Saturday, February 22, 2014

Can you accelerate science with more money?

Chip makers typically know what technologies they are going to use to develop the next two generations of chips.  So why don't they skip the intermediate generation and go to the second one - it would save time and money.  They don't because the technos to create a chip is more than the science - many of the things that you need to know to produce the second generation are learned by producing the intermediate one.  So a doubling of Research or Development spending wouldn't appreciably change Moore's law.

I wonder how far this notion extends because people are always arguing for 'more' research money to 'accelerate' progress.  But often progress can't be achieved in one area until certain things are solved elsewhere.  For example, in AD 1500 you could have put 1000 boffins full time on figuring out Celestial Mechanics and Gravity.  It wouldn't have helped because the tool you needed to figure it out:  the calculus wasn't invented until later, unsurprisingly by the same guy that came up with Classical Mechanics (Newton). (Leibniz invented calculus independently as well).

So when people make this argument, take it with a grain of salt.

No comments:

Post a Comment