The Financial Times has an interesting article on the supposed dominance of the CCP in AI. Via Reuters:
Scary stuff, but how true is it?
First, Chaillan appears to be right about increasing Chinese dominance in AI. A paper from last year found that in AI research:
So it looks like the CCP has us beat on quality and, especially, the quantity of research, with the US-dominated global academic system not quite recognizing it yet. Insofar as the US is still competing, it’s doing so by importing two-thirds of its AI researchers from other countries. (Send me your huddled masses yearning to breathe free… so long as they’ve got a graduate degree in comp sci.) Chinese scientists have also beat Google to a breakthrough in quantum computing and are supposedly able to solve certain math problems about 100,000,000,000,000 times faster than our fastest supercomputer.
So how did we let this happen?
The great advantage of the political economy of the CCP is that it makes it easy to mobilize the resources of the nation towards clearly defined goals. In America, many individual actors, such as academic institutions, congress, corporations, and the administrative state, have their priorities, which they are unwilling to sacrifice for inscrutable AI research. In China, it is much easier to solve these coordination problems: Comrade Xi can decide that AI research is an important priority and throw as many resources at the problem as necessary to produce, or at least appear to produce, results. The US government is capable of doing the same thing when it needs to, as evidenced by the Manhattan Project and Operation Warp Speed, but the Chinese model of governance lends itself to these sorts of projects. This is one of the system’s appeals, and the reason why prominent American journalists, such as Thomas Friedman, have advocated that the US imitate parts of the Chinese system while eschewing its authoritarianism.
The drawbacks of this system are that, while a centralized system is excellent at producing well-defined outcomes, it is awful at long-term innovation. If everyone is taking marching orders from the top, then there is no room for people at the bottom to research and innovate. True, paradigm-breaking innovation is not the sort of thing that a central politburo can order; it often appears in the form of a guy tinkering in his garage or with a professor deserting from the reigning orthodoxy.
Future historians may consider this our “missile gap” moment, similar to when the Soviet Union had apparent technological dominance over the US in terms of missiles. This, combined with their apparent dominance in space, led to a perception of Soviet technological superiority. But the Soviets proved incapable of the bottom-up innovation that spurred the digital revolution in the US, and thus began to lag further and further behind in the tech race as the 20th century drew on.
The Chinese model of authoritarian capitalism is designed to avoid these sorts of problems. If the market is allowed to set prices and allocate resources with the government determining the ends to which these resources are dedicated, then, in theory, the CCP should be able to harness the power of the market to achieve its ends. In practice, however, these competing priorities often create a Jack Ma or Evergrande situation, where the CCP has to choose between economic growth and maintaining its own power.
However, even on the object level considerations, we shouldn’t be too worried about Chinese tech dominance. For instance, semiconductors are a commonly cited area of supposed Chinese dominance, but a recent report found that:
The US semiconductor industry is estimated to contribute 39 percent of the total value of the global semiconductor supply chain… Japan, Taiwan, Germany, and South Korea manufacture the state-of-the-art 300 mm wafers used for 99.7 percent of the world’s chip manufacturing… China controls the largest share of manufacturing for most natural materials. The US and its allies have a sizable share in all materials except for low-grade gallium, tungsten, and magnesium. China controls ~2/3rds of the world’s silicon production, but the US and allies have reserves.”
The rest of the report reinforces this, with the Chinese controlling most of the raw materials, while the US and allies control most of the technology. (One surprising finding is that we are more dependent on the Netherlands than on China for our semiconductor supply chain.) This hints at one final reason to be skeptical of the coming Chinese Tech Dominance stories: our allies. The US and China aren’t competing in a vacuum; they are surrounded by other developed countries, most of whom are much more comfortable under the US-led world order than a Chinese-led one. Even countries like India and Japan, who might challenge US hegemony in other circumstances, are sufficiently antagonized by China that they are likely to continue to prefer the US. Even if the US’s relative standing in technical dominance declines, so long as China cannot match the rest of the developed world, we will still have an advantage.
So how worried should we be about Chinese tech dominance? On the one hand, we need to step up our game, whether on AI or hypersonic missiles, to ensure that the Chinese don’t gain a military advantage against us. On the other hand, we should keep in mind that the inherent flaws in the Chinese political-economic system, the current state of industries such as the semiconductor industry, and China’s lack of developed allies all likely mean that the supposed era of Chinese dominance is hardly foreordained.