The Combination Ceiling

The Combination Ceiling names the structural constraint that Ikujiro Nonaka's framework imposes on any account of AI productivity gains: the Combination mode of the SECI spiral—the reconfiguration of existing explicit knowledge into new explicit knowledge—is the mode that large language models accelerate with unprecedented power, and it is also the least creative mode of the spiral, entirely dependent on the other three modes for both its raw material and the conversion of its outputs into genuine organizational knowledge. Combination takes explicit inputs and produces explicit outputs; it cannot generate the tacit knowledge that makes those outputs meaningful, because tacit knowledge is deposited through Socialization and Internalization—modes the machine does not and cannot perform. The ceiling is reached when an organization's Combination outputs increase while its tacit foundation thins: features ship faster, code compiles, analyses arrive on schedule, but the practitioners producing them have less embodied confidence in their architectural soundness, less felt sense of system behavior, less capacity to evaluate the machine's outputs