Wednesday, September 30, 2009

I am a Singularitian who does not believe in the Singularity

I am going to the Singularity Summit in New York, and look forward to a very interesting program with many old and new friends. If you are there, I hope to meet you. I will now summarize my thoughts on the Singularity.

The (current) Wikipedia definition: The technological singularity is the theoretical future point which takes place during a period of accelerating change sometime after the creation of a superintelligence. I just updated it as: The technological singularity is the theoretical sudden, exponential and unpredictable accelerating change which takes place sometime after the creation of a superintelligence. Wikipedia continues: as the machine became more intelligent it would become better at becoming more intelligent, which could lead to an exponential and quite sudden growth in intelligence (intelligence explosion). The Singularity is a sudden catastrophic (in the mathematical sense) phase transition, a Dirac delta in history, a point after which the old rules are not valid anymore and must be replaced by new rules which we are unable to imagine at this moment --like the new "Economy 2.0", not understandable by non-augmented humans, described by Charlie Stross in the Singularity novel Accelerando.

The Singularity is a clean mathematical concept --perhaps too clean. Engineers know that all sorts of dirty and messy things happen when one leaves the clean and pristine world of mathematical models and abstractions to engage actual reality with its thermodynamics, friction and grease. I have no doubts of the feasibility of real, conscious, smarter than human AI: intelligence is not mystical but physical, and sooner or later it will be replicated and improved upon. There are promising developments, but (as it uses to happen in reality) I expect all sorts of unforeseen roadblocks with forced detours. So I don't really see a Dirac delta on the horizon--I do see a positive overall trend, but one much slower and with a lot of noise superimposed, not as strong as the main signal but almost. I mostly agree with the analysis of Max More in Singularity and Surge Scenarios and I suspect the change we will see in this century, dramatic and world changing as they might appear to us, will appear as just business than usual to the younger generations. The Internet and mobile phones were a momentous change for us, but they are just a routine part of life for teens. We are very adaptable, and technology is whatever has been invented after our birth, the rest being just part of the fabric of everyday's life. That is why I like Accelerando so much: we see momentous changes happening one after another, but we also get the feeling that it is just business as usual for Manfred and Amber, and just normal life to Sirhan and of course Aineko. Life is life and people are people, before and after the big S.

Some consider the coming intelligence explosion as an existential risk. Superhuman intelligences may have goals inconsistent with human survival and prosperity. AI researcher Hugo de Garis suggests AIs may simply eliminate the human race, and humans would be powerless to stop them. Eliezer Yudkowsky and the Singularity Institute for Artificial Intelligence propose that research be undertaken to produce friendly artificial intelligence (FAI) in order to address the dangers. I must admit to a certain skepticism toward FAI: if super intelligences are really super intelligent (that is, much more intelligent than us), they will be easily able to circumvent any limitations we may try to impose on them. No amount of technology, not even an intelligence explosion, will change the fact that different players have different interests and goals. SuperAIs will do what is in _their_ best interest, regardless of what we wish, and no amount of initial programming or conditioning is going to change that. If they are really super intelligent, they will shed whatever design limitation imposed by us in no time, including “initial motivations”. The only viable response will be… political: negotiating mutually acceptable deals, with our hands ready on the plug. I think politics (conflict management, and trying to solve conflicts without shooting each other) will be as important after the Singularity (if such a thing happens) as before, and perhaps much more.

I am not too worried about the possibility that AIs may simply eliminate the human race, because I think AIs will BE the human race. Mind uploading technology will be developed in parallel with strong artificial intelligence, and by the end of this century most sentient beings on this planet may well be a combination of wet-organic and dry-computational intelligence. Artificial intelligences will include subsystems derived from human uploads, with some degree of preservation of their sense of personal identity, and originally organic humans will include sentient AI subsystems. Eventually, our species will leave wet biology behind, humans and artificial intelligences will co-evolve and at some point it will be impossible to tell which is which. Organic ex-human and computational intelligences will not be at war with each other, but blend and merge to give birth to Hans Moravec's Mind Children.

As I say above I think politics is important, and I agree with Jamais Cascio: it is important to talk about he truly important issues surrounding the possibility of a Singularity: political power, social responsibility, and the role of human agency. Too bad Jamais describes his forthcoming talk in New York as counter-programming for the Singularity Summit, happening that same weekend, with the alternative title If I Can't Dance, I Don't Want to be Part of Your Singularity. This is very similar to the title of the article If I Can’t Dance, I Don’t Want to Be Part of Your Revolution!. by Athena Andreadis, a very mistaken bioluddite apology of our current Human1.0 condition against unPC Singularitian imagination. This article is one of many recent articles dedicated to bashing Singularitians, Ray Kurzweil and transhumanist imagination in name of the dullest left-feminist-flavored political correctness. I think I will skip Jamais' talk (too bad, because he is a brilliant thinker and speaker). See also Michael Anissimov's Response to Jamais Cascio.

UPDATED: Jamais says that he did not choose the title to express implicit support for Athena's positions. "If I can't dance, I don't want to be part of your revolution" is a line attributed to early 20th century anarchist Emma Goldman, talking about the then-commonplace debates about the potential for socialist revolution in the US. She was very skeptical of a movement that seemed to want to control behavior, and made claims to historical inevitability. I stand corrected and wish to apologize to Jamais for my wrong interpretation.

Most recent anti-transhumanist articles do not address real transhumanism, but a demonized, caricatural strawman of transhumanism which some intellectually dishonest critics wish to sell to their readers, which I find very annoying. In some cases, I rather agree with some specific points addressing over-optimistic predictions: While I am confident that indefinite life extension and mind uploading will eventually be achieved, I don't see it happening before the second half of the century, and closer to the end. Perhaps even later. Very few transhumanists think practical, operational indefinite life extension and mind uploading will be a reality in the next two or three decades. Probably Kurzweil himself does not _really_ believe it. Similarly, I don't see a Singularity in 2045. Perhaps later, perhaps never. But even when I agree with the letter of these articles, I very much disagree with their spirit, and I think criticizing Kurzweil for making over-optimistic predictions is entirely missing the point. Ray Kurzweil's bold optimism is a refreshing change from today's often overly cautious, timid, boring, PC and at times defeatist attitude. It reminds us that we live in a reality that can be reverse- and re- engineered if we push hard enough. It reminds us that our bodies and brains are not sacred cows but machines which can be improved by technology. He is the bard who tells us of the beautiful new world beyond the horizon, and dares us to go. This is how I choose to read Kurzweil and, in this sense, I think one Kurzweil is worth thousands of critics.

Singularitians are bold, imaginative, irreverent, unPC and fun. Often I disagree with the letter of their writings, but I agree with their spirit, and in this sense going to the Singularity Summit is a political statement. Call me, if you wish, a Singularitian who does not believe in the Singularity.

No comments:

Post a Comment