[FoRK] Is the singularity now only 13 years distant?

Eugen Leitl <eugen at leitl.org> on Fri Mar 30 11:17:48 PDT 2007

On Fri, Mar 30, 2007 at 05:36:14PM +0000, Russell Turpin wrote:

> I remember the heady 90s. The internet was going to change everything,

It did, and it's by no means over yet. I intend to live there, 
eventually. (Of course it no longer will be the Internet as we know it).

> even my refridgerator. Nanotechnology was about to make clothes washers
> obsolete, and soon would cure cancer and heart disease. And even those

It will, one day. We don't have nanotechnology yet. You just bought
into the hype. (Don't fret, the mainstream has just discovered
machine-phase, and are quite breathless and flushed about it).

> advances would be overshadowed by the singularity, which was destined

The singularity is *caused* by the sum of the above, the good
old NBIC convergence, in fact. None of the nano, bio, info, cognitive
advances have happened yet. And most of them will likely happen much
later than anticipated, and elsewhere.

> to overtake us by no later than 2020. All the curves proved that.

If you swallowed the wrong brand of koolaid, yes. Some people
(Kurzweil, ahem) still peddle the same doctored goods. Don't
blame the snake oil salesmen, blame yourself. Cargo cult
is alive and well in futurist circles.

Here's a nice linear semi-log plot for you: the difference
between CPU speed and memory bandwidth is that, too.
> Anyone want to forecast the next singularity?

Trying to forecast something critically depending on a 
series of inventions is... rather foolish. The critical component
of progress autoamplification is AI, and specially AI running
evolution on fast-forward (superrealtime). You need both AI 
and reasonable (a mole of logic gates is reasonable) hardware for that.
E.g. a human equivalent is topping out at about 10^17 sites 
(some 10^2..10^3 bits each), and 10^23 ops on these sites 
(which boils down to refreshes at 10^3 Hz).

Put that into your Moore doublings pipe, and puff it.

Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
ICBM: 48.07100, 11.36820            http://www.ativel.com
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE

More information about the FoRK mailing list