We keep making the same mistake: seeing exponential growth and assuming it will continue forever.
Moore’s Law. Nuclear stockpiles. The Black Death. Population bombs. Dot-com valuations. COVID deaths. AI capabilities. Every exponential story follows the same arc: explosive growth, breathless extrapolation, then reality reasserts itself.
What we call “exponential growth” is actually a series of logistic curves stacked on top of each other. Moore’s Law wasn’t one smooth exponential—it was six distinct technological breakthroughs, each following its own S-curve: planar transistors (1959), MOSFETs (1964), silicon gates (1971), high-density channels (1977), advanced integration (1989), deep-UV lithography (1990s). Together they created the illusion of sustained exponential growth.
The pattern is always the same. Early phase: explosive growth that looks exponential. Late phase: constraints kick in, growth slows to a power law. The exponential was never real—it was just the steep part of a logistic curve viewed through the wrong lens.
Nature doesn’t do exponentials. It does power laws and logistics. Exponentials represent disequilibrium—systems temporarily escaping their constraints before reality catches up.
This matters because exponential thinking makes us systematically wrong about the future. We project today’s growth rate indefinitely and miss the real question: What happens when this logistic curve hits its ceiling? What new breakthrough enables the next S-curve?
The AI discourse is drunk on exponential thinking. Capabilities will explode indefinitely! The singularity approaches! But if history is any guide, AI development will follow the same pattern: periods of rapid improvement followed by plateaus, new breakthroughs enabling new curves, constraints we can’t yet see eventually asserting themselves.
The useful question isn’t “How long will this exponential continue?” It’s “What stable state emerges after this transition period ends?” Better yet: “How do we navigate the transition itself?”
Exponential extrapolation has never worked. The future belongs to those who understand logistics.