The technology sector is currently navigating a period of extreme volatility, characterized by record-breaking revenues alongside massive workforce reductions. A growing theory suggests that this instability is fueled by what some industry leaders call ‘AI psychosis’—a state where executives, detached from the granular realities of daily operations, develop unrealistic expectations regarding the capabilities of artificial intelligence.

Box founder Aaron Levie recently highlighted this disconnect, noting that CEOs often interact with AI only through high-level prototypes or simplified tasks. Because these leaders are removed from the ‘last mile’ of work—such as debugging code, verifying contract terms, or identifying model hallucinations—they frequently mistake the potential of AI for immediate, plug-and-play operational efficiency. This lack of hands-on experience leads to a dangerous overestimation of what autonomous agents can achieve without human oversight.

This executive optimism is manifesting in tangible, often disruptive ways. In 2026 alone, over 115,000 tech workers have been laid off, with many companies citing AI integration as a primary driver. Some leaders, such as ClickUp CEO Zeb Evans, have openly restructured their organizations to prioritize AI agents over human staff, aiming for a ‘100x’ productivity model. However, current academic research, including studies from UC Berkeley and the National Bureau of Economic Research, suggests that there is no robust evidence linking current AI adoption to aggregate productivity gains, pointing instead to a ‘productivity paradox’ where perceived benefits far outweigh actual output.

As organizations rush to automate, experts warn that the bottleneck is simply shifting to the executive level, where leaders are becoming overwhelmed by the sheer volume of AI-generated content. With MIT researchers predicting that AI will not reach human-quality performance across most text-based tasks until at least 2029, the current trend of aggressive automation may be premature. Without a grounded understanding of AI’s limitations, the industry risks trading sustainable growth for organizational chaos.