Xi Jinping’s paranoid approach to AGI, debt crisis, & Politburo politics — Victor Shih

Dwarkesh begins with an apparent paradox: roughly 85% of government spending in China is local or provincial, compared with about half of US spending at state and local levels, which can make China look more decentralized than its authoritarian image suggests. Victor Shih says China was indeed highly decentralized from the mid-1970s to the mid-1990s. Local governments raised and spent substantial revenue, giving them incentives to attract foreign and domestic investment, offer tax breaks, and support a period of largely private-sector growth.

The 1994 tax centralization shifted the balance. The central government took control of the value-added tax and, over subsequent decades, most other major tax categories, then returned some revenue to provinces. Formal discretion over spending did not mean real fiscal autonomy: transfers could be conditioned on carrying out centrally preferred projects. Local governments later gained some independence through land sales during the real-estate boom, but the central government's 2022 intervention in the land market increased their dependence again.

Many Politburo members have engineering, economics, or science backgrounds, including figures who worked in state-owned enterprises or the military-industrial sector. Shih distinguishes technical expertise from the political skill required to rise through the Communist Party. Some technically trained officials also prove politically capable, but credentials alone do not predict who will govern effectively. He cites Li Keqiang's economics education, as well as engineers such as Ma Xingrui and Zhang Guoqing, while noting that other officials' technical training can be quite specialized.

Shih argues that political acumen is essential for advancement. Princelings may gain an advantage because their families can teach them how elite politics works. Graduates of top universities are overrepresented in the Politburo but do not dominate it, and selection remains partly unpredictable. Dwarkesh contrasts the leadership's strong technical credentials with errors such as prolonging Zero-COVID and asks why expertise did not correct the policy. Shih says technical advice can reach the top, but the party's instinct to preserve its power can outweigh public-health or economic considerations.

Shih gives two reasons Chinese leaders may make choices economists criticize. First, students are often tracked into STEM or the humanities, and science students may never study supply and demand. A required course on government ideology has become more prominent, but Shih says it teaches the state's preferred positions rather than economics or other practical skills. Even senior officials with advanced technical degrees may therefore lack a strong understanding of markets and society.

Second, authority has become more concentrated under Xi Jinping. Politburo members once had greater room to manage their own domains, but Shih says they now must follow Xi's preferred policy direction or risk being purged. This limits the scope for expertise or local experience to change course when a central decision proves costly.

Shih describes a highly hands-on decision process. Xi receives briefing books with policy options and must remain engaged across many fields because his decisions effectively become law. Shih says Xi's daily schedule includes frequent meetings, reportedly on policies for roughly 270 days a year, and that his post-lecture remarks can become policy. These accounts portray a leader deeply involved in details rather than delegating decisions to a broad cabinet.