Prime Time Replay:
Andrew Leonard
Technology journalist and author of
Bots: The Origin of New Species
MsgId: *emedia(1)
Date: Mon Sep 8 20:17:25 EDT 1997
From: DaveThomer At: 206.214.136.113
Tonight E-Media welcomes Andrew leonard, technology journalist and author of "BOTS: THE ORIGIN OF NEW SPECIES." In his book, Leonard examines the evolution of the automated programs that have become vital parts of the Net -- and, according to some in the computer world, precursors to truly intelligent artificial agents. Leonard explores the culture of the Internet, where the values of anarchy and anonymity have created an environment where bots can flourish, and shows how we are headed for a future where bots are both the agents of chaos and our best defense against it.Tonight's interview will be open to questions from the audience beginning at about 9:20. Remember to include your name in your post, as you will otherwise be identified only as "guest."
Good evening, everyone, and welcome to E-Media. I'm Dave Thomer, and tonight I'd like to welcome Andrew Leonard, author of "BOTS: THE ORIGIN OF NEW SPECIES." Welcome, Andrew.
MsgId: *emedia(7)
Date: Mon Sep 8 21:00:05 EDT 1997
From: Andrew At: 140.174.162.218
Glad to be here. Thanks for having me.
MsgId: *emedia(8)
Date: Mon Sep 8 21:01:20 EDT 1997
From: DaveThomer At: 206.214.136.113
Before we start talking about the book, I'd like to discuss the evolution of your own career for a minute. Techno-journalism wasn't always what you had in mind, was it?
MsgId: *emedia(10)
Date: Mon Sep 8 21:04:11 EDT 1997
From: Andrew At: 140.174.162.218
Oops. Got caught by the refresh. Actually, I started out as a China hand. Lived in Taiwan, spoke chinese, imagined myself in the romantic role of foreign correspondent. But freelancershave to go to where the market is, and right here in the Bay Area, right now in the 90s, technoculture is how a writer can pay the rent.Just got booted by Netscape. Sorry again. Anyway, I started reporting about the Internet about six months before it became a bonafide mainstream phenomenon, and that little advantage has given me a leg up ever since.
MsgId: *emedia(12)
Date: Mon Sep 8 21:07:26 EDT 1997
From: DaveThomer At: 206.214.136.113
In general, what impressions have you formed of the "technoculture"? How is it different than what the less-than-completely-wired experience?
MsgId: *emedia(13)
Date: Mon Sep 8 21:08:28 EDT 1997
From: AndrewLeonard At: 140.174.162.218
The narrative of Bots actually parallels my career. It started out as a column for a San Francisco weekly newspaper, then became a feature story for Wired magazine, and finally, a book. I'm still waiting for calls from Hollywood, though.The technoculture is where people and computers interact. And the real story is that the same stories take place in the technoculture as anywhere else. The same greed, avarice, unexpected joy, strange mysteries. The key to reporting the technoculture is to look for the elements of the good story, not to be entranced by the latest bell and whistle or the next "revolutionary" breakthrough.
MsgId: *emedia(16)
Date: Mon Sep 8 21:13:34 EDT 1997
From: DaveThomer At: 206.214.136.113
Regarding BOTS -- I really enjoyed the book; it seemed to capture just the right mix of the philosophy, psychology, and technology involved. Since "Bots", as you say in the book, is such an elusive term, would you mind giving us a brief ideas of what YOU mean by "bots?" And how did you get interested enough in them to write a book?
MsgId: *emedia(17)
Date: Mon Sep 8 21:15:50 EDT 1997
From: AndrewLeonard At: 140.174.162.218
Bots are: supposedly artificially intelligent, autonomous, software programs that we endow with some level of anthropomorphic character. I got interested in them because when I first started running across them, it seemed like the cyberpunk science fiction of authors like William Gibson and Neal Stephenson was coming to life right before my eyes.Let me add -- most importantly -- bots are ideally supposed to do something for us in the wilds of cyberspace -- defend us against junk mail, shop for us, find information for us -- but every bot that has been created has also been abused for purposes both malicious and downright greedy.
MsgId: *emedia(19)
Date: Mon Sep 8 21:19:10 EDT 1997
From: DaveThomer At: 206.214.136.113
The anthropomorphic element particularly caught my attention -- it was a great illustration of the fact that it seems many people really want to believe that we can create something that will relate to us, and will go to extraordinary lengths to do so. The stories about people carrying on extended conversations with chatterbots was fascinating. Were you surprised to see stories like this? And why do you think we act like this?
MsgId: *emedia(20)
Date: Mon Sep 8 21:21:07 EDT 1997
From: AndrewLeonard At: 140.174.162.218
Once I had interacted with chatterbots extensively myself, I was somewhat surprised at the stories of people who had fallen in love with bots like the famous Julia, or had taken the psychotherapist Elizabot at face value. Because, in my own opinion bots are quite stupid...But, then again, I was coming into these interactions with all the intensity of a paleoanthropologist discovering a live dinosaur -- I knew far too much about my subject to be fair. Many people are encountering bots in situations where they do not expect to encounter them, and thus they are willing to forgive a lot of outright stupidity, simply because of the medium. Additionally, people, in general have a tendency to populate the world with creatures of their own imagination. The Japanese gave every rock and tree a spirit. Californians created pet rocks. I myself have been known to curse at my computer. So to give a software program personality is not that much of a stretch.
The really interesting questions come when we talk about programs actually having emotions, or thinking, and we're not just projecting, we're talking about some real definable quantity. There are a lot of artificial intelligence researchers and corporate product managers attempting to figure this out, but so far, success has been limited. Artificial intelligence, even for so rudimentary a creature as a bot, is a tough nut to crack, much tougher than most computer scientists anticipated.
MsgId: *emedia(24)
Date: Mon Sep 8 21:26:18 EDT 1997
From: DaveThomer At: 206.214.136.113
Reading about the bots from the comfort of my chair, they seemed fairly stupid to me too . . . but there are many who feel that we're on track to develop truly intelligent, autonomous agents, and that bots are a great stop in that process. You touch on the debates in artificial intelligence in the book, and talk about hopes for AI have falled . . . do you think bots will suffer the same fate?If you have questions for Andrew Leonard, ask them now. Remember to post your name with your question!
MsgId: *emedia(26)
Date: Mon Sep 8 21:29:28 EDT 1997
From: AndrewLeonard At: 140.174.162.218
I think bots are a real test case. Bots, and their close cousins, intelligent agents, are where the grant money, corporate attention, and research stars are no focused. Microsoft alone is pouring billions of dollars into bot-related technology. I definitely don't think bots will ever duplicate human intelligence -- because human intelligence grows out of the embodied experience of being human. But I think it's quite likely that bots could eventually have their own bottish intelligence -- like, say, the way dolphins are intelligent, or E.T. was intelligent. And I'd say it's not entirely beyond the bounds of possibility that we could create something that could redesign itself. And then the whole game would be up.Because once bots could design their own, new, improved bots, their learning curve might shoot up so fast as to be well, nigh inconceivable. Science fiction writer Vernor Vinge calls such an even a "technological singularity." I kinda hope I don't live to see it. I don't talk much about it in my book, because, my book is about what is happening now -- cyberpunk NON-fiction. And bots aren't yet mutating better, smarter bots, at loose in the fields of cyberspace.
MsgId: *emedia(28)
Date: Mon Sep 8 21:31:29 EDT 1997
From: DaveThomer At: 206.214.136.113
If bots did develop their own intelligence, their own frame of reference -- would we be able to interact with them? We certainly wouldn't be able to treat them as nothing more than tools anymore -- what would they be?
MsgId: *emedia(29)
Date: Mon Sep 8 21:33:55 EDT 1997
From: AndrewLeonard At: 140.174.162.218
Well, that's a good question? Would we understand what goes on in the mind of a bot? Of course, if they were really smart, they'd figure out a way to make us understand. I think we'd find a way to be able to interact with them -- because part of their essential nature is as interface between human and computer, between the biological and the digital. Right now, bots are our cyborg extensions into the realm of cyberspace, and one would like to think that they'd never forget their roots.
MsgId: *emedia(30)
Date: Mon Sep 8 21:35:28 EDT 1997
From: DaveThomer At: 206.214.136.113
I just hope they wouldn't resent all that time being bossed around . . .
MsgId: *emedia(32)
Date: Mon Sep 8 21:36:47 EDT 1997
From: AndrewLeonard At: 140.174.162.218
Yeah, grumpy bots could really screw up the Net. But your question does get to the nub of an important issue. Since so much of what is hyped as the essential importance of intelligent agents and smart bots is their putative ability to help us deal with info overload -- to understand what truly is and isn't meaningful to us, then their whole utility lives and dies on this idea of effective communication. And that is one of the most difficult things in the AI field -- natural language communication.
MsgId: *emedia(31)
Date: Mon Sep 8 21:36:35 EDT 1997
From: guest At: 206.26.227.220
What do you think of bot learning. Will there be designer personalities like president and movies stars, albeit shallow impersonations? Will bots spend a long time training with their clients, perhaps a long period of cultivation like raising a child?
MsgId: *emedia(33)
Date: Mon Sep 8 21:39:37 EDT 1997
From: AndrewLeonard At: 140.174.162.218
I think bot learning will get better and better. I think a lot of interesting work is being down with genetic algorithms that allow software programsto breed better versions of themselves according to how well mutated versions solve specific problems. This is a particularly hot area in computer games. There are already gamebots that use genetic algorithms to learn your strategy and become better opponents over time.I think bots and humans will be constantly embracing each other in close, personal relationships. And a really effective program should learn from you, should benefit from traiing. The question is will we have the patience to train a bot. I hardly have enough patience to train a voice operated computer interface.
MsgId: *emedia(34)
Date: Mon Sep 8 21:41:13 EDT 1997
From: AndrewLeonard At: 140.174.162.218
I apologize for my many typos. But I put them in there just so you can tell I'm not a bot. (Actually, Julia, the chatterbot, was programmed to create typos and sometimes pause for several seconds to increase her believability.)
MsgId: *emedia(35)
Date: Mon Sep 8 21:42:07 EDT 1997
From: DaveThomer At: 206.214.136.113
Perhaps I'm completely off base, bit it has always seemed to me that you couldn't create an entity that would use natural language effectively until you created an entity that could directly experience the things to which language refer. But bots are designed to exist in cyberspace -- what are the chances that we'll ever get them to really know what we're talking about, so that a Web search for, say, information about the music group R.E.M. doesn't also pull up sites devoted to sleep research?It's interesting that gamebots would be where a lot of the newest bot techniques are beng employed -- it seems there are no lengths we won't go to in order to entertain ourselves.
MsgId: *emedia(37)
Date: Mon Sep 8 21:45:42 EDT 1997
From: AndrewLeonard At: 140.174.162.218
To really know what we're talking about...That's the 64,000 dollar question, isn't it? Talk to a natural language processing researcher and they'll say it's just a matter of incremental progress. Sooner or later we'll crack the code of linguistic meaning and figure out a nice set of algorithms that maqke it all comprehensible. But there are others who are more skeptical. Part of the equation is that we have to learn how to talk to bots, as well as having them learn how to understand it. So you should be smart enough to ask your bot too look for information about the "music group" "REM" -- and not just REM. Eventually, this could be a problem though -- we'll all become as dumb as bots because that's how we're trying to think.
I believe computer/arcade/playstation gaming is a bigger business, in terms of annual revenues, than Hollywood filmmaking. So its no wonder that a lot of applied research goes on there.
MsgId: *emedia(39)
Date: Mon Sep 8 21:48:35 EDT 1997
From: guest At: 206.26.227.220
I guess the implications of the low intelligence, language problems is that bots will not be spontaneous, but predictable literal servants much like other software applications. I wonder what would happen if contact with them could be filtered through other people or they only responded once a year. Would we ever see them as profound? Perhaps only engineers will appreciate bot adventures, the tales of wayward data packets and circuitous algorithm development. Sounds like a cult to me. Sola
MsgId: *emedia(40)
Date: Mon Sep 8 21:50:03 EDT 1997
From: AndrewLeonard At: 140.174.162.218
One really interesting question is whether, as computer networks continue to merge entertainment, news, financial markets and everything else into one big seamless network -- what could happen if a gamebot escaped from a game environment into, say Wall Street. Yikes!
MsgId: *emedia(42)
Date: Mon Sep 8 21:51:00 EDT 1997
From: DaveThomer At: 206.214.136.113
It's like the difference between talking to a child and talking to a normal adult . . . as a child learns to speak, you have to explain everything, because the child can't make the inferences and draw conclusions from incomplete data that an average adult can. I've always thought it an amazing faculty; I'm not surprised we're having trouble duplicating it.But I wanted to ask about how the bot story illustrates our behavior . . . the development of bots seems to showcase the full range of human potential, from incredibly altruistic to downright sadistic. What do you think we can learn from the way we use bots, or would like to use them in the future?
MsgId: *emedia(43)
Date: Mon Sep 8 21:53:09 EDT 1997
From: AndrewLeonard At: 140.174.162.218
Oops. sorry about the double post. In response to the earlier question -- a cult? Hmm. I think the story of bots is about a lot more than wayward packets. Microsoft's Office 97 is loaded with primitive bots -- the infamous dancing paperclip help system is just one step on the way to fully autonomous helpbots that always get in the way. And it's been somewhat controversial. Not everybody wants a bot getting between them and their computer.If you've wondered about a dramatic increase in junk mail this year, you can blame a particular subgenus of bots -- the dreaded spambot, which spends all its time wandering the Web looking for email addresses and then sending messages to whatever it finds. Spambots are a populist issue, in my belief.
MsgId: *emedia(44)
Date: Mon Sep 8 21:55:13 EDT 1997
From: guest At: 206.26.227.220
There will be bot police, bot hunters and bot inquisitors, some human and some themselves bot.
MsgId: *emedia(45)
Date: Mon Sep 8 21:56:29 EDT 1997
From: AndrewLeonard At: 140.174.162.218
In answer to Dave's question -- I think the essential thing to learn from the bot experience so far is that they are inextricably linked to their human creators. Right now bots are an interesting sideshow in the fast-developing Net, and they show us at all our best and worst -- annoybots and warbots underline how humans can be jerks in any environment. But Julia the chatterbot is just another example of human creativity and fun. As cyberspace matures, as it becomes more populated with humans, it will duplicate all the viral intensities of the real world -- and so will the bots. Bot police, indeed. Bot inquisitors, no doubt. Forewarned is forearmed.That seems like a fairly good note to end on -- unless there are any more questions?
MsgId: *emedia(47)
Date: Mon Sep 8 22:00:38 EDT 1997
From: DaveThomer At: 206.214.136.113
What scares me is that those viral tendencies will be unleashed in a world with fewer rules and less accountability -- the Net ideals of anarchy and anonymity have created a much different environment than the physical world. Will those values survive?
MsgId: *emedia(48)
Date: Mon Sep 8 22:02:14 EDT 1997
From: AndrewLeonard At: 140.174.162.218
Those values are under attack now, and the war has only just begun. I have no idea whether they will survive, but I'm not optimistic.
MsgId: *emedia(49)
Date: Mon Sep 8 22:06:00 EDT 1997
From: DaveThomer At: 206.214.136.113
Well, I'd like to thank you again for being here tonight, and best of luck to you with the book. "BOTS: THE ORIGIN OF NEW SPECIES" is now available from HardWired Press. That's it for this week's E-Media -- thanks for being with us, and we'll see you next week on www.omnimag.com. Check out the site for a full schedule of this week's programs. Good night!
MsgId: *emedia(50)
Date: Mon Sep 8 22:06:07 EDT 1997
From: AndrewLeonard At: 140.174.162.218
Well, folks, I gotta go, my three year old daughter is demanding some quality daddy time, and I don't have a bot I can pass the job off on to, nor would I want to, anyway. Thanks again for having me!
Home || Prime Time || Live Science || Machine Dreams || Project Open Book || SF-Fantasy-Horror
Continuum || Antimatter || Mind-Brain Lab || Interactive IQ || Gallery || OMNI ToonsQuestions, comments and suggestions can be mailed to the webmaster.
Copyright © 1998 by Omni Publications International, Ltd. All Rights Reserved.