3 ms·

> Endogenous growth theory, ... implies that long-term economic prosperity depends on the spread of knowledge and technological innovation What if Moore's law

> Endogenous growth theory, ... implies that long-term economic prosperity depends on the spread of knowledge and technological innovation What if Moore's law wasn't really an exponential, but more of a step function over the past 50 years. What if it was a one-time fluke that greatly advanced society, but we're gearing up for a long haul of flat or linear growth? It explains a lot, including why China is able to quickly catch up to the U.S. in tech. If U.S. tech was really growing exponentially, it would be impossible for China to gain ground even if it was growing at an equal exponential rate. It also explains why iPhone sales are down. It isn't because of lower consumer demand or market saturation, those are just symptoms of a larger problem. It's because the rate of innovation is slowing down. We've relied on Moore's law of exponential growth for three generations, but it's sadly dying (if not dead). The ramifications will affect everything.

nopinsight 8y ago
Catching up is usually faster than pushing the state of the art further. I think the smartphone industry’s slower growth is more the result of diminishing returns from the form factor than the cessation of Moore’s law.
- speedplane 8y ago
  > Catching up is usually faster than pushing the state of the art further. Not so for exponential growth. e^x will always be greater than e^(x-t).
  - melling 8y ago
    You missed the part where they can simply borrow to catch up: “One is that as Chinese technology converges on that of advanced countries, the room for rapid improvement through borrowing declines”
  - nopinsight 8y ago
    The base for an economic growth model does not need to be e and generally would not be the same for different economies. The base for catching up economies is usually higher than developed economies (say 1.2 vs 1.1). In any case, a simple exponential equation is too simplistic for real economies.
tim333 8y ago
Moore's law is dandy and all that but doesn't have a huge effect money wise. Back in the 80s you might spend £1000 on a pc and these days you might spend that on an iphone which does more cycles per second but still most of your money is going on rent food and non tech things. Also while Moore's "the number of transistors in a dense integrated circuit doubles about every two years" may be ending, the more significant rising number of cycles per second per dollar will probably continue one way or another. I daresay the effect of tech money wise will do a step function if we hit super human AI. Rather than saying to a builder can you work on my house and him saying maybe Tuesday week I can do something, which doesn't change much if he has an iphone of not we'll be able to say robot, build me another 1000 robots and then a huge palace. Maybe.

nopinsight 8y ago

Catching up is usually faster than pushing the state of the art further. I think the smartphone industry’s slower growth is more the result of diminishing returns from the form factor than the cessation of Moore’s law.

speedplane 8y ago
> Catching up is usually faster than pushing the state of the art further. Not so for exponential growth. e^x will always be greater than e^(x-t).
- melling 8y ago
  You missed the part where they can simply borrow to catch up: “One is that as Chinese technology converges on that of advanced countries, the room for rapid improvement through borrowing declines”
- nopinsight 8y ago
  The base for an economic growth model does not need to be e and generally would not be the same for different economies. The base for catching up economies is usually higher than developed economies (say 1.2 vs 1.1). In any case, a simple exponential equation is too simplistic for real economies.

tim333 8y ago

Moore's law is dandy and all that but doesn't have a huge effect money wise. Back in the 80s you might spend £1000 on a pc and these days you might spend that on an iphone which does more cycles per second but still most of your money is going on rent food and non tech things. Also while Moore's "the number of transistors in a dense integrated circuit doubles about every two years" may be ending, the more significant rising number of cycles per second per dollar will probably continue one way or another. I daresay the effect of tech money wise will do a step function if we hit super human AI. Rather than saying to a builder can you work on my house and him saying maybe Tuesday week I can do something, which doesn't change much if he has an iphone of not we'll be able to say robot, build me another 1000 robots and then a huge palace. Maybe.