No less authorities than Eric Schmidt, former CEO of Google, Alexandr Wang, founder and CEO of Scale AI, an AI infrastructure titan, and Dan Hendrycks, director of the Center for AI Safety, have produced a document on "Superintelligence Strategy."

This white paper takes for granted the emergence of future superintelligence, and looks at the national security implication.

They propose a framework, directly inspired by the Cold War era Mutually-Assured Destruction, called "Mutual Assured AI Malfunction" (MAIM). MAIM works on the principle that any state attempting to develop superintelligence that might threaten other powers' security can expect its AI projects to be sabotaged or disabled by rivals. The authors argue that MAIM is already the default strategic situation because of the relative ease with which AI projects can be compromised through sabotage or cyberattacks.

The authors argue that MAIM creates a strategic stalemate that prevents any single nation from achieving an AI monopoly that could threaten global security. They propose formalizing understandings about MAIM to prevent misinterpretations of sabotage actions that could lead to broader conflict. This includes establishing clear escalation ladders, improving cyber capabilities to enable less destructive forms of sabotage, building data centers in remote locations, and potentially developing AI-assisted inspection mechanisms to verify compliance with agreed limitations.

While "MAIM" is the centerpiece of the paper, they also have a series of proposals. Concerning economic competitiveness, the authors emphasize the vulnerability created by dependence on Taiwan's semiconductor industry. They argue that nations must invest in domestic AI chip manufacturing facilities, despite higher costs, to ensure supply chain resilience in the event of Chinese aggression toward Taiwan. This mirrors past strategic investments in critical technologies like uranium enrichment during the Manhattan Project. The authors contend that as AI becomes more integrated into economies, access to advanced AI chips will increasingly define national power.

When it comes to the military, the authors recommend securing drone supply chains, as many nations remain dependent on Chinese manufacturers for key components. They also advocate for carefully integrating AI into military command and control systems and cyber operations while maintaining human oversight of critical decisions to prevent "flash wars" triggered by AI mistakes.

The whole thing is well worth your attention.

Policy News You Need To Know

#HigherEd #Antisemitism — The Trump Administration is cutting $400 million in federal funding from Columbia, specifically over its abetting of antisemitism, The Free Press reports, and it's "only the beginning," according to their source. This is very welcome news, but we'll add this from Claremont's Jeremy Carl: "if they expand it to cutting off funding for universities pushing anti-white racism, it will be a real game-changer."

#Budget — The CRFB has a blog item on estimates on the "dynamic impact" of the House's budget plan. Dynamic scoring, as I'm sure you know, tries to take into account, when ascertaining the future consequences of a budget, its impact on economic growth, which, if positive, will in turn shrink the deficit. It's sound macroeconomics, but by playing with the math, you can come up with some pretty crazy stuff. And it looks like that's what the House is doing.

#Reg — The CFPB, it looks like, is on its way out, thanks to clever squeezing of its budget by OMB. However, it's not dead yet, and Pinpoint Policy Institute has an item attacking its new overdraft rule.

#FreeSpeech — Heritage has correctly noticed that the EU is using its regulatory powers, combined with the size of its domestic market, to try to impose its rules globally, and not just on its own citizens. The EU calls itself a "regulatory superpower." The idea is this: regulatory compliance is expensive; the EU is such a huge market that companies can't afford not to comply with EU regulations; in order to save money and have greater convenience, global companies will end up abiding by EU rules everywhere, since they have to abide by them in a huge market anyway. This includes in the area of online free speech. Heritage's Daniel Cochrane and Mary Mobley are therefore right to say that "Europe wants to be the world's speech police," and that America should resist this.

#Tax — Interesting item from the National Taxpayers Union. Did you know that the rules on how businesses can deduct interest from their corporate income tax bill changed in 2022? Since 2022, businesses can no longer deduct significant costs related to depreciation and amortization. A new bill would restore the previous status quo, which would probably help investment.

#Tax — Another interesting tax item, this time from the excellent folks at the Tax Foundation: they point out that highway funding has exceeded highway revenues in recent years, and note the shift to electric vehicles, which, they argue, has "made the gas tax increasingly obsolete."

#DaylightSavings — The issue of daylight savings has recently come up in one of President Trump's impromptu Oval Office press gaggles. Here's Cato's Jeffrey A. Singer with the case against.

#DepartmentOfUnintendedConsequences — Reason's C. Jarrett Dieterle has an intriguing theory: the spike in egg prices could be related not just to the avian flu-motivated culling of birds, but also state-level "cage-free laws." These laws "ban the sale of any eggs within state boundaries that are not laid by cage-free hens." They certainly, all else equal, must make eggs at least marginally more expensive, but could they really be contributing significantly to the recent price spike? You be the judge. One part of the piece stuck out at your correspondent, who certainly has no a priori objection to regulations meant to ensure animals are raised humanely: "Cage-free laws could be viewed as justifiable if they created discernibly more humane habitats for the chickens. But as anyone who has viewed pictures of a 'cage-free farm' can attest, they often consist of hordes of chickens jammed inside a large space, versus a caged farm, which consists of hordes of chickens each jammed in small individual spaces. Is one actually any better than the other?"

Friday Essays

The inestimable Santi Ruiz, author of the excellent Statecraft newsletter, has "50 Thoughts on DOGE" and they're well worth your time. He has taken a step back from the firehose of news and, from a perspective broadly sympathetic to DOGE's aims but also mindful of the need for state capacity, drawn this compendium which strikes us as the best thing written on DOGE since January 20.

Ruiz's essay also linked to this older essay, by Zvi Mowshowitz, explaining with great acumen the techniques which mainstream journalists use to lie while technically remaining faithful to their particular set of so-called ethics, and giving some advice on how to navigate through their misinformation. It's one of those great pieces of writing that explains something you already know, but does it with such clarity and perspicacity, putting into words things that existed hitherto in your mind only as evanescent thoughts, that you still feel like you have learned a great deal. Sample quote: a journalist "can call anyone an expert. Expert consensus means three people. ‘Some investors’ and similar phrases mean two (as does ‘surrounded by.’)"

Arena magazine, a new and very interesting magazine out of Silicon Valley dedicated to celebrating America's "builders", has a striking essay by Dean W Ball making the optimist and patriotic case for AI, titled "Life, Liberty, And Superintelligence." Just that. Ball asks: "What would it take for America—or any country—to realize the benefits of AI [within the next decade]? How would America and its allies regain unquestioned global supremacy, and would the process of doing so itself provoke a war? Perhaps most importantly, how will average humans—the ones who do not know about the latest advancements in AI, the ones who merely want “life, liberty, and property,”—contend with the arrival of a new ‘superintelligent’ entity?"

Speaking of media outlets out of Silicon Valley, Pirate Wires is not a highbrow magazine, unlike Arena, but they do produce very good investigative journalism on occasion. This piece is based on "exclusive interviews with with 20 'trust and safety' workers at Meta, TikTok, Google, and other major tech companies" to chart the history of the online speech censorship and finds that "from the beginning, it was inconsistent, ineffective, and focused more on posturing and internal politics than actually searching for truth."

When you mention the West's woes with birth rates, you sometimes hear the answer that the solution to the problem is more immigration. Aporia Magazine runs the numbers, quite thoroughly, to demonstrate that "Immigration Does Not Solve Population Decline."

At Chronicles, Mark Bauerlein addresses the idea of "viewpoint diversity," the Supreme Court jurisprudence-derived concept that is seen as a kind of helpful compromise in the culture wars. This is mistaken, Bauerlein argues.