Exponential Assembly Scaling
Serial assembly throughput decreases as structure size grows — more time is spent traveling, not building. Self-replication and hierarchical assembly together reverse this: throughput increases exponentially with scale (Abdel-Rahman et al.).
The math
For a cube of side 2^N voxels:
- Serial (gantry or carrier): assembly time O(2^4N), throughput decreasing
- Self-replication only: assembly time O(2^3N), throughput constant
- Hierarchy only: assembly time O(2^3N), throughput constant
- Both combined: assembly time O(2^2N), throughput O(2^N) — increasing
The crucial insight: neither self-replication nor hierarchy alone is sufficient. Self-replication parallelizes but robots still travel far. Hierarchy reduces travel but a single large robot is still serial. The combination eliminates both bottlenecks.
Biological precedent
Ribosomal self-assembly in biological systems achieves throughput O(2^log(E)) — increasing with the number of elements. Only recursive + hierarchical robotic assembly matches this trend. Serial robotic assembly, no matter how fast, cannot.
Implication
Large constructions do not require large machines — they require machines that can make more of themselves at multiple scales (xettel).