Weekly Digest: Five Things Worth Your Time

What we're watching

The Federal Reserve held rates at 5.25-5.5% at this week's FOMC meeting, as expected. Chair Powell's press conference was notably non-committal on the September path, emphasizing 'data dependence' without providing the forward guidance markets were hoping for. The August jobs report and the August CPI print will now be the decisive inputs for whether September brings a cut. Markets are pricing a 60% probability of a September cut, which will move significantly with the next two data releases.

Vietnam and the Philippines reached a bilateral maritime agreement this week that includes joint development zones in disputed South China Sea areas. It's a small but meaningful signal that Southeast Asian nations are finding ways to manage territorial disputes without defaulting to either confrontation or full acquiescence to Chinese claims. The US-China competition in the region is increasingly being shaped by the agency of smaller nations rather than just the two large powers.

Metaverse-adjacent investments by major tech companies have now been largely written down or quietly discontinued. Meta's Reality Labs has lost over $50 billion since 2021. Microsoft disbanded its industrial metaverse team. The consensus has shifted from 'when' to 'whether' for immersive spatial computing at scale. The AI pivot has completely absorbed the strategic attention and capital that was briefly allocated to VR/AR buildout.

One number

420 million. That's the estimated number of people globally who will enter the working-age population (15-64) over the next decade, nearly all of them in sub-Saharan Africa, South Asia, and parts of the Middle East. For context, the entire US labor force is roughly 165 million people.

This demographic reality is the most important long-run variable in global economic development. Whether those 420 million people enter productive formal employment or remain in low-productivity informal work will determine the distribution of global growth, the stability of governments across multiple continents, and the direction of migration flows that shape politics in wealthy countries.

One idea

The difference between risk and uncertainty, as defined by economist Frank Knight in 1921. Risk involves outcomes that are unknown but whose probability distribution is known or estimable (rolling a die). Uncertainty involves outcomes where even the probability distribution is unknown (predicting a geopolitical crisis).

Most financial models treat uncertainty as if it were risk: they assign probabilities to scenarios even when the basis for those probabilities is essentially invented. Knight's distinction matters because the tools for managing risk (diversification, hedging, insurance) don't work for genuine uncertainty. Managing uncertainty requires flexibility, redundancy, and the preservation of optionality rather than optimization for a specific probabilistic scenario.

Worth reading

The FT's year-in-review on China's economy: what the property sector deflation has actually meant for Chinese household balance sheets, and why the government's stimulus response has been more cautious than most Western observers expected: https://ft.com/content/china-economy-property-deflation-review

A reported piece on the second-order labor market effects of AI: the new jobs being created in AI training, data labeling, and model evaluation, mostly in the Global South, that are becoming a significant new category of digital work: https://nytimes.com/2025/07/ai-labor-global-south-new-jobs

A long essay on why the semiconductor industry's 'ecosystem' structure, where the top 10 companies depend on thousands of specialized suppliers, makes it simultaneously robust and fragile in ways that neither purely market nor purely geopolitical analysis captures well: https://economist.com/technology/2025/semiconductor-ecosystem-fragility