The Singularity

> Since the board has been relatively quiet the past week or so I thought
> I would take the opportunity to solicit opinions on the "singularity",
> as applied to the accelerating rate of technology development.
>
> http://www.kurzweilai.net/meme/frame.html?m=1
>
> If something like the "singularity" were to occur, what was be in
> potential impact on the future development of orbiting habitats?

Kurzweil is a clever and creative guy, and says much of interest --
including the notion that change is happening in such a way as the next
hundred years may seem like thousands of years of change compared to the
last hundred years. While I agree with much of the thinking behind the
notion of a singularity (exponential growth in computing value per dollar,
emergence of AIs changing the landscape of the economy and society),
Kurzweil specifically makes a classic error of thinking human intelligence
is beyond evolution. To elaborate, he bases a lot of his optimism on the
notion of human brains being downloadable into some futuristic network,
then proceeds to ignore the evolutionary consequences. For example, he
ignores the possibility that parasitic AI processes that will overwrite
sluggish human intelligences not well adapted for network living.
(Remember, more than 90% of species in the ecology are parasitic.) Or
another way to look at it is, why should the network tolerate huge clunky
programs wandering around just because they have some ancient link to
(likely long since dead) meat-space brains? I'm not saying there aren't
ways to approach those issues, just that Kurzweil seems to ignore them (at
least as of a few years ago), and I think does so to his own peril if the
rest of what he says is true.

One issue of the singularity I disagree with is that I personally think
large corporations or large bureaucracies function as a form of AI (though
made up of human parts, with people often acting mainly like amoral
transistors in that context). So, I think what we have seen in the world
over the last fifty years is already an example of AIs at work. This may
be true even for a few hundred years before that with the various East
India type exploitation companies, and perhaps even a few thousand if you
see kingdoms as AIs -- ahistory of AI behavior including slavery and the
decimation of various native peoples including in the Americas. It's most
often not a pretty picture -- and what good comes from it is generally
despite the system.

NASA is a prime example -- it spent a hundred billion dollars for a space
station, more than O'Neill wanted for habitats, and really has not much to
show for it in terms of realizing positive value for humanity as opposed
to just perpetuating itself. Sure, some minor successes, but a hundred
billion dollars worth? Langdon Winner in his book "Autonomous Technology"
talks about "Reverse adaptation" where large bureaucracies change their
environment to create needs for their continued existence. There are those
few at NASA who do space settlement work anyway (which should IMHO be
NASA's main mission), but it is important to recognize that the dynamics
of the NASA AI as it were have to do with its own perpetuation and
survival -- not directly producing anything of value for humanity. I could
pick on other AI bureaucracies which are worse as far as human survival
(e.g. Enron, FEMA(?), Microsoft, the US Military), but NASA seemed more
topical for this list.

And to head off the stock reply of -- corporations are made of people so
they serve human ends -- I can only point to something like the Nazi
bureaucracy, which internally was very polite and kept accurate records,
yet did horrendous things, where the best one could do from inside was a
bit of foot dragging lest one be accused of being unpatriotic or worse and
be killed. In the case of any modern large public (US) corporation, any
executive who talks about doing anything except pursuing the most short
term profit maximization is going to be fired. Any employee who does not
substantially fit their role will also be, at best, fired. So, the system
functions with its own dynamics despite the individual desires or ethics
of the players. Changing such a system to direct it to more humane ends is
a very complex problem. I'm not saying all bureaucracies are actively
hostile to humanity -- just that they tend to be at best indifferent while
defining the landscape in which humans need to carve out a survival niche.

And yes, I'm sure a bunch of free market people will chime in on how
productive such amoral systems can be -- which is part of the chorus that
keeps that whole system going. To head that off, I would cite
John Maynard Keynes himself:
http://en.wikiquote.org/wiki/John_Maynard_Keynes
"For at least another hundred years we must pretend to ourselves and to
every one that fair is foul and foul is fair; for foul is useful and fair
is not. Avarice and usury and precaution must be our gods for a little
longer still." from: "The Future", Essays in Persuasion (1931) Ch. 5

Well, in Kurzweil's accelerate time frame, it has been much longer than a
hundred years in accelerated time units since 1931, so perhaps, if humans
are to survive, the time is now for new (and perhaps more ethical or
inclusive) ways of doing things. I hope the AIs of the future have more
humane ethics than these bureaucratic AIs of the past.

One other aspect of the singularity and technological change -- it would
not take much more than a self-replicating robot with the intelligence and
size of a cockroach to spell the end of humanity (e.g. how would you
manage to keep them out of your home and from killing you in your sleep by
drilling into your ear canal?) So, there is no guarantee making
self-replicating AIs will lead to space settlements. Further, developing
space settlements without first solving many problems back hear on Earth
will still leave the Earth as a place that generates violent technologies
which might overwhelm space settlements (even self-replicating ones).
There are a lot of potential paths, but the one of an elite in space with
a warlike and envious Earth with continuing resource conflicts I do not
think is a stable one. As Bucky Fuller wrote, whether it will be "Utopia
Or Oblivion" will be touch and go up to the end. And clearly better
computers (and even AIs) and advanced technology will play a big role in
making space settlements easier to do -- and once the reach the level of
only costing a billion dollars or so to start (like launching one
self-replicating seed) then they are at the level where some rich
individuals can do them as a hobby or a spiritual pursuit.

Still, do I think the potential negative consequences of the singularity
for humanity is yet another good reason for developing space settlements?
Yes I do. Though not to such a high degree that we should unduly reduce
trying to minimize its negative impacts by making the Earth a better place
at the same time (and I think the two are compatible, both helped by
realizing Bucky Fuller's "Design Science" ideals).

A bit rambling, but all I have time for. :-)

--Paul Fernhout