(Post updated from a December, 2010, post. Relates to my Search Rudy’s Lifebox page and my book The Lifebox, the Seashell, and the Soul.)
A lifebox is meant to be interactive software that allows a user to feel like they’re having a conversation with the person whom the lifebox supposedly emulates. I have an idea about how to create lifeboxes. In a nutshell:
• Amass a database containing all of a given person’s recorded words and images.
• Use realtime search algorithms to mine the data for nuggets keyed to questions.
• Use syntactic rules to merge each search’s output into a coherent answer.
In reality, this is how I usually answer questions in conversations. Rather than thinking hard, I just search out vaguely related thought-snippets and stick them together. And of course I modulate my mental searches according to my short-term memory of the conversation thus far.
In recent years, several commercial ventures have taken up the idea of helping people to create virtual software versions of themselves. For whatever reason, these developers never seem to mention my name. Yet they are in fact talking about the “lifeboxes” I’ve been discussing for decades. See the paperback, ebook, and online ecitions of my book The Lifebox, the Seashell, and the Soul .
I go into considerable detail about lifeboxes in my 1999 novel, Saucer Wisdom, in my non-fiction tome of 2005 The Lifebox, the Seashell, and the Soul. I have that book online as a web page, and you can read the relevant “Lifebox” section here,.
See also my 2009 article “Lifebox Immortality” which I co-authored with Leon Marvell. And creating software emulations of specific human minds is a notion I discussed in my 1982 novel Software .
A lifebox is really a kind of chatbot. In December, 2010, I met an interesting couple, Bruce and Sue Wilcox. Their chatbot Suzette just won the 2010 Loebner Prize for doing the best job at the Turing Imitation Game, that is, the game of the chatbot program trying to convince a human judge that the chatbot is human too. They talk via an instant-message interface. If a program could reliably and consistently win at the Imitation Game, we’d be included to say it had achieved human-like intelligence. Looking at the chatbot site describing Suzette, I was surprised to see how widespread and popular this programming exercise has become.
I’ve always thought it telling that in Turing’s 1950 article proposing this test, he begins by talking about a different kind of test—in which someone interrogates subjects via instant-messaging and tries to decide whether they are male or female. I’ve integrated my thoughts about this into my novel in progress, Turing & Burroughs, in which Turing does in fact impersonate a woman.
[These images have essentially no connection to the material being discussed.]
The day after I met the chatbot programmers, the dapper and fanciful writer Mark Dery put a new essay online, “Hate is All Around: The Politics of Enthusiasm (And Its Discontents)”, and near the end (on page 6) he mentions my writings about my concept of the lifebox. He also got off what seemed like very funny lines, as when he characterized a petulant remark by Sarah Palin’s daughter Bristol as “a characteristic display of the social grace and subtlety of mind that have made her mother so universally admired on the world stage”.
Given that I’d been talking about the lifebox with the Wilcoxes, reading the Dery article was enough of a push for me to finally make an alpha-release version of a lifebox…and here’s the same link I mentioned above: the Search Rudy’s Lifebox page. I haven’t really done any work on the Search Rudy’s Lifebox tool since December, 2010, but I still think about it.
The reason my alpha version “Rudy’s Lifebox” does function a little bit like a chatbot emulating me, is that I have, over the last ten or twenty years, been placing really large amounts of my writing online.
Note that AI is not in the picture here at all. Rather than tilting at the Quixotic task of writing logic software to imitate human thought, it’s much simpler to beat the chatbot emulator problem to death with big data and fast hardware.
This said, Rudy’s Lifebox is not going to write the next chapter of my novel for me, nor rough out my next painting. So is my lifebox project a waste of time? Well, Rudy’s Lifebox can function as an aid to my fading memory. For instance, remembering just now that there was something about “wasting time” in a book, Be Not Content, that I love, I entered be not content waste time into my Search Rudy’s Lifebox box, and found this from an old blog post of mine:
I’m always worrying about wasting time, right, and I saw a great line in Be Not Content, the author-narrator Abel Egregore expresses this fear to one of his stoner friends, who guffaws, “Time? How can you waste time?” And I get a little enlightenment there. Time and space, the all-pervasive ineluctable modalities. What’s to waste? You use one second per second no matter what you’re doing. A wonderful teaching.
Up here (or down here) in some version of real time—my family and I decorated our Christmas tree today!
And once again, here’s that Search Rudy’s Lifebox link one more time. Do try it and make a comment over there, or on this here “Making a Lifebox” page.
December 14th, 2010 at 6:14 am
A search for “Gnarl” gives 10 pages of hits so far.
December 14th, 2010 at 2:13 pm
You would want an intelligent google, that read through it and cross referenced indexed it by meaning information as opposed to just keywords, and then a chatbot that could interface to that database and speak out your views and opinions. I’m working on it, but it will be quite a while.
December 14th, 2010 at 2:22 pm
Good comment, Bruce, and good to hear from the chatbot master.
I have a feeling that the “intelligent google” engine underlying a convincing lifebox chatbot will not have to be a complex and AI-like as is commonly imagined.
If we did indeed need for the program to in some sense ferret out the meanings of the data phrases, we’d be facing an intractable task, and perhaps even a formally unsolvable problem.
My sense is that some fairly cheap tricks can do the job well enough—as has historically been the case for every successful AI application so far: it’s all cheap tricks.
In practice, when people ask me questions, I very often don’t really reply to the sense that they may have had in mind. Instead I go with a quick mental search based on the keywords from, not only the current question, but the last two or three questions and answers. And if my answer seems in some sense irrelevant—well, that’s the nature of talking to a human being.
People do change the subject and drift off topic all the time.
Note that the Search Rudy’s Lifebox link tends to return “too many” answers if you only put in one or two keywords. One useful trick might be, in the background, to keep adding more and more keywords from the recent conversation until you get down to only a small number of search results. Then pick one you haven’t used yet and extract a couple of consecutive sentences as the answer, possibly massaging their grammar a bit to be in the right tense and so on.
December 14th, 2010 at 2:35 pm
Hi Rudy, we met in Milan a few years ago (FuturShow 2004) and I have not forgotten your concept of “lifebox immortality”, very similar to my concept of “virtual personality” (presented in an article of 1992 on the Italian newspaper “La Stampa”). I am amazed that you waste your time with the infamous Turing Test, which has a negative contribution to research in the field of artificial intelligence. Dear Rudy, simulation of (human) intelligence is not intelligence!
I then took a look at your Search Rudy’s Lifebox. You not believe that a Google search box is something similar to immortality, right?
Please watch my http://www.semplicity.com, then test my http://www.intellibook.net and tell me what you think about.
PS: please note Semplicity as Stockholm Challenge 2010 Finalist in the CULTURE category http://www.stockholmchallenge.org/challenge-2010/culture
December 14th, 2010 at 8:55 pm
Isn’t that the point? We are just a bunch of cheap tricks?
December 15th, 2010 at 3:15 am
Google say their mission is “to organize the world’s information and make it universally accessible and useful.”
But I think we should be organizing our own information and make it as universally accessible as we ourselves determine.
At the moment Google and Facebook etc, are making money out of our information. Why can’t we do this ourselves? Chatbots with P2P streaming and an open source search thats what we we need!
December 15th, 2010 at 9:31 am
It would be nice to see a list of input keyword rankings by frequency. If there are several keywords that you use very frequently (maybe they are in your top 100 or 1000 by rate), this doesn’t cut down the number of results much, but if you’ve only used a word a few times, this word determines the hits.
December 15th, 2010 at 10:41 am
Actually, the task is not quite so bad. Unlike “chat”, your archived data is mostly legal complete sentences, which means they can be be parsed into the basic subject,verb,object components. Lookup based on that would not be so bad, partcularly as I already have a verb hierarchy of 1/3 of the english language verbs and a noun hierarchy of some 50,000 nouns. So after that, more or less standard chatbot patterns could map “what pets have you had” into the more generic “what pets * you ~own” which could locate sentences indexed on you possessing in some way an object which is an animal of some kind.
December 15th, 2010 at 12:47 pm
It may be possible to waste time and space simultaneously.
“In his proposal, time and space can be converted into one another, with a varying speed of light as the conversion factor. Mass and length are also interchangeable, with the conversion factor depending on both a varying gravitational “constant” and a varying speed of light (G/c2). Basically, as the universe expands, time is converted into space, and mass is converted into length. As the universe contracts, the opposite occurs.”
http://www.physorg.com/news199591806.html
December 18th, 2010 at 1:18 pm
RR: I’m with Bruce Wilcox. I like your first stab at a Lifebox, but at the risk of lecturing Nikola Tesla on alternating current, I thought the essential element in the Lifebox was its interactive ability, not to mention its heuristic ability. Elsewise, a writer’s collective works are, de facto, a Lifebox, right? I want to be able to access an interactive simulacrum of Rudy Rucker. To be sure, it will be based on Rudy’s collective writings, interviews, lectures, diary entries, doodles, paintings, ephemera, and errata, but it will be more than the mere Google-searched sum of those parts because it will interact with me, hopefully in a far more evolved fashion than mere ELIZA mimicry, and it will learn as it goes. I want a virtual Rudy I can question, discuss, debate with. Imagine if we had a PKD lifebox, the simulacrum of Phil’s gnostic dreams!
December 18th, 2010 at 4:09 pm
Thanks for the comments, all.
Mark, I admit that what I have here is really just a kind of Rudy-data engine, and it would need a smooth chatbot on top if it as an interface.
Bruce makes the point that getting a chatbot to be convincing is indeed easier if you have a very large data base of “personality sentences” to draw upon. The successive linked search results would give a reasonably smooth sense of interacitivity. Making a chatbot simulation of someone’s particular lifebox database is an easier problem than making a standalone chatbot of a more universal nature.
The program could indeed “learn” over time…in sense of adding new links or of training a neural net. At the very least, for a given user, the program could store transcripts of prior conversation with that particular user so as to avoid senile repetitions.
I had actually heard of there being a program like this based on Phil Dick’s notes, I think it accompanied his (now lost) artificial head.
But, yes, these programs would not at this point be creative, other than via the Surrealist juxtaposition mode.
December 22nd, 2010 at 6:51 am
Hello Rudy!
Just today I read about your LifeBox, wich is super interesting and I’ll read/study along holydays. Curiously today, while cleaning my Google Reader, I face another life box:
http://www.lifeboxcompany.com/
Maybe a sign? o_O
Congrats for your wonderful work!
Cheers
December 25th, 2010 at 12:25 pm
Happy holidays, Mr. Rucker. It seems to my almost 79 years old mind that there’s only one quest, it’s to find out who are liars and who are truth tellers in the world. Not many individuals doubt their own truth. Lifebox r us.
It also seems almost certain that the universe is not operating like a computer, but there is some aspect of our minds/brain/body unified that does though. This ‘intelligent operator’ is probably la feed back process, we alter and are altered by it. I’ve seen changes but also the change-proof side of life, in very every day situations, ‘like’ I moved along a moebius band as I aged.
William Blake seems to me to have understood everything, and he wasn’t ashamed to admit he didn’t ‘think’ his thoughts, they were given to him, he was the recipient.
December 28th, 2010 at 10:48 am
What a life! If there is a computation behind what people say (to me, about me) it results in specific words so maybe there is a generalized formula and its installed in the males I’ve interfaced with throughout my life: “You can’t stick to a subject.” “You are digressing.” “You are off topic.”
The topic of reality, and the search for the real one was
initiated in my life when I was about 14, when someone in authority told me I’d imagined something that seemed to me to have really happened.
I’m not sure what kind of formula would compute the immense interest I had when in the 1950’s I viewed the Play of the Week Rashomon and after it ended, a thought was produced that the play very likely explained why people quarrel so about reality. The understanding that the play produced in a fraction of a second probably, has endured for 5 decades now. The individuals in the play each experienced a different event, their understanding was so different in every detail that it seemed reality was non-sense.
That introduction brings me to my question: What kind of search could I do that would give information about what is the meaning in this statement from Jung’s Red Book which is now available as a pdf. It’s in The Way to the Self chapter:
“In I9I8, Jung wrote a paper entitled “On the unconscious,” where
he noted that all of us stood between two worlds: the world of
external perception and the world of perception of the unconscious.”
I found myself unable to get meaning from ‘the ‘world of perception of the unconscious’, I wondered about how to interpret it as Jung intended it. Did he mean “the world of perceptions in the unconscious” or ‘from the unconscious’ or ‘through the unconscious’ and if it’s unconscious how is it visible? Lots of questions…
How would you understand the statement or since I’ve not read all of your book, have you written about the ‘world of perception of the unconscious’ in some other form?
December 28th, 2010 at 9:41 pm
“Lifebox Immortality,” which is my half of an essay I wrote with Leon Marvell, appeared online in h+ magazine this week, edited by my former Mondo co-conspirator, R. U. Sirius.
February 24th, 2011 at 12:13 pm
A suggestion I’d like to offer for getting started with a virtual lifebox would be “Know Thyself”, from the Oracle at Delphi. (As of Feb, 2012, H+ is no more and now the article is on my website as PDF file: Rucker/Marvell “Lifebox Immortality”.)
March 2nd, 2011 at 7:16 am
It’s not often that someone quotes “Be Not Content.” Thanks Rudy.
August 2nd, 2011 at 2:26 am
Yo Rudy,
I searched for ‘Dave’, no results were found!?! That’s not cool brah…
If you want to up your gnarl maybe consider adding a character based on me in this new Turing book you’re hacking at the mo, I live like 5 minutes away from Bletchley Park after all!
Stay wiggly
Dave
March 2nd, 2012 at 11:02 am
contributing:
https://github.com/siznax/lifebox
at the time of writing i get 403 on your h+ article. here’s a copy:
http://tomobot.net/95
also, an HTML copy of Turing’s paper:
http://tomobot.net/3u
thanks, Rudy, for sharing your ideas, and doing something about them.
March 2nd, 2012 at 6:02 pm
Steve, I’m guessing that H+ is no more. I’ll change the reference in my earlier comment too. The article is on my website as PDF file: Rucker/Marvell “Lifebox Immortality”