The twilight of tech unilateralism

Suddenly, America is in a whole different world

The new issue of Foreign Affairs is out, and the cover article, The Weaponized World Economy, is by Abe Newman and myself. We argue (a) that America needs to regear for a world in which it isn’t the only power with technological and economic chokepoints, and (b) that even if the Trump administration recognizes the challenge in some loose sense, it is hopelessly ill-suited to bring through this transformation.

In a little-noticed speech in June, Secretary of State Marco Rubio hinted at the administration’s reasoning. China had “cornered the market” for rare earths, putting the United States and the world in a “crunch,” he said. The administration had come to realize “that our industrial capability is deeply dependent on a number of potential adversary nation-states, including China, who can hold it over our head,” shifting the “nature of geopolitics,” in “one of the great challenges of the new century.” Although Rubio emphasized self-reliance as a solution, the administration’s rush to make a deal demonstrates the limits of going it alone. The United States is ratcheting back its own threats to persuade adversaries not to cripple vital parts of the U.S. economy. Other powers, too, are struggling to figure out how to advance their interests in a world in which economic power and national security are merging, and economic and technological integration have turned from a promise to a threat.

If you want to read it - and you are not currently a subscriber to Foreign Affairs - I recommend that you click through soon. It’s outside the paywall now, but likely to scuttle back inside its shell in a couple of days.

But what I want to write about (I promise, this newsletter will be shorter than the last one), are the commonalities between this piece, Graham Webster’s WIRED article on the chip choke strategy from a few days ago, and Eric Schmidt and Selina Xu’s op-ed in the New York Times today, arguing that we should stop obsessing over AGI. The three are actually of a piece. A year ago, the US appeared to have a coherent world view that tied together a particular perspective on AI with a particular view of American power. Now, both the AI story and the power story are crumbling. What next?

Back in January, I had a conversation with Ed Luce, which briefly laid out the narrative that pulled AI, semiconductors and the weaponized world economy together. The relevant bits from my side of the conversation:

there’s an excellent case to be made that Eric Schmidt is the most influential American foreign policy thinker of the early 21st century. … Over a few years, Schmidt reshaped America’s understanding of national security through talks like the one you heard, and his leadership of the National Security Commission on Artificial Intelligence.

People used to talk a lot about the so-called “Washington Consensus” — the rules of economic neoliberalism that guided the thinking of the World Bank, IMF and other DC based institutions. Now that economic neoliberalism is moribund, I think that there is a New Washington Consensus, which Schmidt has done more than anyone else to shape.

Instead of multilateral institutions, you should now look to the assumptions of the emerging mind-meld between Silicon Valley and national security policymakers, to see how America wants to shape the world. These assumptions can be boiled down to four claims: that competition between the US and China is everything, that AGI is right around the corner, that whoever gets to AGI first is likely to win, and that America’s big advantage is its chokehold over the chips that you need to train powerful AI.

Those who believe these claims, including Biden’s national security adviser Jake Sullivan, and Trump’s former deputy national security adviser Matt Pottinger and Anthropic founder Dario Amodei, argue that America should focus on slowing down China’s build-up of AI, by denying it access to specific powerful chips. If America can only get to AGI first, it can build up an overwhelming long-term advantage. …

I’m sceptical myself, both because I don’t believe we’re on the verge of AGI, and because I suspect that it is much harder to control future technologies than national security thinkers such as Sullivan contend. But there is another problem. Dan Wang suggests in his fantastic forthcoming book, Breakneck: China’s Quest to Engineer the Future, that China is focusing on building the physical technologies of the future, such as renewable energy, while America obsesses over virtual possibilities and problems. If AGI turns out to be a bust, and the Trump administration does everything it can to squash renewables, America could find itself in real trouble.

More on Dan’s excellent book soon, but the crucial points are that (a) this mindmeld was very, very important to American policy, and (b) it is visibly crumbling away. When I described this New Washington Consensus’s focus on the AGI proposition, I was relying on informal conversations with people who I knew in the intersection of DC foreign/tech policy. Now, Graham Webster has done proper interviews and come back with the goods.

the Biden administration was placing a high-stakes bet that the United States could use its leverage to hold China back and that the losses in terms of forgone US exports to China and collateral damage to bilateral ties would be worth it. … several people who served in the Biden administration say that a more novel concern was also behind the big bet. Key officials believed AI was approaching an inflection point—or several—that could give a nation major military and economic advantages. Some believed a self-improving system or so-called artificial general intelligence could be just over the technical horizon. The risk that China could reach these thresholds first was too great to ignore.

The only thing that I would add to his account is my surmise that the “self-improving system” hypothesis was the most crucial underpinning of the bet. If the U.S. could get to AGI first - and AGI was capable both of making itself better and advancing technological progress on a variety of fronts, then the losses in terms of leverage over China would be trivial. The U.S. would have created a technological cornucopia, a self-licking ice cream cone of technological advantage that would secure its long term dominance.

As Webster’s account suggests, there is a bigger institutional story to tell about how America came to make this “high stakes bet.” I’ve already mentioned Schmidt’s importance; Webster emphasizes the role of CSET, a Dustin Moskowitz-funded institution within Georgetown that brought together rationalist/EA arguments on the crucial importance of AGI with analyses of semiconductor supply chains and China’s vulnerabilities. There is a great dissertation to be written about the history of how this all came together.

But if the AGI bet is a bad one, then much of the rationale for this Consensus falls apart. And that is the conclusion that Eric Schmidt seems to be coming to.

Just a few months ago, he was boosting AGI in the Wall Street Journal.

the contours of an AGI future are beginning to take shape. AI systems capable of performing at the intellectual level of the world’s top scientists are arriving soon—likely by the end of the decade. … The magic would really kick off—and it does sound like magic—if the systems reach a point at which they become scale-free, meaning that they could train themselves on self-generated data through a process known as recursive self-learning, relying only on electricity to advance … The advent of AGI could herald a new renaissance in human knowledge and capability. From accelerating drug discovery to running whole companies, from personalizing education to creating new materials for space exploration, AGI could help solve some of humanity’s most pressing challenges.

Now, he and Xu are making a very different argument:

It is uncertain how soon artificial general intelligence can be achieved. We worry that Silicon Valley has grown so enamored with accomplishing this goal that it’s alienating the general public and, worse, bypassing crucial opportunities to use the technology that already exist. In being solely fixated on this objective, our nation risks falling behind China, which is far less concerned with creating A.I. powerful enough to surpass humans and much more focused on using the technology we have now. … while A.I. capabilities have made extraordinary leaps since the debut of ChatGPT in 2022, science has yet to find a clear path to building intelligence that surpasses humans.

To be clear, I’m less interested in scoring points than in talking through the practical consequences. I greatly value many of the people who I disagree with on these questions, and it is obviously better that people ruthlessly abandon problematic bets than that they double down on them. But a few short months ago, the United States believed that it could use its chokehold on advanced semiconductors to create a grand global scheme for controlling the development of AI, and secure its long term dominance. Now, that is in tatters, and not just because of Trump. The technological bets that underlay the grand strategy of unilateral domination look to be going bad. Even if the same people were in charge now as in January, they would have to deal with some extremely difficult questions.

Of course, they are not in charge, and expertise driven technocratic over-reach may, it turns out, be a less vexing problem to solve than an administration that seems to treat expertise as a direct threat. The assumption that America held all the good cards was a bad one. The current proposition - that America ought throw away the cards that it has - is much worse. As Abe and I argue in Foreign Affairs:

Every administration is forced to build the plane as it flies, but this is the first one to pull random parts from the engine at 30,000 feet. … One might have expected that the United States would respond to the age of weaponized interdependence as it responded to the earlier era of nuclear proliferation: by recalibrating its long-term strategy, building the institutional capabilities necessary to make good policy, and strengthening its global position. Instead, it is placing its bets on short-term dealmaking, gutting institutional capacity to analyze information and coordinate policy, and poisoning the economic and technological hubs that it still controls.

Now, however, the United States is spiraling into a rapid and uncontrollable drawdown of its assets, pursuing short-term goals at the expense of long-term objectives. It is increasingly using its tools in a haphazard way that invites miscalculations and unanticipated consequences. And it is doing so in a world in which other countries are not only developing their own capacities to punish the United States but also building technological stacks that may be more appealing to the world than the United States’. If China leaps ahead on energy technology, as seems likely, other countries are going to be pulled into its orbit. Dark warnings from the United States about the risks of dependence on China will ring hollow to countries that are all too aware of how willing the United States is to weaponize interdependence for its own selfish purposes.

I do have more to say about all of this, but it’s maybe better said on the back of a discussion of Dan Wang’s new book (which you should absolutely buy). More soon.