From: Carey Lening (carey@tstonramp.com)
Date: Mon Apr 03 2000 - 21:06:38 PDT
I have to admit, it was a bitch and a half to get into the conference. Once it hit Slashdot, the factor really did set in. There was well near 600 people in the panel room, with a overflow room of 100 or more sitting down the hall. Ryan and I barely eeked out a seat, but through some clever finangling on his part, we were able to see the discussion up close. It was -well- worth it. I was glad we made the drive (all 7 hours) and slept in the car to go.
None the less.. for those individuals ( I know a few were actually interested in the discussion and wanted to attend but couldn't for some reason or another [RK]) I took fairly detailed notes and have decided to FoRK it on over. Douglas Hofstadter was the moderating host and the 8 panelists were as follows (I'm clipping from Rohit's original posting, as there wasn't much introductory descriptions from Hofstadter, and I couldn't procure a flier...
...
Primary speakers:
Ray Kurzweil (inventor of reading machine for the blind, electronic
keyboards, etc., and author of "The Age of Spiritual Machines")
Hans Moravec (founder of Carnegie-Mellon University's Robotics Institute,
and author of "Robot: Mere Machine to Transcendent Mind")
Bill Joy (co-founder of, and chief scientist at, SUN Microsystems)
John Holland (inventor of genetic algorithms, and artificial-life
pioneer; professor of computer science and psychology at the U. of Michigan)
Panel members:
Ralph Merkle (well-known computer scientist and one of today's top figures
in the explosive field of nanotechnology)
Kevin Kelly (editor at "Wired" magazine and author of "Out of Control",
a study of bio-technological hybrids)
Frank Drake (distinguished radio-astronomer and head of the SETI
Institute -- Search for Extraterrestrial Intelligence)
John Koza (inventor of genetic programming, a rapidly expanding branch of
artificial intelligence)
Forthcoming are the notes... Enjoy
Ray Kurzweil: Reverse engineering of the Human Brain.
Neural nets: Paradigm shifts every few years.
21st century -- Within that century developmental time will have elapsedin what would have time wise been 20,000 years.
- Exponential growth -- makes the future exciting
- Haven't encountered machines that come close to the human brain
* Progression of Moore's Law: How long is this trend going to go on?
Machines on an exponential growth
-- Moores law is 5th paradigm.
S Curve
Change in paradigm occurs right before old one dissapears
--6th paradigm -- 3rd demension: Nanotubes
$1k worth of circuits == human brain computation, and it will increase exponentially.
Reverse engineering of brain == BIG
Kursweil estimates that by 2029, there will be nanobots.
Downloading vs. Understanding what is in the brain.
* OUTCOME: AUGMENT HUMAN INTELLIGENCE * (Cyborg?, he didn't call it this, but I think that s what he he was hinting at the entire time. He went back to this often in the discussion of the other panelists as well as the audience discussion -BB)
He also envisoned virtual reality that involved all the senses, and found it inconceivable to separate the good tech from the negative technology.
The only way Kursweil found it was possible to inforce good technology waas through totailitarian means of control.
Bill Joy: The relinqueshment of technology, or 'slow down and think'
-genetic nanotechnological, Bio: == Great wealth, great ability --> POwer
-Joy: "Be unsentimental about technology."
Problem areas of technology:
1. Weapons of mass destruction.
2. Knowledge can be a weapon: Knowledge created weapons of mass destruction.
The real risk: Democratising abilities for individuals to cause harm.
-Do we really want to stop knowledge then? (me)
Genetic Engineering
Designer pathogens by 2028.
(For any who haven't read, Joy recomended Robert Wright's Stolen Continents)
- Info is public
-Info is cheap
-Computers are good
-Nanotech is much the same.
WHAT DO WACKOS do?
-Pestilences
Robotics:
Wild successor species
-Joy estimates that robotics are less than 30 years away -- and wil lbe the change in paradigm.
Furthermore that 200 delusional techies are all that are needed to wreak some major havoc on the world.
Whats the fix??
== Relinqueshment of technology: Don't develop things that are too dangerous.
We must anticipate or at least consider the possibilities of the realities of extremists.
* The consequences of our truth Security
What are the technical and ethical rammifications?
Hans Morveck: Robotics
-Humans have an innate fear of change.
We have immune systems to deal with these bad things ( me: I believe Morveck was actually talking about immune systems to deal with bad change or bad technology, but he was kind of fuzzy interms of what he was actually talking about. )
-What we need are well funded researchers to fund the robotics race (me: I thought that most likely would be a necessary lead in to corporate / commercial interests. )
Morveck noted that it wasn't always being done by military any longer.
"Automate, Emigrate, evaporate"
Robotics path:
-Automation beyond information -- tactit physical tasks to be done by robots instead of humans. (EArly Example: The Stanford Cart)
3d Mapping will show up on industrial vehicles.
1000 Mps to do small Robotic tasks: Which he related is essentially the brainpower of a guppie.
John Holland: Need a significant theory on tech development.
- Concious machines
-Not on a 30-100 Year time scale
CNS = HIGHLY controlled
Humans have very narrow snapshots of life. --> Sacades
Don't know how to make sacades work '
-Humans articulate by metaphors and analogies, not by boolean Logic like machines.
- More machines === the less solving of the "big" problems.
-Software has not followed Moore's law.
=Evolutionary Technique
* where do I start? What do I start with?
* Not easily defined
-pruning vs. Stopping
-How can you predict? Even on theory?
-Too much reliance on theory.
Kevin Kelly: Literally answered the question Will spiritual robots Replace Humanity by 2100?
-Social Observer
-Society has taken 400 Generative (Mother/daughter) cycles to get to the point that it has.
- IT will take 4 generations to morph into something else. (INCREDIBLE time increase)
+ Humanity, what is humanity? What are we good for? IF anything at all.
The central question is : people will look back in 2100 and have noted a 100 year identity crisis. We don't know who we are.
* Replacement of the species is VERY rare, rather, what will happen is that we will meld into what ever we create.
-Works the same with technology in general.
* robots are our mind children.
-We like regular parents have to let them go.
-The Less we vary, the less revolutionary we are.
-We as humans are constantly taught to be good citizens.
--- In the spiritual realm: Why wait, we have AI?
- Making new kinds of intelligence through AI.
Kelly: "I think technology has its own agenda."
-The question is, what does technology want?
* wants to get smaller (Moore's Law)
* Wants to get faster (Ray's Law)
* Wants to do what humans do (Kelly's Law)
Humans will do whatever we want.
His answer to the above question, will spiritual machines replace humanity? No WAY!
Frank Drake: SETI -- Relationship to ETI.
-He feels the relationship is far more advanced.
-Guidance: Who will we be?
-Related to SETI:
-- Idea: Very shortly computers will be surpassing humans : gives us guidance to what to search for.
Guidance == Guidance for design of equipment.
Demensions of Phase space
* Position in Sky
* Radio Frequency
* Bandwith
* Polarization
* Sensitivity
* Duty Cycle
What can we do/ have we done?
-Telescopes cost the same as the weight of telescope and hamburger (SETI theory)
Ralph Merkle: Nanotech
- Joy has one of the worst strategies according to Merkle.
Reasons:
1) We cannot have informed discussion without Active research
2) We can only stop research if we know about it -- public research in democracies
a) Give less scrupulous nations economic advantage, so they cannot develop the same tech.
3) Turning our back on technology that could save lives. While we in the western world have many ways to save lives, other countries may not.
* self Replication
What are the dangers?
-Our only model is with other living systems.
Machines are NOT living systems.
Merkle's Idea of Broadcast Architecture:
-Living systems replicate by DNA
-Broadcast Architechture: Blueprints are in a central machine and are broadcast out to replicate.
-Smaller
-Simpler
-non-Biological
-Replication is cut off if you remove the central system.
-What happens if the BA enters the Machine?
-Deliberate Abuse is always an issue.
-Why does nanotech have to be principally viewed as a weapon (He was focusing endlessly on the function of nanotech as a weapon, weither it would be offensive or Defensive (me)
-Theoretical + computational mechanisms of artificial Self - replication.
John Koza: computers + Biology
- Computers can do human things? Yes?
- Like Biology? No.
- When a computer appears to do things that are human like
Human brain operates at 10x 12 (Ten to the twelfth)
10x 12 = 1 brain second.
Panel Discussion:
-MAchines dont' need to do everything a human does.. in other words, we don't need to make them do all functions of a lizard if we just want the visual accuity, or breathe if we just want it to swim well, etc.
-Interaction between mps + Genetic Algorithms.
Kurzweil: Notes that the bad form of tech is already apparent: computer viruses.
Joy: "OS size follows Moore's law, function is fixed."
-Q: How did evolution miss producing the "Uber Machine"
The panel also reccomended reading from Dr. Dobbs.
I made note that the audience ( who seemed to be comprised mostly of Theology Majors) had more of a clue about the topic at hand than most of the panelists.
--------
Well I think thats it ... All in all, it was a GREAT lecture, and I really got a lot out of it. Enjoy.
--BB
This archive was generated by hypermail 2b29 : Mon Apr 03 2000 - 21:08:43 PDT