A critical asymmetry exists in the US-China competition: It is far harder for the U.S. to rebuild its complex manufacturing ecosystems and tacit process knowledge than it is for China to improve its scientific research capabilities, where it is already making significant strides.

Critical manufacturing expertise is not easily codified in manuals; it's tacit knowledge embedded in experienced teams. Offshoring production leads to an irreversible loss of this 'process capital,' hindering a nation's ability to innovate and scale complex industries, as demonstrated by the transfer of German rocket scientists after WWII.

The belief that a country can innovate while others produce is a fallacy. Innovation is a consequence of production; solving small, banal production problems leads to revolutionary breakthroughs. By outsourcing manufacturing, the US has seeded its innovative capacity to rivals like China, who graduate from production to creation.

Critical process knowledge often exists only in operators' heads—the nuances and undocumented tricks that ensure success. Systematically capturing this 'tribal knowledge' during tech transfer is crucial for preventing hard-to-diagnose failures at the new site.

China's durable advantage isn't just its massive workforce but the collective "process knowledge" generated on factory floors. This expertise in solving countless small manufacturing problems cannot be easily written down or encoded in equipment, creating a powerful, hard-to-replicate competitive moat.

Arm's CEO argues the US has lost its 'muscle memory' for 24/7 manufacturing. The core issue is cultural: manufacturing isn't seen as a prestigious career, unlike in Taiwan where working for TSMC is highly esteemed. This cultural gap is a major hurdle for onshoring efforts.

Manufacturing faces a crisis as veterans with 30+ years of experience retire, taking unwritten operational knowledge with them. Dirac's software addresses this by creating a system to document complex assembly processes, safeguarding against knowledge loss and enabling less experienced workers to perform high-skill tasks.

When enterprises hire external firms, they outsource not just costs but also institutional knowledge. AI platforms can reverse this by capturing learnings from external engagements, building a proprietary 'brain' for the company and keeping knowledge in-house.

A company's unique operational processes and judgments—its "tacit knowledge"—is a new form of IP called "token capital." Feeding this to external AI models is a one-way door to losing competitive advantage. This capital must be cultivated and protected internally.