On Mon, 26 Mar 2001, Emlyn wrote:
> I'm betting on bio & computer science to beat any of these; I think
> biological and intensive computional methods will beat out so-called
> intelligent design, to all the big stuff. Of course, these methods will be
There's no point in compartmentalizing the different areas of human
enterprise, they're all one team.
> extremely intelligently designed in themselves, and the direction of humans
> in the process will be vital, up to the point where gaussians become
Unfortunately, I've ceased to expect innovation from Comp. Sci. types.
That field stagnates across the border, a real pity.
> Much of my justification for this point of view is that we have a big push
> in the biotech area now, which doesn't look likely to abate soon. This
> emerging field is necessarily computationally intensive; see this fluffy NYT
> article, which talks about the fact that biology is using the most computer
> time at the moment of all the sciences:
And they're only doing sequence yet. Very little structural work so far.
> Top quote from the article:
> "If I were 21 years old, I probably wouldn't go into computing. The
> computing industry is about to become boring. I'd go into genetic
> engineering." - Larry Ellison
I'd go into nanotechnology.
> Major computational resources, the smarts to use them, biotech
> (cough*nanotech*cough) and a LOT of dollars will combine into something
> which I think will beat the cognitive science approach to AI/IA/etc. We are
Are you talking about robust AI Minsky or Lenat style? That's pure snake
oil they're peddling.
> going to get the aeroplane before the bird. And, it sure as hell is not
> going to be hand coded, although the systems that generate the systems that
> generate it may be, maybe.
I agree very much. However, you'll observe that most of your peers don't
(yet). So it will take a while still, unless some unknown hotshot whips up
a killer demo, galvanizing the rest of them into frantic action.
Meanwhile, I'm not holding my breath.
This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:59:43 MDT