Re: Nano question?

From: Michael M. Butler (mmb@spies.com)
Date: Mon Feb 24 2003 - 09:49:12 MST

  • Next message: Michael M. Butler: "Re: The Pentagon's New Map"

    Nathanael Allison wrote:
    >
    > It seems highly
    > improbable that a nano assebler can be created and be functional in a
    > non-isolated enviroment. Why? because evolution would have already
    > created them a thousand times over.

    Yep.

    > It would be much more likely to incorporate a nano assembler in some
    > type of existing organism.

    If you are speaking of a general purpose mint: *Why* do this? The risks of
    runaway go up vastly. Why not just create as many (small) isolated environments
    ("sandboxes") as are deemed prudent, put the GP mints there and keep them starved
    for some rare feedstock?

    Even if you are talking about specialized assemblers, it's probably smart to include
    multiple limitations into both the assembler and the biological source code.
    And make the breeding of assemblers a throttleable process; consider carefully
    whether most assemblers ought to be able to reproduce themselves at all.

    > However this would be very difficult for the
    > nano asseblers to do anything too complex as the organism would
    > determining how they assembled together. It might require many stages of
    > many slightly altered micro organism to get one simple effect.

    True. So there's a tradeoff. Some stuff doesn't get efficiently made by bug chains.

    > When they do get these nano asseblers up and running in labs it will be
    > interesting to see their evolution occuring even at that level. Maybe
    > after the first one replicates 10 to the 100th power they will evolve
    > enough. But will they still be funtional? If we select the functional
    > ones to survive will they evolve enough?

    Evolution by mutation is risky. It appears that's what you're talking about here.
    I don't want a general purpose assembler mutating but still functioning, and I'm
    pretty cautious-minded about green goo in general.

    The "standard model" of molecular nanotech that some folks like to dismiss has it
    that AIs will help to design better and better assemblers, along with better and
    better AIs. We might get SkyNet or COLOSSUS, or we might get Friendly AI. Or the
    bus might be late while we stand in the cold fog by the streetlamp.

    I expect back-pressure from society, and a few surprises. Unfortunately, the
    surprises I anticipate are sometimes contradictory. Oh well.



    This archive was generated by hypermail 2.1.5 : Mon Feb 24 2003 - 09:51:36 MST