Fri 27 March 2026

It is clear, as The Atlantic argues, that the Middle East Crisis Turns AI Boom Into Systemic Macro Risk. The AI boom now sits on a fragile stack of exposed supply chains, soaring capex, and heavy debt, so a prolonged Middle East conflict could disrupt energy, chips, and data-center buildouts, squeeze hyperscalers and private credit, and spill from tech into the broader economy and markets.

This is really big! China has overtaken the U.S. in total token usage as next‑generation AI agents, like those built on OpenClaw, consume vastly more tokens than traditional chatbots. Because these agents process millions of tokens daily, small cost differences have major financial impacts. China’s advantage comes from lower energy costs and more efficient model architectures, allowing companies such as MiniMax and Moonshot to offer tokens for just $2–$3 per million — compared with about $15 for Anthropic’s Claude Sonnet 4.5 — giving Chinese labs a growing edge in the global AI race.

Anthropic's latest economic impact report, while reassuring that AI has not yet triggered widespread job displacement, subtly underscores a burgeoning skills divide that could exacerbate workforce inequalities. As an AI firm, Anthropic tempers CEO Dario Amodei's dire predictions of 20% unemployment by noting stable unemployment rates across AI-exposed and non-exposed roles, yet early data reveals power users—predominantly in high-income regions and knowledge-heavy occupations—deriving outsized productivity gains from tools like Claude. This uneven adoption, where novices lag in leveraging AI for sophisticated tasks like iterative feedback, risks entrenching advantages for the already skilled, potentially sidelining younger entrants and amplifying geographic disparities. Skeptically, the report's emphasis on monitoring frameworks may serve promotional ends, but it rightly flags the need for proactive policy to address diffusion lags before displacement accelerates in the next five years. Key points: • AI has not caused significant job losses yet, with stable unemployment across exposed roles. • Early adopters gain a productivity edge by using AI as a 'thought partner' for complex tasks. • Skills gap is widening, favoring users in high-income countries and knowledge-worker hubs. • Potential for rapid future displacement, especially in entry-level white-collar jobs, up to 20% unemployment. • Uneven AI adoption could amplify existing inequalities rather than equalize opportunities. Expert question (counterfactual): What if targeted upskilling programs and equitable access policies could close the AI skills gap faster than anticipated, mitigating displacement risks without relying solely on market-driven adoption?