Acho que existe outro motivo também. Muitas pessoas sao incapazes de compreender como a ordem e a complexidade que se vê na
natureza puderam surgir espontaneamente. Estamos muito acostumados a associar mecanismos complexos com a açao de um
agente inteligente. Além disso, o resultado da processo evolutivo, por ser algo adaptado ao ambiente, acaba sugerindo a um observador
descuidado uma intencionalidade. Isso porque, quando vemos um organismo que age de forma eficiente e precisa, nao podemos ver
todos os seus ancestrais que
nao eram tao eficiente e precisos, nem conseguimos notar os "defeitos" do organismo atual (como
a localizaçao incômoda da próstata no ser humano). Se pensarmos na seleçao natural independentemente da biologia, vemos que ela
é um mecanismo realmente contra-intuitivo e, em certos casos, até meio assustador. Por exemplo:
Newsgroups: rec.arts.sf.science
From: "Thomas Womack"
"Forrest" wrote
> Here's a query that only intelligences can help me with, since it's
> currently too vague in my mind to distill into a net-searchable
> formulation.
>
> Somewhere I read the (true) story of a...researcher, maybe an
> engineer...who in order to solve a problem...of a nature I cannot
> recall...set up a neural network...of some variety...that eventually
> solved the problem, but in such a fashion that it was difficult to
> understand how it worked.
The problem was a toy one: discriminate between 1kHz and 10kHz signals.
The mechanism used to solve it was a genetic algorithm: prepare many strings
of random noise, use them to program an FPGA [1], breed from the strings
that got closest to solving the problem.
The solution produced turned out to be very much smaller than a human
engineer would have produced; you could prune it down to what looked like
the essential parts, but it then stopped working. It turned out that one or
two components which weren't connected to the main part of the circuit
nonetheless were critical in making the solution work, by setting up various
electric fields in the silicon.
It's a classic anecdote in the genetic-algorithm field: searching for
'genetic algorithm FPGA' might work.
Does that help?
[1] An FPGA is essentially a computer chip of programmable design; you lay
down components, like the ones found in microprocessors, on a regular grid,
and provide a programmable network of switches to decide which components to
interconnect and in what configuration.
Tom