CONCEPT
The Convergence Machine
Edward de Bono’s diagnosis of the default human-AI
collaboration: a bilateral self-organizing system in which the builder’s patterns and the machine’s training distribution reinforce each other, producing increasingly polished output that occupies an increasingly narrow region of possibility space.
The default AI collaboration is not a creative partnership. It is a
convergence machine. The builder brings a problem shaped by her existing patterns. The AI brings its training distribution, in which the high-probability paths were carved by the statistical patterns of millions of prior outputs. The two converge on the response that their combined pattern most naturally produces—competent, smooth, and indistinguishable from every other response that two similarly configured systems would have reached. The output is the first idea at computational scale: not wrong, not useless, but shaped entirely by the channels already established in both systems.
Edward de Bono’s analysis of the self-organizing pattern system explains why this convergence is not a failure of the collaboration but a structural property of it. Any system powerful enough to organize information at scale will tend toward its own center; the
self-organizing pattern trap is not a bug in
large language models but