Ben Goertzel converses with Hugo de Garis on his transhumanist argument for the reality of a Creator.
BEN GOERTZEL: Hugo, you’ve recently published an article on KurzweilAI.net titled “From Cosmism to Deism”, which essentially posits a transhumanist argument that some sort of “God” exists, i.e. some sort of intelligent creator of our universe – and furthermore that this “creator” is probably some sort of mathematician. I’m curious to ask you some questions digging a little deeper into your thinking on these (fun, albeit rather far-out) issues.
Could you start out by clarifying what you mean by the two terms in the title of your article, cosmism and deism? (I know what I mean by Cosmism, and I described it in my Cosmist Manifesto book, but I wonder if you have a slightly different meaning.)
HUGO DE GARIS: I defined these two terms rather succinctly in the kurzweilai.net essay so I’ll just quote those definitions here. Deism is “the belief that there is a ‘deity’ i.e. a creator of the universe, a grand designer, a cosmic architect, that conceived and built our universe.” Cosmism is the “ideology in favor of humanity building artilects this century (despite the risk that advanced artilects may decide to wipe out humanity as a pest).”
Artilects are “artificial intellects, i.e. godlike massively intelligent machines, with intellectual capacities trillions of trillions of times above the human level.” Deism is to be distinguished from theism, which is the belief in a deity that also cares about the welfare of individual humans.
BEN: Previously you have talked about “Building Gods” as the ultimate goal of artificial intelligence technology. So is your basic argument in favor of deism that, if we humans will be able to build a god once our technology is a bit better – then maybe some other intelligence that came before us also was able to build gods, and already did it? And maybe this other intelligence built (among other things) us?
HUGO: Yes, pretty much. The traditional arguments in favor of a deity (as distinct from a theity, which I find ridiculous, given that last century 200-300 million people were killed in the bloodiest century in history – so much for a loving deity) are less strong and persuasive in my view than the artilect-as-deity argument. The rise of the artilect is based on science, and the extrapolation of artilectual intelligence to trillions of trillions of times above the human level, seems very plausible this century. If human beings (e.g. Prof Guth at MIT) have theories on how to build baby universes, then perhaps artilects could actually build them, and hence, by definition, become deities (i.e. creator gods). That was the main point of the kurzweilai.net essay.
BEN: I see… This is actually very similar to the “simulation argument” made by Nick Bostrom, Stephen Wolfram, Jonathan vos Post and others – that since creating computer simulations as complex as our universe is probably possible using advanced technology, the odds seem fairly high we’re actually living in a simulation created by some other intelligences. But Bostrom, for instance, focuses on the simulation itself, whereas you seem to focus on the intelligent entity involved, the assumed creator of the simulation.
And this brings up the question of what intelligence means. What’s your working definition of “intelligence”? In other words, how do you define “intelligence” in a way that applies both to human minds and to the potential super-intelligent universe simulation-creating deity that you hypothesize?
HUGO: As a common sense, man-in-the-street definition, I would say, “Intelligence is the ability to solve problems quickly and well.” As a research scientist, I am made constantly aware, on a daily basis, of the fact that intelligence levels differ greatly between individuals. I spend my time studying PhD level pure math and mathematical physics, trying to wrap my head around the works of Fields Medal winners such as Ed Witten, Michael Freedman, Richard Borcherds, etc, all mostly likely with extraordinarily high IQs. Believe me, with my only moderately high intelligence level, it gives me “brain strain”. So it’s easy for me to imagine an ultra-intelligent machine. I only have to imagine a machine a little bit smarter than these genii. I am in awe at what these genii create, at what the best examples of the human intellect are capable of. I am in awe. However at a neuro-scientific level, we don’t know yet what intelligence is. A five-year-old can ask questions about the nature of human intelligence that are beyond state-of-the-art neuroscience to answer, e.g. “What was so special about Einstein’s brain that made him Einstein?” “Why are some people smarter than most?” “Why is the human brain so much smarter than the mouse brain?” I dream of the creation this century of what I label “Intelligence Theory” (IT), that would provide real answers and understanding to such questions.
We should aim at a universal definition of intelligence that would be applicable to all levels of (humanly) known intelligence. It is an interesting question how far up the superhuman intelligence level a human concocted IT could go. One would think that the finite level of human intelligence, by definition, would preclude humans thinking of an IT at a level that an artilect could manage.
BEN: Following up on that, one question I have is: if there is some superintelligence that created the universe, how similar do you think this superintelligence is to human intelligences? Does it have a self? A personality? Does it have consciousness in the same sense that humans do? Does it have goals, make plans, remember the past, forecast the future? How can we relate to it? What can we know about it?
HUGO: My immediate reaction to that question is that with our puny human brains, we very probably can’t even begin to conceive of what an artilect might think about or be like. If we think that a universe-creating, “godlike” artilect has the human like attributes you list above, then that might a “category mistake” similar to a dog thinking that human beings are so much smarter and capable than dogs, that they must have many more bones lying around than dogs do. One thing that is interesting about this question though, is that by conceiving of the artilect as a scientific-based creation, we can begin to attempt answers to such questions from a scientific perspective, not a theological one, where theologians are all too likely to give all kinds of untested answers to their particular conception of god. Is a consciousness, or sense of self a prerequisite to the creation of superhuman intelligence? These are interesting questions, that I don’t have answers to. Perhaps I haven’t thought deeply enough about these types of questions.
BEN: In a nutshell, how does your deism differ from conventional religions like Christianity, Islam, Judaism and so forth? And how does it differ from Buddhism, which some have argued isn’t really a religion, more of a wisdom tradition or a practical philosophy?
HUGO: Traditional religions such as the above, that were invented several thousand years ago, after the agricultural revolution and the growth of cities (with their occasional genius level priest-theologian) I find ridiculous, utterly in conflict with modern scientific knowledge. The cultural anthropologists of religion have shown that humanity has invented on the order of about 100,000 different gods over the broad sweep of history, and across the planet. These many gods are so obviously invented (e.g. New Guinea gods have many pigs, etc) that their human origins are obvious. However, the fact that every primitive little tribe has invented its own gods makes one deeply suspicious that human religiosity is in fact physiologically based, and hence has Darwinian survival value (e.g. if you can believe in a loving god, you are less likely to commit suicide in a cold, indifferent, callous universe, so religiosity-inducing genes would be more selected for).
Deism, on the other hand, especially with the artilect-as-deity argument, is much closer to modern science in its conception. The very real possibility of the rise of the artilect this century virtually forces anyone confronted with the argument to accept its plausibility. Our sun is only a third the age of our universe, and there are a trillion trillion 2nd generation stars that we can “observe” that probably have generated life and intelligence. Once a biological species reaches an intelligence level that allows it to have mathematics and science, it is then only a small step for it to “move on” to the artilectual stage, whose potential intelligence is astronomically larger (pun intended) than any biological level. An artilect of the distant past in an earlier universe may have designed and built our universe. It would have been our deity.
BEN: Traditional religions serve to give people comfort and meaning in their lives. Do you think that the form of deism you advocate can serve the same purpose? Does it serve that purpose for you – does it make you feel more meaning in your existence, or in existence in general?
HUGO: I look down on traditional “religionists” as ignorant deluded fools. The fact that where I lived when I was living in the US, namely Logan, Utah, there were hard-science professors who were converted Mormons, showed me that the human brain is modular, with some compartments isolated from other parts, e.g. the religious areas from the critical analytical scientific areas, so that these professors were unable or unwilling to destroy their religious beliefs with just a little analytical scientific thinking. I don’t have much patience with people who have low “RQs” (reality quotients). If I present these religionists with the idea that many tens of millions of theists last century were killed in the bloodiest century in history, they just block thinking about its implications. If I show them the evidence that humanity has invented 100,000 gods, they do the same. I don’t deny that if one is able to believe in a loving god, it might be comforting. Especially to someone who is, in Darwinian terms, sexually unattractive, and gets no human partner, so remains unloved – especially older widows, whose supply of men has run out due to the greatest of human realities, death. But emotional comfort and high RQ may not be compatible. If forced to choose, I prefer not to be a gullible child. A non-theist deist looks at the massively indifferent universe as a given. Having “faith” is no argument to me. Believing something simply because one wants to believe it allows one to believe in the “tooth fairy.”