daily notes

previous notes →

Must read

Measuring AI AdoptionJohn Haltiwanger, interviewed by Tim Sablik

The transcript uses a Census Business Trends and Outlook Survey covering about 200,000 firms and finds adoption concentrated by sector and firm size, with about 70% of users reporting augmentation only and 2% reporting AI job cuts.

Sentiment

Capability evidence with deployment constraints +0.18

71 source items · 55% editorial confidence

Must read

Signals

1/2 Michael Froman: "Perhaps the most significant, if implicit, subsidy is Chinese manufacturing firms’ access to Beijing’s state-directed financial system, which channels vast credit to prioritized sectors, enabling Chinese firms to expand without the same concern for profit and return as their international peers." This helps show why rebalancing is so difficult. Raising interest rates and cutting off financing to unprofitable producers would set off a self-reinforcing process in which a large number of producers would have to close down facilities and fire workers which, in turn, would reduce domestic demand further. https://t.co/dq6jtEwkt9 via @ForeignAffairs

discussion1 selected reply
@michaelxpettisreply ↗

2/2 This is how the US "rebalanced" in the early 1930s, and if China were to do the same, it would be painful not just for China but, as its imports dropped sharply, also for its trade partners and especially for countries who in the past 25 years, as commodity prices soared, had recentered their economies around commodity production. China and the world would be better off with a gradual rebalancing, spread out over one or two decades, but while this would have been relatively easy ten or fifteen years ago, rising debt and increasing pressure on trade partners have made it increasingly difficult. https://t.co/1B3KgMQsGF

The post quotes research arguing that Chinese bank credit directed to firms and investment can preserve production capacity without fixing weak consumer demand. It presents a mechanism for why cheaper credit may worsen supply-demand imbalances, not a forecast of the next policy move.

Sentiment

Rates, credit, and infrastructure risk -0.08

417 source items · 55% editorial confidence