[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Mathematician-to-silicon compilers and the Law

>> >publishing a paper on the algorithm also transmits that knowledge making

1. In the *best-case*, the published paper transmits some knowledge to the
reader :-)

2. In ECAD, tools try to go from that published Paper to circuits.  
In SoftE, tools try to go from english specs to runnable prototypes. 

Meanwhile, the humans who do various steps of this translation
get paid reasonably well.

2. But eventually normal english prose will be *mechanically translatable*
to a machine-ready format.  Which means that free speech law starts 
mixing with machine law (patents, export restrictions, posession

Because the Instructions become the Thing.

You can sell instructions on converting your rifle to full auto, or
stealing cable TV.  They're protected speech.   But suppose you could feed
a properly formatted blueprint into a universal fabricator and get a full
auto rifle out?

Crypto source code fed to a universal computer gives you a regulated item;
what  happens to the law when regular english is just a few clicks away
from being an executable?  A chip?

There is a fundamental issue that the law doesn't handle at all consistently,
because it previously didn't matter much, but which every CS Ugrad knows: 

There is no difference between software, and hardware. Only such annoyances
as cost, development time, throughput, flexibility, etc. 

(Not everyone can see machines as instantiations of ideas.  Politicians
have particular problems linking abstractions to atoms.  Fortunately, those
can see machines as ideas can turn their ideas into machines.)

I see two extrapolated futures.  In one, the end result is restrictions on
free speech in certain areas (e.g., crypto), for the benefit of national
insecurity and the Children.  In the other, the Law will have to allow you
whatever posessions you like, and only punish *actions* after the fact.
Given recent historical trends...


Imagine biologists couldn't publish new enkephalin sequences because
someone might put them into a programmable yeast and make unlicensed