[ExI] Moore's Law

[Eugen Leitl]

On Wed, Mar 20, 2013 at 01:59:49PM -0700, Adrian Tymes wrote:

> > http://www.semiwiki.com/forum/content/1388-scariest-graph-i-ve-seen-recently.html
> 
> I have seen such graphs before.  It often turns out that they get their data

Great; I haven't, and not for a lack of looking. I've also received
several independent confirmation that we've entered difficult terrain,
so I think that was not a false blip.

> by comparing current prices - e.g., the current price of older tech, that has
> had years put into refining it and getting the cost down, vs. the current
> price of newer tech, that is much fresher from the lab.  I don't see anything
> in the article that suggests this is not the case.
> 
> This is a classic argument against any new solar tech, with the implicit
> assumption that there won't be successful R&D to lower the costs - or
> that costs can come down with larger manufacturing blocks - and
> therefore that solar can never become cost-competitive with fossil fuels.
> 
> >> present day - at least for actually released hardware.  There is no
> >
> > Yeah, I like to call things before most people notice.
> 
> So have a lot of others who said Moore's Law was dead, over the
> past few decades.

You might notice that until the end of the last year 
I haven't. 
 
> > Nope. You need about nm^3 for an addressable bit.
> 
> en.wikipedia.org/wiki/Spintronics suggests one can make an
> addressable bit in less size than that.  Or effectively so:
> multiple bits per atom.

How are you going to read out that state? How are you going
to power the system, and cool it? It's pretty easy to tweak
a nitrogen vacancy in diamond, but unfortunately you need
a roomfull of equipment for that. Do try to build an array
of bits each less than 1 nm across, using full atomic
detail. No longer plenty of room at that particular bottom.
 
> >> http://en.wikipedia.org/wiki/Moore%27s_law has data that shows it
> >> continued at least through 2011.  So if Moore's Law has ended, one
> >
> > We've got March of 2013. 2011 is ancient history.
> 
> Not so much.  The late-'90s dip lasted longer than 2 years.

Look, if you're not buying it, I don't particularly care.
It definitely does influence my model of how the next couple
decades will look like. See 
http://lp-hp.com/blog/2012/08/09/the-easy-stuff-is-over/
for another person's view on the situation.
 
> > I say rather firmly that you haven't been been looking at the data.
> 
> And I say you are looking at a smaller subset of the data than is
> relevant, and seeing a pattern that more data casts doubt upon.

It's okay to be contrarian, if you're being intelligently so.
Feel free to shoot us data which show that the price scaling
will continue. It would be good news, since it would buy us
a few more doublings at constant time.

However, I fear you won't be able to find data to cheer me up.