There are multiple potential paths to a technological singularity. Although many people think a singularity would result if computers acquired sentience and general intelligence, a singularity might also happen if methods were found to increase the IQ of individuals and populations. In Singularity Rising: Surviving and Thriving in a Smarter, Richer, and More Dangerous World , the economist James Miller argues that a technological singularity is highly probable in the 21st century. The cognitive performance of humans can be enhanced by drugs, environmental changes, neuro-feedback, and even implanting electronic devices in or near the brain. In an interview with Sander Olson for Next Big Future, Dr. Miller describes a future world in which parents use genetic enhancing technologies to increase the capabilities of their offspring, the best ways to bring about a friendly AI, and the signs that could appear indicating that a singularity is imminent.
Question: You recently attended at the Singularity Summit 2012 conference. What did you take away from that conference?
The participants were a high-powered group. I’ve been attending since 2007, and I’ve noticed that there is quite a bit more acceptance of these concepts than even five years ago. I also noticed that there were more potential investors there than in previous years. We are making steady AGI progress, and interest in this field is clearly increasing.
Question: In Singularity Rising, you discuss multiple possible paths towards a technological singularity. What do you think is the most likely scenario?
That’s a tough question I avoid addressing in the book in favor of my thesis that there are so many plausible paths to the Singularity and such tremendous economic and military incentives to achieving both AGI and human intelligence enhancements that the odds of a singularity are high. Plus, making progress on one path, such as boosting human intelligence with smart drugs, makes it easier to walk (or stumble) down other paths such as creating an AGI intelligence explosion.
Question: Will there be any signs that a singularity is imminent?
I would look for signs in the stock market. People might stop funding construction of real estate, or people might stop investing in their retirement. If people realize that fundamental, transformative changes are coming it will completely change the manner in which they invest. Or if we see developed nations such as the U.S. experiencing 10% annual growth rates in GDP, that would be another sign.
Question: But the collapse of the Soviet Union and the rise of the internet were not preceded by any such signs.
There are more factors pushing us towards a singularity, and these factors will start to become obvious. A few rich, well-informed individuals set the tone for the markets, and these individuals will grok that transformative changes are on the horizon.
Question: So you see a technological singularity as virtually inevitable?
I’m an economist, and economists are never supposed to use the word inevitable. The runup to the singularity will be quite dangerous, and we may destroy ourselves beforehand. But assuming that scenario doesn’t come to pass, I see a 90% probability of a singularity within a hundred years and a significant chance of one within fifty.
Question: You base your predictions largely on Moore’s law. If Moore’s law soon ends, as some are predicting, how would that impact your views on a coming singularity?
If Moore’s law ends that would reduce the likelihood of a singularity. Although increasingly fast chips are clearly desirable, the most important factor is the decline in the cost of computing resources. From all my research, the cost of computing should continue to decline for a long time, even if no breakthroughs in post-CMOS silicon technology emerge. But there are other paths to a singularity that do not depend on Moore’s law, so I see a singularity as likely even if Moore’s law ends.
Question: Peter Thiel has argued that technological progress may actually be slowing.
I hold Thiel in the highest regard, and consider him an authority on these matters. But I think that the trends Thiel discusses will reverse themselves. There are drugs in development that could potentially raise IQ, and that alone could collectively increase global intelligence.
Question: You write extensively about the potential of certain drugs, such as adderall, to improve memory, and concentration. Are drugs the best way to improve cognitive abilities?
In the shorter term, we could use neuro-feedback techniques. The patient looks at a television screen, and the application trains your brain to produce different frequencies desired. This approach can improve flexibility, and can temporarily enhance certain senses. When I used it, I noticed that my sense of sight was somewhat improved.
Question: How exactly does neuro-feedback work?
Neuro-feedback works by rewarding your brain for producing or inhibiting selected frequencies. So it could encourage your brain to produce certain alpha-waves, such as those that induce relaxation or concentration. You put on a skullcap, and watch a screen. I tried this technique and was able to improve my vision. Things became sharper and more beautiful. This technique might be used in as little as a decade to improve the learning capabilities of most students.
Question: Ray Kurzweil has argued that IBM’s Watson computer beating the world’s best Jeopardy champions represents a major AGI accomplishment. Do you agree?
Being an economist, I take an economic perspective. If all Watson could do is what it did on Jeopardy, then that does not constitute a real AGI advance. But I am intrigued to see what effect Watson has on the marketplace, and in fields such as health care. Watson could end up being a transformative technology even if it is not a true AGI advance.
Question: Could technologies such as genomics have the potential to substantively affect the AGI field?
Genomics does have the ultimate potential to use genetics to create more intelligent individuals. This won’t become commonplace within the next twenty years, but it will happen within the 21st century. It will likely eventually be feasible to combine many of the positive and beneficial genes in an embryo or at least to create an extraordinarily fit human who has very few deleterious genes. I predict that within 15 years, it is more likely than not that 10% of new parents in rich countries will use genetic enhancing technologies to improve the expected “quality” of their children.
Question: Do you think that implanting electronics in the brain is a viable path towards increasing intelligence?
The cost of making the technology safe for the first few individuals will be quite high, but once the technology is perfected the cost will plummet. And if Moore’s law keeps going, the intelligence of these implanted-individuals would continually increase. If this happens, that would be a clear and irrefutable sign of a singularity, at least if we don’t first destroy ourselves.
Question: If you were a billionaire, how would you spend your money?
I would spend my money specifically on promoting friendly AGI. Creating friendly AGI is even more difficult than creating AGI, but could make the difference between creating an artificial intelligence that values humans and one that doesn’t.
Question: Do you foresee any technological breakthroughs within 20 years?
Within twenty years there will be clear signs that a singularity is coming. AI will increasingly permeate society, and it won’t be as limited or as brittle as it is today. But I also think that there will be substantial reductions in poverty in the developing world during the next two decades. There is a clear correlation between prosperity and IQ rates, since richer people are generally better fed, better educated, and healthier than poor ones. So I see the collective IQ of our planet measurably increasing every decade during the 21st century.