4 ms·

This is the first I hear about the "slowdown of Moore's Law." Do you have any data to share? Thx.

teraflop 8y ago
Moore's law originally stated that the number of transistors per IC would double every 18 months. For a long time this was a remarkably accurate empirical law, but it has slowed down dramatically in the last decade or so. As an individual example: in 1998, Intel's top-of-the-line CPU was the Pentium II with about 8 million transistors. In 2008, they released the Core 2 Extreme QX9770, with about 820 million transistors; this is a 100-fold increase in 10 years, which is almost exactly what Moore's law predicts. Fast-forward another decade to 2018, and Intel's highest-end Xeon chips have something like 7 billion transistors, which is less than a 10-fold improvement in the same timeframe. EDIT: that said, I think it's a real stretch to go from "the slowdown of Moore's law" to a sweeping claim about innovation in general being more difficult. There are a ridiculous number of different fields in which someone could start a business; relatively few of them are dependent on enormous, ever-increasing amounts of computational power.
- tracker1 8y ago
  I'd go even further to say, that we're at a point where not EVERYTHING needs to be smaller... our mobile devices are so small and difficult to hold, we're putting things on the back to make them easier to hold. They're so thin and fragile, we're putting thick cases on to protect them... I'd love to see a production cell phone that is already thick with a relatively huge battery, more rugged case, and a slightly better protected display. I'm willing to bet there's a market for this, just based on otterbox sales, and the sales of backpack battery covers. Similarly a home media player like the roku, shield tv or fire stick are plenty small. That has nothing to do with being able to innovate and create newer or better experiences.
modeless 8y ago
From 1987 until 2011, Intel created a new manufacturing process every 2 years like clockwork. In 2013, their 14nm process was delayed one year to 2014. The next process, 10nm, is delayed 3 years, assuming it is not delayed again before the currently planned late 2019 release. https://en.wikichip.org/wiki/intel/process https://en.wikichip.org/wiki/intel/process

Moore's law originally stated that the number of transistors per IC would double every 18 months. For a long time this was a remarkably accurate empirical law, but it has slowed down dramatically in the last decade or so. As an individual example: in 1998, Intel's top-of-the-line CPU was the Pentium II with about 8 million transistors. In 2008, they released the Core 2 Extreme QX9770, with about 820 million transistors; this is a 100-fold increase in 10 years, which is almost exactly what Moore's law predicts. Fast-forward another decade to 2018, and Intel's highest-end Xeon chips have something like 7 billion transistors, which is less than a 10-fold improvement in the same timeframe. EDIT: that said, I think it's a real stretch to go from "the slowdown of Moore's law" to a sweeping claim about innovation in general being more difficult. There are a ridiculous number of different fields in which someone could start a business; relatively few of them are dependent on enormous, ever-increasing amounts of computational power.

tracker1 8y ago
I'd go even further to say, that we're at a point where not EVERYTHING needs to be smaller... our mobile devices are so small and difficult to hold, we're putting things on the back to make them easier to hold. They're so thin and fragile, we're putting thick cases on to protect them... I'd love to see a production cell phone that is already thick with a relatively huge battery, more rugged case, and a slightly better protected display. I'm willing to bet there's a market for this, just based on otterbox sales, and the sales of backpack battery covers. Similarly a home media player like the roku, shield tv or fire stick are plenty small. That has nothing to do with being able to innovate and create newer or better experiences.

modeless 8y ago

From 1987 until 2011, Intel created a new manufacturing process every 2 years like clockwork. In 2013, their 14nm process was delayed one year to 2014. The next process, 10nm, is delayed 3 years, assuming it is not delayed again before the currently planned late 2019 release. https://en.wikichip.org/wiki/intel/process https://en.wikichip.org/wiki/intel/process