The singularity is near says this, but it seems that the rate of Moore's Law has been pretty constant (cost version: x2 performance/price every 18 months). Whe

The singularity is near says this, but it seems that the rate of Moore's Law has been pretty constant (cost version: x2 performance/price every 18 months). When Ray extends this back to discrete transistors, vacuum tubes, relays, and mechanics, there's a case for an acceleration in rate - but not within silicon. If hardware is updated at different rates in different categories, I think it's due to factors specific to that category. Counter-example: the update rate for game consoles is much slower now than it was previously.