doubled in real terms since the 1980s. There are well-chronicled challenges with inequality, as well as an open question whether a doubling over forty years is fast enough growth. But Americans are making more money than ever before — and there is little evidence that reindustrialization will accelerate that income growth or offer a more productive life to more people.
Endless Frontiers had the same can-do spirit one expects from its Texas milieu, with ambitious founders taking on intricate business challenges. One founder I talked to has created the first rare-earth mine that is fully decoupled from China, and I was quite compelled by his vision. The optimism was infectious.
Yet, the difference between June 2024 and today is — obviously and painfully — the occupant in the White House. By the tail end of the Biden administration, we were talking about new factory investments, the impact of the CHIPS Act and the Inflation Reduction Act on greenfield projects, and whether there was a budding renaissance in American manufacturing. With the switch in administrations, we now talk about whether tariffs will be up or down in the next hour.
Even reading my own phrase, “premeditated strategy,” gets at the complete sea change in just the past few months. Was Biden’s strategy working? There were positive signs, although I remained somewhat skeptical of an impending reshoring of supply chains. But there was a strategy, there were mostly competent people running the right program offices (even as I complained bitterly about some of the distractions laden in these programs), and perhaps most importantly, there was global macro stability, which is conducive to long-term financing of physical plants.
Last year, “I had to break with the narrative of the conference” because nearly every attendee was ebullient about the prospects for American manufacturing. Not unlike a crypto conference, participants had visions of grandeur about reinstalling America as the preeminent force in physical goods.
Now, the narrative had walked over to my side. The absolute chaos of the past few weeks had clearly stunned attendees. Secretaries of defense, CIA directors, senior flag officers, major corporate CEOs — all of them were flummoxed. Effusive optimism has been replaced with weary pessimism. Endless Frontiers convened a very smart, extraordinarily experienced and pragmatic group of people. They know what Trump just did to any seed of reindustrialization, and that is to kill it.
For decades, Trump has advocated for eliminating America’s trade deficits and has excoriated East Asian nations like China, Japan and South Korea for their currency controls that buttress dollar supremacy. What’s missing from Trump’s simple analysis is an understanding of why these countries have managed their currencies so tightly.
Price stability — and global macro stability more broadly — is the linchpin for industrial capacity. New factories take significant time and capital to build, and so projections of their future profits are pivotal to their approval. When everything is stable, this analysis is simple and decisions are made rapidly. Instability, however, raises the threshold for committing to an investment due to the stochastic risk factors that must be included. If interest rates, inflation or currency exchange rates are unpredictable, it’s next to impossible to make a decision on whether a factory is viable or not. Better to wait.
China and other East Asian industrial centers solve for stability by actively managing their currencies and macroeconomies through an extensive toolbox. Last summer, Brad W. Setser wrote on how China has continued to peg its currency to the U.S. dollar, which is obvious from macroeconomic analysis. The South Korean won and Japanese Yen show much larger gyrations, but their governments similarly protect their markets for industry by maintaining as stable an external environment as possible.
Turn to the United States, and I’m crestfallen to say that “stability” is the adjective furthest from the reality of our markets. The most critical ingredient for bringing back American manufacturing is more distant than it has ever been.
Which obviously leads to the central challenge for America: It needs a long-term, consistent industrial strategy. This strategy can take a lot of different forms, and can prioritize a range of different goals. The key though is that it’s a strategy (where resources are used to match goals in a logical way) and that it stays consistent for long enough to make a difference. Critically, investors would have to perceive that the strategy was here to stay, either enshrined in law or with such widespread bipartisan support that it would be expected to carry the day regardless of which party was in power. Michael Kratsios, who heads Trump’s Office of Science and Technology Policy, gave his first public address at the conference, where he said that “a Golden Age is only possible if we choose it."
That’s the rub though. It’s precisely the lack of consensus around choosing reindustrialization that undermines it before it even gets started. In conversation after conversation, the questions around American power returned to the more profound question of what the country even wants, and how we can come to a broad societal agreement on it.
Even with my weary pessimism, let me be frank: America’s advantages are superlative. Our access to natural resources, a two-ocean border, the smartest minds in the world, a free-market capitalism that rapidly moves money to the most valuable opportunities, the greatest research labs in the world — we are just fucking loaded with advantages.
We don’t agree on how to use any of our assets though. I just read Paul Fussell’s Class (thanks to Peter Hébert’s repeated recommendations), a magnificently hilarious sociology of America. So my analogy here may just be scrambled by Fussell’s caustic and wry depictions of behavior. America isn’t middle rich, where everyone at the funeral understands that arguing over the inheritance through lawyers is a loser’s game straight to bankruptcy. We’re filthy rich, and we understand that aggressive and avaricious attorneys ready to fight a blood sport to the death might result in a lifetime of leisure. (Insert Succession comment here from someone who watched the show.)
That’s a cultural problem impossible to solve within the confines of a gorgeous and gregarious outdoor conference.There is no use talking about reshoring supply chains when interest rates may skyrocket in the next few weeks. No one can make a thoughtful investment when economic policy can change on a whim.
Case in point: Last weekend, Trump reportedly authorized Nvidia to export its lower-spec H20 chips to China. That decision was reversed by Tuesday. Nvidia then announced an impairment of $5.5 billion in its first-quarter sales to China from the regulatory change, resulting in its market cap dropping more than $250 billion.
That’s not a strategy — that’s just chaos. At the conference this week, a former secretary of defense codenamed “Chaos” was walking about, and a founder and former CIA director from a company called Chaos was chatting with participants as well. Maybe chaos is the answer to America’s manufacturing decline — I just wish I knew how.
How can we make the internet fun again?
This week, Laurence Pevsner and I interviewed Renée DiResta, the prominent internet research specialist now based at Georgetown. She had a great piece in Noema earlier this year on “The Great Social Media Diaspora,” which explored the changing contours of social media with the rise of the so-called fediverse.
🔊 Listen to “How can we make the internet fun again?”
This interview has been condensed and edited for clarity.
Danny Crichton: We first met you in person at our Riskgaming session in Washington D.C. We were hosting this game, DeepFaked and DeepSixed, which focuses on AI election security. This was right before the election in November. And one of the threats that comes out of the game is a voice hack, in which the Russians or North Koreans download clips of local pastors from YouTube, make a voice print of them, and then deliver phone calls saying “Hey, this is Reverend so-and-so from the local church. I just want to make sure you're voting on Tuesday. It's really important for the issues that we do.”
In the end, we survived 2024 from a deepfakes perspective. But do you think that changes either in '26 with the midterms, or '28 and going forward?
Renee DiResta: My friend Katie Harbath has this phrase, “panic responsibly.” You don't want to say, "Oh, the sky is falling and everything's going to be terrible." For the 2024 election, it was important to point out the risks. And recognizing what could happen, how do we think about mitigating it if it does, or making the public aware? That's responsible risk management.
And the technology is only going to continue to become more democratized and easier to use. So there are going to be fewer guardrails on what's possible. You can't really put that genie back in the bottle. There's no way to regulate model outputs because of their myriad uses.
So then the question is how we let people who want to say, “I am real,” say it. There's no compulsion here, but maybe it’s social ecosystems or companies, for example, requiring some sort of credential in order to prevent voice or face spoofing.
But the main point is that it's important to be aware of the trends and the risks, and then to decide rationally and communally what measures we consider appropriate and reasonable for mitigation.
Read more on our Riskgaming Substack next week.
The OB: Jason Crawford on The Techno-Humanist Manifesto
This week on The Orthogonal Bet, Lux’s scientist-in-residence Sam Arbesman interviewed Jason Crawford of Roots of Progress, who is serializing his “The Techno-Humanist Manifesto” on Substack. I interviewed him back in August when the first chapter was published, and now Sam gets the latest on its reception as the project near its midway point.
Lux Recommends
- Our editor Katie Salam recommends a research paper from the Max Planck Institute on “Artificial neurons organize themselves.” “The new artificial neurons, known as infomorphic neurons, are capable of learning independently and self-organized among their neighboring neurons. This means that the smallest unit in the network has to be controlled no longer from the outside, but decides itself which input is relevant and which is not.”
- Sam enjoyed Benjamin Breen’s article, “When Jorge Luis Borges met one of the founders of AI” (also catch Breen on our Riskgaming podcast episode from last year). “If you compiled an enormous dataset of everything Borges read, and combined it with an exquisitely sensitive record of every sensory experience he ever had, could you create a Borges LLM? It’s a question that I think a lot of my peers in the humanities would not take much interest in, today (the two cultures, again). But I love that Borges himself was thinking about it.”
- I should have sent this recommendation out last week, but ‘Hilarius Bookbinder’ wrote a plangent note in “The average college student today” that seems to have really struck a chord with the internet crowd, including me. "That’s right, ChatGPT. The students cheat. I’ve written about cheating in ‘Why AI is Destroying Academic Integrity,’ so I won’t repeat it here, but the cheating tsunami has definitely changed what assignments I give. I can’t assign papers any more because I’ll just get AI back, and there’s nothing I can do to make it stop. Sadly, not writing exacerbates their illiteracy; writing is a muscle and dedicated writing is a workout for the mind as well as the pen.”
- While we wait decades for a new Penn Station, Sam points to this piece on “How Japan Built a 3D-Printed Train Station in 6 Hours.” “The new station’s components were 3D-printed elsewhere and assembled on site last month, in what the railway’s operators say is a world first. It may look more like a shelter than a station, but building one the traditional way would have taken more than two months and cost twice as much, according to the West Japan Railway Company.”
- Shaq Vayda liked Yasemin Saplakoglu’s look in Quanta on “Intelligence Evolved at Least Twice in Vertebrate Animals.” "A series of studies published in Science (opens a new tab) in February 2025 provides the best evidence yet that birds and mammals did not inherit the neural pathways that generate intelligence from a common ancestor, but rather evolved them independently. This suggests that vertebrate intelligence arose not once, but multiple times. Still, their neural complexity didn’t evolve in wildly different directions: Avian and mammalian brains display surprisingly similar circuits, the studies found.”
- Finally for this weekend via Sam, the value of human ingenuity and skill in the protection of our most important resource — cheese — in “Why Tap a Wheel of Cheese?” “‘We hit all of the surfaces of the cheese, and from the sound of the hammer against the surface we can imagine in our mind an Xray of the wheel and how it is internally. For a perfect wheel with no defects, the paste is completely compact, no empty space. If there are structural defects, such as little fissures or cavities, we can hear that the sound of the hammer is different,’ explained Stocchi.” Now I’m hungry.
That’s it, folks. Have questions, comments, or ideas? This newsletter is sent from my email, so you can just click reply.
