AI Adoption, Not Capability, Is the Real Bottleneck Holding Back Progress

A commentary responding to the 'AI as Normal Technology' thesis argues that technological advancement and real-world adoption operate at fundamentally different speeds, with organizations struggling to keep pace with rapid AI development. The author contends that adoption remains the critical bottleneck, as integrating AI requires changes to processes, responsibilities, and organizational culture, not just access to capable models. In software development, for example, AI has accelerated coding but not overall product development, shifting the challenge from implementation to decision-making about what to build. The piece also highlights that AI adoption will be slower in high-stakes fields like healthcare, law, and finance, where issues of legitimacy and accountability matter as much as reliability. A key concern raised is that AI may accumulate indirect influence over human decisions even without formal authority, blurring the line between human judgment and machine recommendation.