The supercycle doesn’t move evenly — it moves layer by layer, and the binding constraint rotates. Five analytical instruments cut different cross-sections of the same terrain.
The Five Instruments
The Three Tribes
Substrate, Surface, Sense — three ways companies relate to AI. Substrate companies rebuild the foundation. Surface companies repaint the interface. Sense companies generate the data. Know which tribe you’re in.
Four Scaling Paradigms
Pretraining, post-training, inference-time, agentic-loop. Each paradigm has different economics and different winners. The paradigm shift from inference-time to agentic-loop is happening now.
Cognitive Range
How far AI capability extends across the spectrum of human cognition. Strategic fluency vs tactical execution. The range is expanding — and each expansion kills a job category.
The FRED Test
Frequency, Repeatability, Error tolerance, Decomposability. Score any task on these four dimensions to predict whether AI automates it. High FRED score = automated first.
What’s Next
Three horizons are converging: the SaaS mutation window closes by 2028, the control plane hardens in 12-24 months, and the geopolitical floor shifts over 5-10 years as new rare earth processing comes online. The supercycle doesn’t move evenly — the binding constraint rotates.
The Bottom Line
The supercycle isn’t one story — it’s nine layers moving at three speeds with a ceiling that can shut any layer down. The toolkit gives you five ways to read it. The question is whether you’re reading it before or after the constraint rotates to your layer.









