Thursday, March 19, 2009

A Conversation with Ray Kurzweil

Darlene Cavalier and Bart Leahy

The questions below were submitted by our readers. Some questions were edited or combined for content. We appreciate your participation in the ongoing discussion about the Singularity. (Note: Hyperlinks added by Bart for reference.) Our thanks to Ray Kurzweil for taking the time to respond to our queries.

General and Technical Questions – What Is the Singularity?

1. What is your “short version” definition of what the Singularity is?

RAY: The Singularity is a future time when the pace of technological change will be so fast and transformative that you will not be able to follow it unless you merge with the intelligent technology we are creating.

2. Singularity University is clearly aimed at helping to shape the Singularity and hasten its arrival. Do exponential trends really need help and, if so, can we really expect to shape them?

RAY: The exponential growth of information technologies will continue inexorably as it has for over a century. However, technology has always been a double-edged sword ever since we developed fire and stone tools. How we apply these technologies and whether constructive applications that overcome human suffering and extend our creativity predominate over destructive applications is not preordained. That is where we can help shape the Singularity.

3. How would we recognize the Singularity happening? Would the change be gradual, or would it be similar to the singularity around a black hole--so enormous, rapid, and widespread that we wouldn’t even realize what had happened until long after we’ve crossed the point of no return?

RAY: The pace of information technology is continual yet exponential. Exponential trajectories have no discontinuities but are nonetheless disruptive. You describe the changes well as “enormous, rapid, widespread,” and constituting a “point of no return.”

4. Given the slow and erratic progress in
AI over the past 40 years, what makes you so confident that machines will become intelligent (in the commonly understood sense) in the next 40?

RAY: I disagree with your characterization of AI. It reminds me of people who go into the rain forest and ask, “where all the species that are supposed to be here” when there are 25 species of ants within fifty feet of them. The species in the rain forest are not seen because they are hidden in the ecostructure. Similarly, AI is all around us yet hidden in our modern economic infrastructure. Every time you send an email or connect a cell phone call intelligent algorithms route the information. Pick up a product, it’s been designed at least in part by intelligent computer assisted design software, inventory levels controlled by intelligent just-in-time inventory systems, assembled in robotic factories. AI software flies and lands airplanes, guides intelligent weapons system, automatically detects credit card fraud, helps you find information on the web, diagnoses electrocardiograms and blood cell images comparable to trained physicians, and much else. If all the AI programs were to stop tomorrow, our modern infrastructure would grind to a halt. That was not the case just fifteen years ago. These were all research projects then. On the research front, AI programs can now play master levels of go, drive cars with no human drivers through complex terrains, recognize songs and images, and so on. Now that we can see inside the brain with very high resolution we are building models and simulations of brain regions. That will accelerate AI in the years ahead.

Ethical and Social Questions

5. Why do you think it’s a good idea for us to create machines that are smarter or more powerful than human beings? Can we expect singularity to shape itself or will we still hold charge of our technological creations?

RAY: The machines are not an alien invasion from Mars. It is part of our civilization which is already a human-machine civilization. Ever since we picked up a stick to reach a higher branch, our tools have been extensions of ourselves. We are the only species that changes who we are based on tools we create.

6. Suppose that things continue in much the way they are now, with increasingly powerful and miniaturized wireless devices making information available wherever we want it. Does that count as a “Singularity?” It is easy for me to imagine, for instance, a brain implant that allows me to conduct Google searches purely by the power of thought—but that merging of biological and digital intelligence seems distinctly different from what you mean by a Singularity.

RAY: Accessing the web from inside our brains is one good example of what we will see in about twenty years.. The machine extensions to our brains will grow exponentially both in hardware and software capability. By the late 2030s, it will be the nonbiological portion of our intelligence that predominates.

7. How do you see the economics of the future working/changing if everything is free?

RAY: Who said everything will be free? We will continue to have open source and proprietary sources of information. When we have
nano desktop factories that can produce physical products from information files and very inexpensive input materials, we will be able to live very well on just open source information. But there will still be an edge and demand for proprietary information. Information technologies have had an 18 percent growth rate as measured in constant dollars for the past fifty years despite the fact that you can get twice as much of it for the same cost every year. This in fact has been the source of true economic growth.

8. How close are we to getting a Star Trek-like “

RAY: In twenty years we will be able to produce physical products from information files using nano desktop factories. This is still short of the holodeck but it will turn virtually the entire economy into an information economy.

9. How far away are we from reaching the Singularity? Is 2045 still a reasonable estimate?

RAY: According to my models, we will multiply our biological intelligence a billion fold through its integration with nonbiological intelligence by 2045. I consider that the Singularity.

10. How do you respond to people who claim that this is just “the geek Rapture?”

RAY: This “criticism” is based on the notion that the “rapture” came first and that we just worked backwards to justify this religious notion. But that is not where the ideas come from. They come from a scientific analysis of technology trends. Religion emerged in pre scientific times and we do need to update our philosophies based on science.

11. What is to become of people who don't want to join your Singularity and just want to remain human as they are?

RAY: First of all, it is human to change who we are. We didn’t stay on the ground, we didn’t stay on the planet, and we have not stayed with the limitations of our biology. Human life expectancy was 23 a thousand years ago. We are the only species that changes who we are and extends our reach, both physical and mental, through our tools. So it is human to change who we are. There will always be early and late adopters, but people are not going to completely dismiss these changes. How many people today complete reject medical and health technologies? When there is a therapy based on blood cells devices that overcome a particular disease, very few if any people will reject it. People put computers in their brains today if they have Parkinson’s Disease. People do not reject this FDA approved therapy due to philosophical issues.

12. Are you pleased or disappointed with the progress made so far? What technologies are “ahead of the curve,” as you see it, which ones are behind?

RAY: My team and I just updated the graphs that were in my 2005 book
The Singularity is Near from 2002 through 2007. The exponential curves have remained precisely on track. It is pretty remarkable when you consider that what we are measuring is the innovation of millions of people.

13. Your book takes a very optimistic view of science and human nature, but neglects the problem of human evil. Given what we've done with previous “great inventions,” don't you worry that individuals or nations could do great harm with all this—accidentally or maliciously? What safeguards would be put in place to prevent this?

RAY: I don’t know why people say I ignore the downsides when I was the one who initiated the debate about promise versus peril. Bill Joy’s WIRED cover story “
Why the Future Doesn’t Need Us” was based on my book The Age of Spiritual Machines, as he acknowledges at the beginning of the article. My recent book, The Singularity is Near, has an extensive discussion of the downsides and what to do about them in chapter 8 (“The Deeply Intertwined Promise versus Peril of GNR”). The safeguards we need are twofold. We need ethical standards for responsible practitioners such as the Asilomar guidelines in biotechnology. And we need a rapid response system, basically a technological immune system, to deal with intentional abuse. We have such a system for software viruses. I am working with the U.S. Army on putting a rapid response system in place for bioengineered biological viruses.

14. What is/will be the relationship between ethics and The Singularity? The rapid growth of science/knowledge leads to many advancements via engineering, but how can/will ethics be applied when mankind can no longer keep pace?

RAY: See my response to the previous question.

15. What are your foundational core values....specifically your belief about God and how that guides the limits of what you do?

RAY: I believe that evolution is a spiritual process in that it leads to greater intelligence, creativity, beauty, and love, all of the attributes that God has been called without limit. I believe we have a responsibility to apply our ideas to overcome human suffering. We have made good progress on this. Just read Thomas Hobbes on what human life was like a few hundred years ago. He described it as short, brutish, disaster prone, disease and poverty filled. Human life expectancy was 37 just 200 years ago.

16. Space exploration is one technology you downplay in The Singularity is Near. Does your new venture with
Peter Diamandis and NASA—the Singularity University—change this perspective?

RAY: Space travel will be of key importance once we saturate the matter and energy in our midst at the physical limits of computation. We will then need to spread out to the rest of the galaxy and universe. But we will not sending missions of squishy creatures, but rather missions of nanobots swarms..

17. What is the purpose of “Singularity U?”

RAY: The purpose of Singularity University is to bring together the most creative students and professors to study exponentially growing information technologies and to apply these ideas to meeting the grand challenges of humanity.

18. What should people do about scientific literacy so that everyone can understand, at least a basic level, the rapidly advancing technology?

RAY: Indeed, scientific literacy needs to be a core goal of our educational system. Other countries are taking that more seriously than we are. About twenty years ago, the U.S. graduated about 60,000 engineers per year and China graduated about 10,000. Now, we graduate about the same level and China graduates about 300,000 per year.

19. What should we be *doing* about all this?

RAY: Celebrate science and engineering as the cool subjects that they are. It is only the exponentially growing information technologies that have the scale to address the major problems of humanity.

20. In 1999, you created a hedge fund called "FatKat" (
Financial Accelerating Transactions from Kurzweil Adaptive Technologies) which began trading in 2006 to recognize patterns in "currency fluctuations and stock-ownership trends" and eventually beat the best human financial minds at making profitable investment decisions. Did FatKat predict the market collapse?

RAY: The FatKat algorithms are designed to only predict a few hours or days ahead and to do that with an accuracy that is somewhat better than chance.

21. Between the
Reading Machine and the newest pocket-sized device designed to aid blind people by reading written text aloud, you’ve demonstrated a remarkable desire to help the blind. I’m curious: what sparked your interest in helping the blind?

RAY: In the mid 1970s I had developed a method that could recognize print in any typestyle. It was a solution in search of a problem. I happened to sit next to a blind guy on an airplane who said that his only real handicap was the inability to read ordinary print. That sparked my desire to apply this technology to build a print-to-speech reading machine..

22. You are making a movie due for release this year called
The Singularity is Near: A True Story About the Future, part fiction, part non-fiction, in which you interview 20 big thinkers like our friend Marvin Minsky. I assume Marvin shares your vision on what Singularity is and will be. Do most “futurists” share your vision? Why or why not?

RAY: There is increasing awareness of my exponential view, but linear thinking is actually hard wired in the brain. So even otherwise sophisticated scientists often project current trends linearly into the future. They just have not studied technology trends. There is a profound difference between the intuitive linear perspective and the historically accurate exponential view. If I take thirty steps linearly (1, 2, 3, 4, 5, …) I get to 30. If I take thirty steps exponentially (2, 4, 8, 16, 32, …) I get to a billion. The latter sequence describes what has already happened. When I was a student at MIT we all shared one computer. The computer in your cell phone today is a million times cheaper than the one we all shared when I was a student, and is a thousand times more powerful. That’s a billion fold increase in price performance since I was a student and we will do it again in the next 25 years. This applies not just to computation but to any technology where we can measure the underlying information properties such as bits moved around on the Internet, genetic sequencing, brain sequencing, and much else.

23. The “part non-fiction” subplot in your movie includes a computer that saves the world from self-replicating, tiny robots. Are you concerned that such microscopic robots will pose a threat to the world?

RAY: Yes, that is called the grey goo scenario, and the narrative thread in the movie illustrates this danger.. I do think we can manage that through a combination of ethical standards to build in safeguards into nanotechnology, as well as a rapid response system that detects threats and immediately deals with them, just like our biological immune system is designed to do. But this is not something we should be sanguine about. We need to be very diligent about it.

24. I understand it had been documented that one of your goals is to bring back your late father using AI. How can that happen?

RAY: Future AI’s will be intelligent to gather all of the information about a deceased person (his DNA from his gravesite, memories of people who knew him, archived records) and create a person (say a virtual person in a realistic virtual reality environment) very similar to that person, basically someone indistinguishable from that person to the people who knew him or her. For this reason I have kept about fifty boxes of my father’s archives, all of his music, letters, photographs, movies, etc. Would this person be the same person as my father, or a new person that just happens to be very similar? You can argue that if my father lived, he would be very different anyway. We change our particles every six months or less, but there is a continuity of pattern. I discuss this philosophical issue in chapter 7 of The Singularity is Near.

25. Lastly, you were clearly influenced by your parents’ and uncle’s careers. Are your children working in science/engineering fields?

RAY My son Ethan, age 29, works as a venture capitalist for Bessemer Ventures in Silicon valley in the area of high tech business. I often talk to him about my business strategies. He is not an inventor but he is fostering technology innovation. My daughter is a writer and artist and is writing a graphic novel as her senior project at Stanford. Interestingly, I majored in both computer science and creative writing at MIT.

1 comment:

lin said...

Establishing ethical criteria to prevent or minimize misuse of the products and byproducts of the Singularity is laudable but potentially flawed in ways that could be fatal for humanity. One weak point of ethical constraints on human wrongdoing is that ethical behavior is voluntary and self-imposed.

Self-restraint and self-sacrifice for the common good are lofty goals that lack effective enforcement. They will not prevent the unscrupulous from holding humanity hostage with some technological innovation. Development of nuclear weapons by Iran and North Korea come to mind.

The exponential growth of information used for good may be eclipsed by that used for evil purposes. As humans, we have a long history of setting aside ethical restraint in favor of political and economic gain. Nanotechnology could be used to contaminate the blood, food, or water supply of much of the planet with self-replicating and self-aware virus-like contaminants.

A need for immediately effective countermeasures to save billions of people could produce solutions that eventually become worse than the original problem. In this case, the results are potentially catastrophic for the entire planet. What will keep us from traveling so far down the slippery slope of expedience that we cannot return from it?