While there's a popular narrative about a US manufacturing resurgence, the massive capital spending on AI contradicts it. By consuming a huge portion of available capital and accounting for half of GDP growth, the AI boom drives up the cost of capital for all non-AI sectors, making it harder for manufacturing and other startups to get funded.

The most significant long-term threat to the supply of critical materials isn't a lack of resources in the ground, but a lack of people. The aging workforce of geologists and mining engineers, with a shrinking pipeline of new talent, poses a greater systemic risk to the industry.

During tech gold rushes like AI, the most skilled engineers ("level 100 players") are drawn to lucrative but less impactful ventures. This creates a significant opportunity cost, as their talents are diverted from society's most pressing challenges, like semiconductor fabrication.

Just as NAFTA brought cheap goods but eliminated manufacturing jobs, AI will create immense productivity via a new class of "digital immigrants" (AIs in data centers). This will generate abundance and cheap digital services but risks displacing vast swaths of cognitive labor and concentrating wealth.

AI will primarily threaten purely cognitive jobs, but roles combining thought with physical dexterity—like master electricians or plumbers—will thrive. The AI-driven infrastructure boom is increasing demand and pushing their salaries above even those of some Silicon Valley engineers.

Developed nations are building massive infrastructure projects like data centers, yet the construction workforce is aging and shrinking. This creates a critical bottleneck, as every project fundamentally relies on excavator operators—a role younger generations are avoiding.

AI is rapidly automating knowledge work, making white-collar jobs precarious. In contrast, physical trades requiring dexterity and on-site problem-solving (e.g., plumbing, painting) are much harder to automate. This will increase the value and demand for skilled blue-collar professionals.

Most AI applications are designed to make white-collar work more productive or redundant (e.g., data collation). However, the most pressing labor shortages in advanced economies like the U.S. are in blue-collar fields like welding and electrical work, where current AI has little impact and is not being focused.

Robert Kaplan suggests the labor market's sluggishness might not be a simple cyclical slowdown. He points to a significant "matching problem" where open jobs don't align with the skills of job seekers. This structural issue limits the effectiveness of monetary policy as a solution.