Welcome to Markets Desk, where we break down what matters across business, tech, and the broader economy.
A sharp piece out of Forbes is raising important questions about how we evaluate AI in sensitive settings. When researchers test AI mental health tools, they typically run isolated, stateless prompts, but real users engage in extended, evolving conversations. That gap means safety assessments may be fundamentally misleading, and the risks far greater than reported benchmarks suggest.
Shifting to the policy frontier, a large-scale survey experiment published through Marginal Revolution examined American attitudes toward human enhancement technologies. With over five thousand five hundred respondents, the research finds that private demand for these tools, things like cognitive or physical augmentation, does not reliably reflect how citizens actually want them regulated. When social costs go unpriced by markets, consumer choice becomes a poor proxy for public preference, a classic externality problem with real governance stakes.
And on the equity side, Amazon heads into its July thirtieth earnings report with a stock that has essentially flatlined on the year, a notable underperformance given the broader market. Investors are watching closely for signals on cloud growth, advertising momentum, and any updated guidance on AI infrastructure spending, all of which could serve as the catalyst the stock has been waiting for.
That's the tape. Markets Desk, signing off the floor.
