From: Robert Bradbury (bradbury@genebee.msu.su)
Date: Wed Jan 12 2000 - 09:35:59 MST
In response to my comments:
> Now, on the biotech front, we are going to have progammable, self-replicating
> machines *very* soon (1-3 years). The only problem will be they will not
> general purpose...
Eugene Leitl wrote:
>
> I think you're way too optimistic here. If we can get protein folding
> licked, then modification is suddenly doable. Then you can look at
> protein/DNA interaction, design new enzymes from scratch, etc.
I may be, but so were the Wright Brothers... :-)
Let me be clear about what I am saying. I'm not talking about designing
any *new* capabilities (new enzymes, etc.). I'm only talking about
taking existing tools and putting them together in novel and useful ways.
> Before that you can't predict the impact of your changes without doing the
> actual (expensive, slow) experiment.
Well, the typical experiment is slow (a day or two), but if you can do 1000
of them in parallel, then all you are concerned with is what 1000 to do and
how to do them cheaply.
> You still need a full cell model
> to tackle the hard stuff like embryomorphogenesis and such, but that's
> the logical next step after solving PFP.
I'm going to let the Zebrafish people work that all out before I go near it.
For my purposes, I don't think I need to understand it, I simply have to
put the pieces together so they work well together. Its a children's
puzzle where you have to build a cube out of a little of funny 3D pieces.
You simply have to prune the problem appropriately so you don't have to
try all of the possible assembly combinations.
>
> If I'd hazard a timeline I'd place reasonably competent genetic
> engineering capability at two decades downstream.
I would probably agree, if you mean "comptency" implies understanding what
the pieces you are working with actually "do". However, I don't care what
they do so long as I can get them to work together effectively. The nice
thing about nature is that many "standard" problems have been solved
multiple times, if one approach doesn't work, just choose another.
Robert
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:26:13 MST