Re: NANO: Hacking assembler security

From: Charlie Stross (charlie@antipope.org)
Date: Mon Feb 14 2000 - 04:23:43 MST


On Sat, Feb 12, 2000 at 08:36:58AM -0500, Delvieron@aol.com wrote:
>
> Trouble is, how do you make a system that can cope appropriately with
> a group like, say, Aum Shinryko, membership in the tens of thousands, who
> want to immanetize the eschaton (by building grey goo), without denying,
> say, smaller groups like, say, Alcor, membership somewhat smaller,
> who want to build their ressurect-o-mat(TM) (for all the corpsicles in
> their freezer)? >>
>
> Instead of a majority vote, how about a majority veto? A proposed design is
> sent out to everyone, and is approved after a length of time unless a
> majority of people vote against it.

I like it but I don't think it'll work.

Points in favour: Torvald's law applies ("given an infinite number of
eyeballs, all bugs are shallow") to spotting hazards.

Points against:

* It's vulnerable to special interest groups trashing someone's project
  for irrational reasons. (Consider the current mess over GM crops in the
  EU for an instance of this sort of thing.)

* It's vulnerable to an effective denial of service attack. If you want
  to smuggle a bad design through, simply hit the approvals mechanism with
  about ten million spurious design requests, all generated by making
  variations off some core structure. This will overload the people
  monitoring the designs, leading to designs being passed by default without
  enough people checking them. (You could work around this by making
  submitting a design a chargable expense, like filing for a patent, but
  then you end up saddling the nascent nanotech design industry with
  a very artificial cost structure and force legitimate inventors to
  subsidize bull-goose loonies.)

* It's inimical to the currently understood concept of a patent -- in
  effect, it gives your business competitors an opportunity to veto your
  best inventions, rather than protecting them. I'm no fan of the current
  patent system, but I think this is going rather too far in the opposite
  direction!

I think with some careful thinking you can probably find ways to tackle
these objections: some ultimate version of a "distributed veto" may well
prove workable. But I think designing these safeguards safely is going
to be rather a tricky problem ...

-- Charlie



This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:03:44 MDT