[ExI] Ray Kurzweil's predictions regarding the Singularity

[Damien Broderick]

On 11/29/2010 9:24 PM, John Grigg wrote:

> According to someone from another list who has a spouse in the
> supercomputer field, we should be seeing a 30 petaflops machine by
> around 2014.  And so Ray Kurzweil turns out to be just a few years
> off!  : )

Here's what I wrote in the 2001 edition of THE SPIKE:

< machines already exist that do several trillion calculations per 
second. So our current devices are 33,000 times too feeble for the job. 
We need to increase computing power from 3 teraflops to 100,000 
teraflops (or, more concisely, 100 petaflops)--a whopping factor. 
Impossible, surely?
	Not at all. Exponential growth cuts numbers like that down to size... A 
mere fifteen doublings meets that goal nicely. >

With annual doubling, that would predict... 2016. For 100 petaflops. 
2014 looks right for 30 petaflops.

Damien Broderick