Kurzweil's Five Paradigms of Computing
The exponential growth of computing survived five completely different underlying technologies—the trend transcends the paradigm.
Explore this event on the interactive timeline →Kurzweil's most powerful argument for the inevitability of accelerating returns: computing power per constant dollar has grown exponentially since 1890 across five fundamentally different technologies. When one paradigm exhausts itself, the next is already emerging. This proves that exponential growth is a property of the evolutionary PROCESS, not of any specific technology like transistors or Moore's Law.
Key Numbers
- Paradigms Survived
- 5 (since 1890)
- Speed Increase
- ~10 trillion-fold
- Sixth Paradigm
- 3D / Quantum (emerging)
Verified Facts
- Paradigm 1 — Electromechanical (1890-1940s): Hollerith tabulators, analog computers. Computation via mechanical gears and electrical switches.
- Paradigm 2 — Relays (1940s): Zuse Z3, Harvard Mark I. Computation via electromagnetic relay switches clicking open and shut.
- Paradigm 3 — Vacuum Tubes (1940s-1950s): Colossus, ENIAC. Computation via heated cathodes in glass tubes. 1,000x faster than relays.
- Paradigm 4 — Discrete Transistors (1950s-1960s): IBM mainframes. Solid-state switching replaced fragile vacuum tubes.
- Paradigm 5 — Integrated Circuits (1960s-present): Microprocessors, GPUs. Billions of transistors on a single chip. This is the paradigm of Moore's Law.
- Kurzweil's prediction: "When Moore's Law reaches its limits (around the 2020s), a Sixth Paradigm will emerge—likely 3D molecular computing or quantum computing." As of 2026, both are in active development (TSMC 3D chip stacking, Google/IBM quantum processors).