Ask what changed fastest in your lifetime and most people point at the phone in their pocket. Smartphones, as we know them, did not exist a decade before Singularity Hub published its 2016 look at the idea behind this. Go back three decades further and nobody owned a computer. The first personal machines were only about 40 years old then, and two-thirds of Americans, according to a Pew report, already carried one in their hands.
Futurist Ray Kurzweil says that feeling of speed is not an illusion. In The Singularity Is Near he calls the cause the law of accelerating returns.
Moore’s Law is only one example
Chips keep getting stronger and cheaper because the number of transistors on a single chip has doubled again and again for five decades. That doubling is Moore’s Law, and it is why a phone can hold so much power at an affordable price.
Kurzweil’s point is that chips are not special. Data storage follows a similar curve. So does the recording of DNA base pairs. In each case, capability has multiplied by thousands, millions or billions while cost fell, and these technologies do not depend on transistor counts at all.
Why the curve keeps steepening
Kurzweil starts with biology. Genes are the blueprints of useful tools, and selective pressure passes the good ones to the next generation. Over geological time, living things grew more capable by building on earlier progress rather than starting from nothing. “Evolution applies positive feedback,” Kurzweil writes. “The more capable methods resulting from one stage of evolutionary progress are used to create the next stage.”
Technology works the same way, only faster. Cells, bones, eyes, thumbs and brains came first, and from thumbs and brains came tools.
Picture a chair built with hand tools, then power tools, then an assembly line. Each step speeds up production. Now suppose each set of tools is also used to build the next, better set. Kurzweil says that is what happens in chip design. In his words, “the first computers were designed on paper and assembled by hand.” Today they are designed on workstations, with the computers themselves working out many details of the next generation, and then built in largely automated factories.
Success also draws money. When a technology works, it attracts bigger research budgets and top recruits, and those resources push it further. Kurzweil describes a second level of exponential growth here, in which the rate of growth itself speeds up.
No single paradigm lasts forever, he adds. Integrated circuits will grow until they run out of room, and then a new paradigm takes over.
What Kurzweil predicts
Kurzweil wrote in 2001 that the overall rate of progress doubles every decade. His conclusion: “We won’t experience 100 years of progress in the 21st century, it will be more like 20,000 years of progress (at today’s rate).”
He has a record to point to. In 1990 he predicted a computer would beat a professional chess player by 1998. Garry Kasparov lost to IBM’s Deep Blue in 1997. The 2016 article added that a computer had since mastered Go, a result some experts did not expect for another decade.
Genomics shows the same pattern. The first human genome was sequenced in 2004 at a cost of hundreds of millions of dollars. By 2016, machines could sequence 18,000 genomes a year at $1,000 each.
