Re: Lyle's Law

Dan Clemmensen (
Wed, 09 Oct 1996 19:23:45 -0400

Damien Broderick wrote:
[SNIP of John vs. Lyle,]...
I agree with your main point. We need to spend less time yelling at
each other and more time trying to identify our actual points
of disagreement.

> *detailed architecture and practise* of certain canonical nano claims, but
> does so in a teasing and finally annoyingly flip manner. John (among
> others) seems to respond as if Lyle's requests for detailed flow-charts are
> redundant and even slightly blasphemous acts of lese majeste.
> This is a shame, because we need the detail even more than the grand
> speculative strokes.
True. Read NANOSYSTEMS, not just "Unbounding the Future", etc. I intend
this as a positive suggestion, not as the "generic canonical aswer to
all pleas for detail." In particular. I'll try hard not to refer to
NANOSYSTEMS in response to a detailed question unless I cite a specific
paragraph. In this case, you are making a broad requiest, so I'm giving
broad answer.

> To consider a recent exchange: I asked whether people in an early nano
> society might be obliged to accept draconian authoritarian controls to
> forestall self-obliterative use of an Unstoppable Salt Mill (you know, the
> fairytale device that explains why the sea is salt - it's down under the
> water even now, cranking out its ceaseless stream of salt). The range of
> responses stretched from (a) it can't happen, because nano assemblers can
> only do X and even that only with great difficulty,
Nope. If nanotech has any validity, then out-of-control naneotech is
a large incremental leap.
> (b) it can and probably
> will but that's okay because we (we?) will be off the planet by then (by
> 2015?!) [and why won't the folks in the habitat or on Mars also bugger
> things up with their handy assemblers?],
Good question. If "we" implies a community of humans, and it they
nanotech to leave earth, then unless they police the entire community,
run the risk of "wild" namotech.
> (c) it's cool, because we (we???)
> will have Transcended.
Quite possibly. Its a race between Trancendence (via the
and the gray goo.
> Do you see why Lyle (category a) might think he was
> dealing with the Santa Claus Fan club?
Nope. I don't understand how people on this list can reject the
premises just because the consequences are both strange and frightening.

> I wish he'd [Lyle] given detailed analyses of why modularised,
> bottom-up-bootstrapping nanomech systems are impracticable.

I think you are asking for a bit too much for this list. I'd be
satisfied wit a simple statement that he considers it impracticable,
a brief sketch of the reasons such as you give below. That would be
sufficient to permit us to understand our differences and examine them
further without rancor.

> I suspect he
> shares my hunch that orders of magnitude are being scamped when someone
> easily gestures toward replication via atom-by-atom scanning,
I've never seen atom-by-atom scanning mentioned seriously except in
science fiction. Do you have a reference?
>or the AI
> sophistication required to do the same job top-down.
This one needs a bit more excplication. We already do top-down design,
fromn the level of city planning down to the level of individual
within a microprocessor. Why is this qualitatively harder when designing
the smallest components from atoms? even if every single component must
be designed by a human, we still get "free" manufacturing. The design
effort is a very small part of the cost of a consumer item.