For nearly twenty years, the Singularity concept has been floating around academic and futurist groups, alerting humanity to an issue far greater in importance than any other forthcoming advance, including nanotechnology and virtual reality. In 1996, Eliezer Yudkowsky published "Staring into the Singularity", interpreting the Singularity as something concrete, feasible and worthwhile that needs doing, as opposed to some speculative science fictional concept or overall sociotechnological trend. In 2000, Yudkowsky formed the Singularity Institute for Artificial Intelligence, the first and only human group devoted specifically to accelerating the Singularity and making sure it is a pleasant one.

In the past decade, most notably in the past three years, when most Singularity advocates were originally exposed to the idea, there has been a decent series of articles and publicity surrounding the Singularity concept. But this is where the problems begin - many popular accounts of the Singularity use incredibly convoluted definitions of it, drawing more attention to the general shock-aspect of transhumanism than the critical decisions humanity will need to make in order to survive the Singularity successfully.

Contributing to these befuddled, sensationalistic Singularity interpretations are dozens of futurists arguing the Singularity as a futuristic scenario to be defended, rather than a cause to be accomplished. As the technological barriers surrounding transhuman intelligence get thinner and thinner, this incorrect version of the Singularity meme continues to propagate, and will get its highest spike of publicity yet when Ray Kurzweil's "The Singularity is Near" comes out later on this year.

There is a positive aspect to this, however; even in their most diluted forms, alternate versions of the Singularity meme sometimes manage to set the groundwork for the proper absorption of the correct ideas. The problem is, where are the people to educate everyone on how best to approach the Singularity? They currently seem to be rare, but are emerging over time. Oxford philosopher Nick Bostrom has recently stepped forward with his paper "Ethical Issues in Advanced Artificial Intelligence", putting into words what he had believed for quite a while; the initial conditions of the first superintelligence are what will determine the human race's fate in the near future.

For those who have read all the arguments, studied them carefully, done the background research, and thought real hard; the conclusion is clear - we're going to need to set the initial conditions for the Singularity to go right, and these initial conditions (roughly speaking) should be benevolence, altruism, and respect for the wishes of others. To this end, Eliezer Yudkowsky came up with the "Friendly AI" idea, based on the premise that AI was the forerunner of several possible Singularity technologies. He also outlined the "Singularitarian Principles", guidelines for a new generation of philosopher-activists devoted to accomplishing the Singularity and preserving its integrity.

Yudkowsky has said, "The stage of development of the Singularitarian philosophy seems adequate to support the Singularity, although I would certainly like to see a lot more introductory material out there. By "adequate", I mean that an individual professing what seems to be the current state of the art in Singularitarian philosophy should hopefully be able to think coherently and rationally about the Singularity and the ethics thereof, and to devote effort to accomplishing good thereby." Most Singularitarians agree; there is enough philosophy out there that we should be able to work towards the Singularity and know what we're doing. After twenty or so years of analysis, it's about time we got our act together and start laying a groundwork for the accomplishment of general AI.

Which is exactly what the Singularity Institute was for. As of 2003, the Singularity Institute has received enough donations to keep Eliezer working on AI issues for the past two years, but building up an account large enough to start a serious AI project has been coming along slowly. The community surrounding the Singularity Institute has certainly been growing, and has attracted some extremely bright individuals, but the commitment is lacking. There are hundreds or thousands of Singularity philosophers, but very, very few Singularity activists; or Singularitarians.

How many Singularitarians exist today? Yudkowsky answers, "In terms of total commitment, I would say at least two, but probably less than a dozen. In terms of being interested enough to put forth major efforts, I would say at least ten and less than a hundred. In terms of being interested enough to make minor efforts, I would say at least fifty and less than a thousand. In terms of talking about it, I would say at least hundreds and less than a million. In terms of having heard of the Singularity and thinking vaguely that it might be a good thing, it could be millions - who knows? It's been showing up outside transhumanist circles lately."

Widespread Singularity awareness is good, but the numbers near the beginning of the quote desperately need to change. The Singularity is not something we should expect to go right by accident, or something we should necessarily expect to happen at all - in this age of exponentially advancing technology, global threats are emerging from all directions, and a Singularity will be averted if a sufficiently large disaster takes place. Creating "Friendly AI" is the most effective currently known strategy to avert the possibility of these disasters, setting the power of transhuman intelligence towards humanitarian ends.

Contributions to the Singularity effort will yield disproportionately more good than contributions to humanitarian efforts ignorant of the opportunity of transhuman intelligence, and contributions near the beginning of the Singularity effort, now, will yield disproportionately more good than contributions taking place later on. Singularitarians agree that there aren't many things on Earth, if any, more important than taking care of this issue as soon as possible.

How can we most effectively help the Singularity? Efforts stereotypically associated with the Singularity, such as the creation of faster computers, may contribute peripherally to the occurence of the Singularity; but not necessarily to that of a pleasant one. For humanity to survive the Singularity, the first transhuman intelligence has to care for us and want to help us; we want an intelligence with the moral complexity to view us as sentient beings worthy of value, rather than simply stalks of protein dotting the landscape - morally identical to rocks and air. Contributions to technology such as computers accelerate the arrival of UnFriendly AI and Friendly AI equally; so unless we preferentially accelerate Friendly AI, we aren't doing the world much good. This makes things much more complicated.

In order to better accomplish this goal of working towards the Singularity, I want to work more closely with other Singularitarians, pooling our knowledge and abilities in service of a common goal. We need to get together and brainstorm, in addition to getting to work on this extremely important challenge. So little is currently being done; as mentioned before, there are many who are interested in the Singularity on a part-time flirtacious basis - they see the Singularity as important, but don't want to see it get in the way of the life they were planning to lead (although it will eventually, even if ignored.) Another recurring problem is the case of he or she who encounters the Singularity movement, reads about the possibility of seed AI, and decides adamantly that they want to become one of the programmers, only to entirely lose interest several weeks or months later, when they find out that they simply aren't smart or experienced enough for the job.

Making the best possible effort for the Singularity doesn't necessarily mean glory, it doesn't necessarily mean continuous excitement, or even doing much outside of a computer. Advocates of the Singularity don't necessarily get to pile into a large and expensive building with hundreds of rooms and listen to professors; we don't necessarily get impressive anecdotes to boast to our friends with, and for every person that agrees with what we say, there will be hundreds who disagree with us fervently. This is the reality of Singularity advocacy - it takes plenty of internal drive, the ability to create your own inspiration and motivation, the patience to defer instant gratification, and the enthusiasm to acquire the knowledge necessary to have any competence whatsoever in what you are doing.

For those that have these rare qualities, I'm as anxious as ever to get in touch and start speaking with you. The typical First World citizen will have between 40 and 80 free hours per week; and most of this will be spent watching television or movies. Do you have five or ten hours per week to invest in the goal of the Singularity, even if those five or ten hours sometimes turn out to be superficially tedious or boring?

I'm proposing creating a circle of activists that can work (i.e., write) together, keep tabs on one another, determine the best methods of contributing to the Singularity, and actually exploit these methods. Motivation, communication skills, and a minimum level of writing and reading ability are a must. You should have already read "Creating Friendly AI", "Levels of Organization in General Intelligence", "Advice for Singularity Writers", and "The Singularitarian Principles" at least once, and be familiar with the work of prominent figures in the Singularity analysis and advocacy communities - John Smart, Ray Kurzweil, and Eliezer Yudkowsky. ("Familiar" will mean you've read and understood most of their online work.) You should have some experience with arguing Singularity issues clearly and convincingly, on mailing lists or in person. Knowledge in scientific or philosophical issues related to the Singularity would be a big plus, the more the better.

This group will advocate achieving the Singularity through Friendly Artificial Intelligence, and have close ties with the Singularity Institute - currently the only group explicitly working on the Singularity. We're open to hearing arguments for alternative approaches to accelerating the Singularity or safeguarding its integrity, but a minimum level of agreement on the specific importance of Friendly AI to the Singularity is essential.

We haven't decided on a name yet, but a modest website is in the works, which will be hosted on acceleratingfuture.com. We are seeking a dedicated domain and servers. Please drop me a line if you can supply them for low cost or for free.

We're already working on several exciting projects in Singularity analysis and promotion, and are eager to get more progress done. Among these projects are 1) researching the relevant variables in Singularity forecasting, such as minimum seed AI complexity, neurocomputational evidence, and likely funding levels, 2) creating lists of online communities and individuals who would potentially be receptive to Singularity-oriented ideas, 3) reaching out to those people, 4) writing introductory material to address recursive self-improvement, Friendliness theory, and the rationales behind Singularity advocacy, and more.

So, would you like to join us? If so, give me an email, along with a 2000-words-or-more piece of your writing on the Singularity topic. Of course, I'd be interested in hearing a little bit about you and your background as well.

Sincerely,
Michael Anissimov