Paradigm capability matrix

Neuromorphic AI is strongest where continuous adaptation and sparse real-time sensing matter more than dense batch inference.

Recursive self-improvement is less a substrate than a dangerous and powerful control loop over capability growth.

Analog AI can win when approximate computation is acceptable and data movement dominates energy.

Hybrid intelligence is the likely practical path: use each paradigm where its physics and economics are strongest.

Maths AI treats intelligence as an evolving mathematical ecosystem rather than a static chatbot: recursive interaction plus proof, code, simulation, compression, and selection pressure.