Will Spiritual Robots Replace
Humanity
by 2100?
A Symposium at Stanford
April 1, 2000
A report on the symposium organized and headed by Douglas Hofstadter. Participants included:
Ray Kurzweil, Hans Moravec, Bill Joy, John Holland, Ralph Merkle, Kevin Kelly, Frank Drake, John Koza.
( The original announcement page is online at : http://www.stanford.edu/dept/symbol/Hofstadter-event.html )
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
I woke at 5AM Saturday morning, left at 6:30 to drive 250 miles in about 4+ hours from east of Lake Tahoe to Palo Alto, found a seat in the Stanford Teaching Center auditorium, not leaving my seat from 11AM till 5:30PM, stayed till 6:30 and then sat driving home, arriving at 11PM after another 250 miles. That might seem like the premiere actions of an 'april fool', but it was anything but. I had spent the day in the company of 9 of the planet's most creative thinkers, with an overflow crowd of 700-800 in attendance. Ideas and attitudes clashed, meshed and impacted .. it was an intense 5 hours of conversation and commitments to ideas ... It was worth every minute, every word.
There was wit and humor from every one of the speakers, and I'll try to capture their personalities for you as well as their ideas.
Hofstadter started off by being very straight forward with everyone. He's a man on a quest. He organized the symposium (the second in an impromptu sequence of such gatherings) not for the edification of the audience, but for strictly personal selfish reasons: to try to find a grail, a 'Theory Encompasse' <as I would name it>, that would make sense of science's wondrous discoveries and human perceptual realities .. all at once and all together.
He said that he stacked the day's deck in who he asked to participate. He wanted people who deep down assumed that energetics and systems have a vitality which either harbors sentience or affects and may even mimic certain characteristics of sentience. It wasn't to be a forum debating 'could AI's be conscious?' but rather, how behaviors and designs are to be encountered, interacted with, valued. Are we dealing with artificiality or evolution? and what if our opinions eventually prove not quite perfect? How should we deal with potential "now" and then, 'later'? Who anticipates what? And why do they?
Unquestionably, every symposium member held an optimistic belief in the future. So that was their commonality, even as they saw different alternatives as desirable, and so challenged one another, either derisively or good naturedly. Tension and respect ... with the ultimate desire being pushing humanity into a healthy meaningful future.
~ ~ ~ ~ ~ ~
Hofstadter's real symposium question: "Who will be 'we' in '93?" (2093, 2493, you name the rhyming year).
I.e., Will what we use to modify and improve organic existence contain seed qualities and abilities which will transform life and humanity beyond what we anticipate? Will emergent qualities supercede 'organic' qualities? Is 'mind' and sentience robust enough to be transferrable from container to container? Could we be creating fabricated species which will replace us, co-exist with us, or be no competition what so ever?
Can we anticipate what the 'dominant life form' of year 10,000 (and beyond) will be?
~ ~ ~ ~ ~ ~
Two cartoons:
One fish to another: "Life, out of water? Are you crazy? How would life breathe out of water!?!?!" [off picture edge: a frog hopping onto land,"ribbit, ribbit"]
One scientist to another: "Life, in a silicon matrix? Thinking? Are you crazy? How can a device 'think'!?!?!" [off picture edge: a year 2150 automaton, "robot, robot"]
~ ~ ~ ~ ~ ~
After this 'invocation' to an enthusiastic audience, came:
Ray Kurzweil. Futurist/prognosticator par excellance'. author: "The Age of Spiritual Machines". Kurzweil is a smooth and sophisticated speaker. His take on AI is that its all an adventure, and siliconics will resolve life's discomforts, open outrageous and wonderful vistas of experiences and possibilities undreamt of in past human history, and just, well, make life a lot more fun and good for the psyche/soul. Virtual experiences will take coming lifeforms to new plateaus, unlimited realities. Nothing but improvements, on the horizon. The more outrageous the predictions, the better. We've always gotten caught short by conservative predictions, so let's go to extremes and assume the best. Technology isn't just salvation, it will be the essense of existence.
Bill Joy came next. Co-founder,Chief Scientist at Sun MicroSystems. Joy is a humanist first and foremost. He knows the 'soul' of siliconic technology, what it is capable of .. for good and for not so good. And he's not convinced that tech-nirvahna is any kind of Nirvahna at all. In fact, its the 'mindlessness' of siliconics that worries him the most. It may end up with vast information-processing capabilities, but, with current architectures (including neural nets) it has no value-base to determine, let alone guide or monitor un-governored behaviors.
We'll assume one set of outputs and referential impacts, and low and behold (as is typical even today) other unexpected results will happen. And if we start rampant manufacturing of self-reproductive and self-modifying pico sized pseudo-lifeforms, we could be swamped in the grey slime of nanobot-mud, doing all sorts of unwanted things. Joy recently wrote an article in WIRED magazine, stunning folks by quoting some insightful and wise ideas by the flipped-out Unibomber, Ted Kuszcinski. Business for business sake, at the expense of human values, is a future not to be so devoutly wished for. What's troubling is how to make the case for conservative science, when the arguments get blurred by 'change for changes sake' and 'its evolution, just a new form of it, that's all'.
It's interesting. Bill was sort of cast as the evangelizing nay-sayer - later in the day he admitted to carrying a bible for reference, and, in parallel, quoted Einstein's full letter decrying the disastrous possibilities of unleashed radioactivity, calling for its controlled limitation. But this tall lincolnesque personage came across to me as a most sensible scientist - cautious about implications he'd seen create havoc in other venues. Late in the afternoon Kurzweil mocked Joy in a fabricated press-release where Joy supposedly announced he was ceasing all software/hardware development because of possible detrimental affects on the world. Everyone laughed - including me - as Bill Joy sat in hurtful ridicule. But I mention the incident not to put down Bill Joy's opinions and stand, but because of what Ray did not half an hour later, just before the proceedings ended. He payed homage to Bill Joy by finally, after 4 and a half hours of gritty view-banter, saying that yes, complete unbridled damn the torpedoes and fullspeed ahead on absolutely every and all siliconic developments/research was probably a naive and sophomoric tack to take. Caution, somewhere somehow, is probably a wise act. I think he was earnest, he didn't have to make that concession. It just came a little late in the day.
John Holland spoke. Briefly and to the point. Holland is an Elder - along with Drake and Koza - in the field of 'other intelligences'. He's seen it all, and takes all of it with a grain of salt. He peppered in good natured soft spoken humor at both extreme views. You could tell that he loves knowledge and learning, pushes no agenda other than 'enlightenment', and is delighted to see conceptual progeny accomplishing so much in so many different specialties. In some ways though, AI seems to have reached an impasse, and he, like many, is waiting for the next paradigmatic breakthrough. Right now, discussing spirituality let alone reliable primal sentience, in siliconic systems is more than a bit pre-mature. He didn't say those words, but his demeanor did.
Hans Moravec, Carnegie Mellon's premiere roboticist took the podium. A very unpolished and methodical speaker at the podium, the rest of the day proved him to be anything but a 'reluctant thrown into the limelight'. Grabbing the dialogue-microphone at every chance during the later roundtable/questions session he came across as an excited wunderkind, totally gleeful in everything that robotics and computer intelligence "might" be capable of .. as if all the promises were just around the corner .. accomplishable maybe even tomorrow morning. You just have to be enthusiastic and believe. All the toys and gadgets will do everything we've ever imagined of them. Walk, talk, think and amuse. They'll wet nurse our kids, clean our houses, perform medical miracles and be our best buddies all through life. Teachers, physicians, friends and playmates. Soulmates from a different part of the periodic chart. la, la, la .. "life will be a breeze ... sweetheart!" Its only a matter of breaking through the MIPS barrier (millions of instructions per second). At 10 mips now, we have to get to 1000 or better to mimic organic capacities.
Kevin Kelly, editor at WIRED, author of "Out of Control". Another soft spoken person, but for my money, the only participant who voiced some truly 'out of the envelope', creative, challenging possibilitiess. I'm not a WIRED reader (yet. that is about to change) but Kelly is a cool breed of the information-tech-news age. He's a real 'possibilities' thinker, even more artist than scientist, but obviously both, and not afraid to dance with several concept partners at once. If he represents the psyche of WIRED, then the rest of us are in very good hands. I apologize that I can't recount specifics of what he said, but WIRED is an easy access, and you ought to read him directly.
Frank Drake is a leader out on the edge of perceptions. Self-effacing, he showed a cartoon of an E.T standing next to him at a blackboard, making a subtle correction to his notable 'Drake Equation' that predicts the likely number of sentient civilizations waiting for us to join them in the universe. (a factor of .2 is deleted, making the population 5 times more likely than predicted. :-} ) Drake stood to explain that SETI-esque technology is only a factor of ten to the third away from total-sky monitoring, in broad spectral regions. Sensitivity and power are being boosted by new array designs and adjustments for alternative signal forms. Also, recognition of alternative comm-methods .. like holographics .. are being studied/developed (at least in the theoretical stages) and will probably be part of the 21st and after century technology used in expanded search programs.
His work and that of nano-tech AI share a common qualities .. subtlety, efficient coding, and finesse. Intelligence at the large and non-local is companion to intelligence at the quantum and local/nonlocal. Drake devoted his life to humanity's searching for soulmates from a different part of the galaxy.
Ralph Merkle. nanotechnologist. Foresight Institute. The most unabashed techno-industrialist I've ever had the pleasure to see in action. "Its a tough world kids. You have to get in there first and fast. Front line research. Front line application. If you don't, someone with another world view (which you may not like) may get there ahead of you." ...that's paraphrasing, but you get the idea. "there's POWER in them thar techno-hills, so don't stop me, I'm doin' it all for YOU and your future, y'know!".
Things are happening in the nano-tech world that Merkle's incessant wry smile all day seemed to hint at, but there was no way he would ever divulge what they might be. He seemed to be here on a recruitment mission actually. His concern isn't sentience, just straight forward commercial development/application. The military-industrial complex is alive, well, and .. safe .. in his contributing presence. And, there in rests the tension between himself and Bill Joy (the two seated happenstantially shoulder to shoulder for most of the day). They tussled verbally several times with neither giving any ground or conceding any issues, though Joy got closest with a terse question, "Is nanotechnology mainly offensive or defensive?" which Merkle fudged about replying. The audience was in a minor frenzy at that point and local techno-sentiment weighed the balance in that direction. Of course, when Merkle described a consumer Utopia where every home would have a Star-trek like replicator, all material needs and desires looked after in some consumer driven Eden, he glossed over the fact that such a future would raze to ashes complex commercial economics ... but the audience seemed to buy into it anyway and the thirst for material ownership relegated Joy's views to relative nuisance status. Bill Joy got the last - very softly and nearly missed word in though: "you {addressing the whole room}don't realize where all the qualities you're talking about [sentience, behaviors, drives, motivations, so on] come from" or what can happen when these different kinds of organizations/systems interact. (Fortunately, a definite coterie of life-values hungry people were present, and Bill Joy was swarmed around in the good sized hallway outside the lecture hall for the better part of an hour after the formalities ceased inside.)
Anyway, last but not least, John Koza, the inventor of genetic programming, spoke during the initial presentation session. What drove into my awareness this day was the fact that he was the most unappreciated thinker there. Of everyone attending, panelists and audience alike, here was the key personage. Because he and his work are at the forefront of organic/siliconic behavior dynamics .. where they are alike and where they are different. If anyone would have a window onto the potential sentience of contrived information systems, it's John Koza. And he was there, highly respected, but glossed over because of the hub-bub caused by the publicized writings of some of the other panelists.
He, more than anyone, could probably quote chapter and verse on the plural-dynamics of multi-leveled and integrated systems. No single dynamics rule. No one set of rules - mathematical or otherwise - prevail. Information is codified and transmitted and reformed in ways that bespeak several kinds of 'sentience' at every alternate level of organization .. co-present and simultaneous. Yet no one ventured to investigate his depth of understanding about these issues. A lost opportunity. I doubt it will be passed over again. Next time, someone will take notice. New questions, new answers will be heard.
In the end, since there was a house full of people reluctant to really venture new thoughts, Hofstadter replied to me as the crowd thinned and admirer's clambered around the podium for contact with him that, no, he hadn't heard anything new today that sparked of the grail, the New Theory of 'life, the universe, and everything' that he's hoping will reveal itself. But he is still hopeful, still searching. [Though, if Bill Joy's closing remark intimated he was on to something different, Doug ought to spend more time with him to find out what it is.]
And here's a funny twist. I, like many, find Hofstadter's
extensive knowledge and philosophical turns very admirable. I handed him my card and
asked if he'd consider contributing something to the Ceptual Institute website. He
differed, saying, "The web? Oh, no. I never get involved with anything having to
do with the internet." [now there's a surprise .. the AI at the small
specialist avoiding AI at the techno-gaian scale?] I smiled, and we shook
hands. He had brought together some extraordinary minds this day. They had
bared their professional souls, beliefs, and commitments. A personal coup would have
been interesting - since the CI website harbors some of what he's looking for. But
most importantly, the accomplishments, shortfallings and re-clarified issues of natural
versus contrived mind-power were brought into rich focus this April first Saturday -
no foolin'. The optimisms and anxieties cavorted in vigorous pas de dux. All
the cards are on the table. And the action is still going on. If you want a voice,
pull up a chair and have a seat. Everybody can play, help design what will be. [I
voice this especially to the several Complexity lists members whom I'm writing this piece
for. The whole day there was only one reference from the audience to
'complexity' ideas per se .. and not even a 'tip of the hat' to fuzzy logic, soft
computing, at all.]
April 5, 2000
By the way, here are the Ceptual Institute's 'house rules'. If you were at Stanford on April 1, 2000 and have other comments or views on what was said and what happened, write them up and I'll post them right along side this rendition. Every panelist and audience member has a standing invitation. No editorial restictions imposed. (well ... hardly any. :-} length and overt grandstanding will be watched for. :-} )
§ § § § § § §
CI Website Sections
| THE INTEGRITY PAPERS ( main papers) GENRE WORKS (world writers) CONVERSATIONS DIALOGUES MINDWAYS POETICS (about Integrity ideas) |