Anthropic's Co-Founder and Top Economist on Doing Research at the AI Frontier

Artificial intelligence is rapidly reshaping the global economy and labor market, driven by the "bitter lesson" that scaling compute consistently produces smarter, more capable models. Jack Clark and Peter McCrory of Anthropic highlight that while macroeconomic data currently reflects only modest productivity shifts, internal metrics show engineers writing eight times more code through automated agents. This technological leap creates a "barbell" hiring pattern, where senior domain expertise is significantly amplified, while early-career roles face potential displacement. Anthropic emphasizes the necessity of third-party auditing and transparency to manage national security risks, such as AI-manufactured bioweapons or cyber threats. As these models evolve into general-purpose technologies, companies must navigate the transition from traditional software implementation to managing an expanding cloud of autonomous agents, balancing rapid innovation with the critical need for safety and institutional alignment.