Newsgroups: comp.ai.alife
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!gatech!swrinde!emory!mind.org!runic!thantos
From: thantos@runic.mind.org (Alexander Williams)
Subject: Re: ALife Park
Organization: Runic Writings UUCP Link: Convoco Hasturam
Message-ID: <D8IyHB.4pA@runic.mind.org>
References: <3o61j0$cbp@ns.cityscape.co.uk> <3oe09l$dfd@gap.cco.caltech.edu> <D8DMxx.11r@runic.mind.org> <3otcdi$abc@gap.cco.caltech.edu>
Date: Sat, 13 May 1995 16:15:06 GMT
Lines: 122

In an arcane scroll, Alexander Williams quotes the holy scripturist
C. Titus Brown, replying to the mystic words as written, saying:

>>   Its one thing to be alert to the possibilities, but completely
>>another to worry about apple trees uprooting themselves to march to
>>your house and lurk around outside your door, waiting for you to
>>leave so they can drop apples on your head.
>
>Hey, I thought EVERYBODY worried about that... ;)

   Only when its Birnham Wood, /then/ I worry.  I worry a /lot/.

>I have more faith in the physics of the universe than I do in the physics
>hand coded by Tom Ray (or myself, or Charles, or Chris...)

   I think I have more faith in the "physics" of the Tierra
simulation code than in RL physics.  The latter tends to be fractal
in complexity, confusiong, and contradictory at different scales.
Tierra, on the other hand, is far more limited in physics, the ASM
language is well-defined and will be well-defined in the net.Tierra
as well, and can not only be comprehended by a single human in a few
hours to a few months but an entirely different set can be conjured
as quickly.
   If the instruction set and architecture /themselves/ evolved,
then you'd hear my hue and cry far and wide.  They don't, only
things constructed from the basic blocks therein are changing.
Thus, I feel highly secure in saying there's no question of danger
effects outside the Tierran architecture.

>Well, I'd be interested in hearing why; Tom himself admits that it's not
>a scientific move, but rather that it's just something he's interested in.

   Mainly because I persist in thinking "scientific moves" are
anything that resembles the scientific method and gives back
information that's half-way useful.  I know, I know, heretically
neophillic, but I was born that way.

>If you pretend to know what tierra will do in a given situation, then
>(to steal from your later text ;) you need to be taken out back and shot...
>Tom & the other people who've studied the thing feel comfortable generally
>characterizing it, but none of us say that we know what it's going to
>do every time.

   Not only can I pretend to know, I can know exactly.  The Tierran
OS and instruction set are computable, you know.  If you give me any
Tierran organism I can step through it and tell you what it does.
If you give me an instruction set that doesn't include a way to
generally access telnet, I can assure you it won't be logging into a
MUD, in the same way that if you don't give it an instruction that
accesses the HD (or a generalized interrupt), it won't be able to
scribble the works of Shakespeare on my home directory.
   You don't need to be able to completely characterize a system's
every behaviour to know what its /not/ capable of, given full
knowledge of its basic physics.  That full knowledge, we possess.

>You seem a bit overconfident about the knowledge you and others possess of
>tierra.

   If t'weren't for overconfidence, we'd never get ANYTHING learned.

>Fine; I actually agree with you.  However, a virus COULD be designed
>specifically for, say, "pine" running on a vt100 terminal.  It would get
>something like 30% of all the college students on the net...

   Well, I don't know about Pine on a vt100, there's nothing surely
that you can /access/ through that vector.  Now, if you futzed
around with the MIME encapsulation and convinced it to execute some
code, then, maybe, just maybe, it'd get some useful access.  I don't
even know if /that's/ possible, given my limited knowledge of MIME.
   Mind you, we've now dropped it from being a true virus to just
being a camoflaged Trojan horse...  And a highly, highly limited one
at /that/.  Single platform, single OS, single mail-reader, single
/style/ of encapsulation.  Given those limitations I'll freely admit
that something of the sort could be conjured.  I'd also say it'd hit
maybe 5% of the studentry on the Net, possibly much less.

>Telling people that it's impossible isn't ever a good idea.  Highly
>improbable that ANY such virus exists?  True.

   Its /impossible/ to create a virus that does what the GT virus
purports to do.  I feel confident in the extreme that my statement's
300% correct.

>Well, at least it's self-deprecating as well :).  (And do you have a
>better explanation for the missing socks?  I didn't THINK so...)

   It all started with the sexual voraciousness of my first
girlfriend.  Developing a fetish for feet, she...  [3meg cut out to
save SANity.]

>Hmm.  Let's see... I 'member me a little Internet worm that took advantage
>of holes in finger and sendmail.  Well, that hardly did any damage, did it?
>[ note: that's SARCASM ]  Sure it did.  And yes, "experts" knew about it.
>But they didn't fix it.  Slightly different scenario, but you rule that out
>too, so why couldn't you be wrong about tierra?  (Which, I might add, has
>been around MUCH less time than either finger or sendmail, has been beaten
>on MUCH less than either of them, and is MUCH more complicated.)

   Tierra is /much/ more complex than finger I'll buy, saying its
/much/ more complex than sendmail suggests to me you've never tried
setting up a leaf UUCP node with demand output of SLIP-based email
and news but secondary paths out two seperate UUCP links down which
differing newsgroups come by both news and email transport...  I'd
rather debug a random Tierran from the 3x10e24th generation on a
CM-5 than do that sendmail configuration again.
   I can say with surety that there exists no instruction in Tierra
to do X.  I am not /even/ so sure about the internals of sendmail,
even today.

>Look, I'm not saying ANY of this is highly probable.  IGNORING the possibility
>is bad, however, and that's what you're encouraging people to do.

   Think of me as alpha-beta pruning on the game-tree of Life.  Some
things are so unlikely (an in fact, impossible) that fretting over
them is contra-efficent.  Thus, worrying about Tierrans outside the
Tierran architecture gets pruned.  Worrying about "turning the
rainforest to plantations" in the Tierran architecture is valid.
-- 
thantos@runic.mind.org (Alexander Williams)     | PGP 2.6 key avail
  Should we shed our mental pants and compare   | DF 22 16 CE CA 7F
  the size of our consciousnesses?              | 98 47 13 EE 8E EC
      Jan Sand to Marvin Minsky                 | 9C 2D 9B 9B
