[extropy-chat] Victoria News: Singularitarians in the City of Gardens

[Hughes, James J.]

http://www.vicnews.com/portals-code/list.cgi?paper=36&cat=43&id=583423&m
ore=

Singularitarians in the City of Gardens

Don Denton/Victoria News

By Keith Norbury

Victoria News

Feb 03 2006

He once dreamed he was a computer, David Coombes candidly admits.

"Basically, I was in a matrix," Coombes says as he sits on the sofa in
the living room of his character home in James Bay. "There were many
lights flashing on and off. It was pretty boring because all it was was
lights going on in foolish patterns."

That dream occurred around 1970 when he was in his mid 20s, a decade
before the personal computer became popular. In those days, a computer
was regarded as a bulky box of flashing lights and whirring magnetic
tape.

He still dreams of being a computer - only the dream is anything but
boring. And the computer he envisages bears far less resemblance to the
desktop machine of today than a modern computer resembles the one in his
dream 35 years ago.

Coombes is the vice-president of the Canadian branch of the Singularity
Institute for Artificial Intelligence, which is devoted to promoting the
creation of friendly, superhuman intelligence for the benefit of
humanity. His home, built six years before the Wright Brothers' historic
flight at Kittyhawk, is the institute's Canadian headquarters. Its
president is Coombes' house guest, 41-year-old divorcee Michael Roy
Ames, who is currently between jobs in his career as software developer
and database administrator.

At the entrance to the house, a brass plaque proclaims "Toad Hall."
Inside, the 19th-century woodwork and the wall decorations of spears and
woven hats from southeast Asian tribesmen juxtapose with the digital
future the two have in mind. In the kitchen, Coombes' wife, Zheny, an
Igorot from Bontoc on the island of Lauzon in the Philippines, is making
haggis.

"She likes it," her husband says. An immigration consultant, he is
moving his office out of the house, which adds to the chaotic
atmosphere.

So far, their group has only five members, although the parent
U.S.-based society boasts about 1,000. Similar groups, such as the
Acceleration Studies Foundation, also exist. The movement is small, but
growing - just what Singularitarians would expect.

Ultimately, they anticipate technology will rapidly advance over the
next few decades to where superhuman computers, or other entities as yet
undreamed, merge with human brains to produce a paradigm shift unlike
life has ever encountered.

"There is suffering - death, pollution - of many kinds. People try to
solve these problems, but it's very difficult. They need a lot of
organization and intelligence to grapple with them," Ames says in the
slight accent that remains from his upbringing in England. "One of the
things that has helped our society solve some of these problems and in
other societies in the past is our technology. We use computers already
to enhance our intelligence."

At first blush, Singularitarianism may seem like a religion and has even
earned from critics the label "Rapture of the Nerds." But
Singularitarianism differs from religion in one crucial aspect: its
belief doesn't depend on a supernatural power influencing human affairs.
Singularitarians rely only on the evidence and patterns scientists have
tested and observed in the natural universe.

Neither Ames nor Coombes considers himself religious. Zheny Coombes
quips that she's the only person from the Philippines in the city who
doesn't attend church.

"I think there very well may be an intelligence associated with the
universe like a higher power, but I don't think that is a religious
thing," Ames says.

He points out that a set of well-defined constants enables the universe
to exist as it does. Even slight variations and the universe, and life
as we know it, wouldn't exist because atoms and molecules wouldn't have
formed.

"I believe there could be a higher power when we make it or when we
become it," Coombes says.

Their movement doesn't have a Bible or a Koran. However, last year
computer scientist Ray Kurzweil published a bestseller, The Singularity
is Near, that detailed his arguments on the matter.

Kurzweil's primary thrust is that technology, particularly computing, is
advancing exponentially. He builds upon the premise of Moore's Law (the
40-year-old observation by Intel founder Gordon Moore that computer
processing power doubles every 18 months) and finds a similar pattern in
all of technology. Indeed Kurzweil goes beyond that - graphing
exponential increases in biological evolution and even the evolution of
the universe from the Big Bang, through the appearance of the first
hydrogen atoms, to complex molecules to the formation of life itself.

According to Singularitarians, these exponential increases in
technological innovation are about to explode in complexity so profound
as to boggle our human brains. The thesis sounds impossible until one
looks at the relatively simple math and graphs in Kurzweil's book and
realizes he is onto something.

Since the creation of the first computer in the 1940s, the processing
power has doubled about 32 times. Kurzweil estimates it will take only
about five more doublings for a supercomputer to "emulate the human
brain." A decade beyond that, today's equivalent of a desktop computer
will have that capacity. By 2080, Kurzweil calculates, $1,000 worth of
computing will be able to contemplate the equivalent of all human
knowledge since the dawn of the species in a fraction of a second. And
that still won't be the end of it.

Long before that - by his projection the year 2045 - the Singularity
will have occurred.

Not only do Singularitarians expect this to happen, they expect to live
to see it, their lives prolonged by genetically engineered organs and
pathogen-destroying nanobots.

By its definition, the Singularity is beyond our comprehension, but
Kurzweil envisions a world where reality and virtual reality become
indistinguishable. Ultimately this consciousness spreads across the
cosmos at nearly the speed of light - or even faster.

The runaway changes leading to that aren't yet noticeable because in its
initial stages an exponential graph appears quite flat. Only at the
"knee of the curve" does it begin to veer sharply upward. We're at the
knee now, Kurzweil says.

Ames doubts any force can stop the Singularity - barring a planetary
cataclysm such as a global thermo-nuclear war or an asteroid collision.

"It's really hard to imagine an obstacle that would throw the
Singularity off the rails," Ames says.

When it will occur remains a subject of debate among Singularitarians.
Vernon Vinge, the mathematician credited with popularizing the idea in a
1993 paper, predicts the Singularity as early as 2030. Others push the
date to early in the next century.

The expression "singularity" derives from a mathematical term to
describe certain types of functions, such as where the answer approaches
infinity as one of the variables approaches zero.

Physics borrowed the term to describe the collapse of matter in a black
hole to a seemingly singular point. The technological singularity refers
not so much to the black hole as to the event horizon surrounding it
beyond which nothing, not even light, can escape. In that sense, the
Singularity represents an event horizon of ideas impenetrable to the
modest power of a biological human brain.

"We both anticipate it and we want it to happen," Ames says of he and
Coombes. "And we want it to be accessible to everyone and not to just a
few countries and a few people."

Even if Kurzweil underestimates human brainpower by a factor of a
million or billion, that will only postpone the Singularity by a few
decades - as long as the increases remain on the same exponential track.
He agrees that any physical system eventually imposes a limit on growth.
However, he notes that integrated circuits aren't expected to reach
their limit until about 2020, by which time he anticipates carbon
nanotubes will have replaced them as the preferred circuitry and will
continue the growth. Beyond that the future is hazier, but he reports
that physicists have already demonstrated the capacity of a single atom
to store 50 bits of information.

Indeed the most vocal arguments against Singularitarians aren't that
this technological explosion won't occur soon - but that it will and it
must be stopped.

Among those sounding the alarm is another computer scientist, Bill Joy,
the founder of Sun Microsystems. In a famous 2000 Wired magazine essay
titled "Why the Future Doesn't Need Us," Joy underscored the perils of
runaway biotechnology, nanotechnology and potentially hostile superhuman
artificial intelligence. Anyone familiar with the movie, The Matrix,
understands the latter danger. Biotechnology harbours the spectre of a
genetically engineered plague. Runaway nanotechnology threatens to
convert all of the Earth's biomass, including human life, into grey goo.
Kurzweil acknowledges these threats but disagrees with Joy on how much
technological innovation humanity will have to relinquish in order to
prevent them.

Too many obstacles will postpone technologies to relieve human suffering
and prolong human life, Kurzweil argues. He is confident careful
protocols and other innovations, such as reverse engineering the human
brain in order to apply its pattern-recognition powers to artificial
intelligence, will safeguard against the dangers. That is also the
philosophy of the Singularity Institute.

"The point is currently we have control and how much control we have in
the future does not just depend on fate. It depends on what we do now,"
Ames says. "The Singularity Institute favours getting in the driver's
seat and driving it, so we can see these problems coming up."

Ames first became interested in the idea during his involvement with
transhumanist mailing lists in the late 1990s. He introduced the idea to
Coombes, who he has known for about 20 years going back to when they
both worked for the provincial government where Coombes was a fishery
scientist.

"I am doing this to solve the world's problems. This is an important
thing to do," says Coombes, who has a master's degree in
bio-geochemistry from the University of Victoria.

"I think it's pretty obvious something is going to happen," Coombes adds
later. "I read a lot of science fiction and technical science magazines
like Discover and Scientific American. I just see it's pretty inevitable
that it's going to happen. And if we don't get organized to reason out
how it comes about, we could be in deep trouble."

Even if we employ superhuman intelligence to defuse the threats of
biotechnology and nanotechnology, that still leaves the intelligence.

"Artificial intelligence will think in different ways than humans," Ames
says. "That is one thing that is hard for people, even educated people,
to grasp. I think they won't have human goals and desires unless we put
them in there."

But what if these super intelligences don't want to have any part of us?

In technical terms, this problem is called goal stability during
recursive improvement. An example of that would be Isaac Asimov's three
laws of robotics as enunciated in his science-fiction books and the film
I, Robot.

"Right now, we don't have the answer," Ames says. "We don't know how to
create that goal stability when the AI copies itself again and again. We
don't have a technical theory of how we can create that stability. This
is one thing we must know how to do in order for whoever is on this side
of the Singularity to connect to whoever is on the other side."

Even if the goals cannot be programmed, the most likely outcome is the
superior intelligence will simply be as indifferent to humans as we are
to insects, Ames says.

On a similar note, Kurzweil expects a super intelligence would devote
such a tiny fraction of its resources to human needs that it would
hardly even notice we exist.

Now what if humans didn't wish to be part of that brave new
super-intelligent reality?

"There'll definitely be a wide range of views and I think the
Singularity Institute is about increasing the number of options," Ames
says. "We are not just going to create artificial intelligence where
everybody just gets uploaded and lives forever."

Among those who have doubts about life in the Singularity is Coombes'
wife, Zheny.

She works as a registered nurse, often with patients near the end of
their lives, and she has difficulty with the Singularity's promise to
extend the human lifespan indefinitely.

"It's kind of a cool idea to be alive forever, but I don't know if I
would like to live forever," she says. "It could be kind of boring,
especially if I lived with people who don't like me very much."

Of course, she could cook all the haggis she likes and even occupy a
virtual reality identical to her present 19th-century home. In the
Singularity anything a human mind can imagine should be possible.

vicnews at vinewsgroup.com