Why Didn't AI "Join the Workforce" in 2025? (calnewport.com)

A year after Sam Altman predicted that AI agents would join the workforce by 2025, expectations have not been met. While Altman and OpenAI's Chief Product Officer, Kevin Weil, envisioned sophisticated AI that could take on complex tasks akin to human employees, the reality has shown a notable gap. Current AI products, such as the ChatGPT Agent, have struggled with basic tasks, failing to meaningfully transform professional workflows. Critics like Silicon Valley skeptic Gary Marcus argue that the technology underlying these agents is still too rudimentary, with large language models lacking the capability to fulfill the lofty promises made. The optimism surrounding 2025 has given way to skepticism, marking a shift in focus for the AI community. As we move into 2026, there’s a call to reevaluate our approach to AI, shifting from speculative predictions to a more grounded understanding of current capabilities. This pivot away from dreams of AI revolutionizing the workforce towards a more pragmatic assessment encourages a realistic dialogue about the implications of existing AI technologies, instead of worrying about unproven potentials. The conversation now centers on leveraging the functions AI can currently provide and responding critically to forecasts that may not align with present-day achievements.