The Singularity is Nearer: When We Merge with AI

by Ray Kurzweil

Singularity will happen around 2045. At the time that book was published, that date lay forty years—two full generations—in the future.

As a result, as I write this, one dollar buys about 11,200 times as much computing power, adjusting for inflation, as it did when The Singularity Is Near hit shelves.

In my 1999 book The Age of Spiritual Machines, I predicted that a Turing test—wherein an AI can communicate by text indistinguishably from a human—would be passed by 2029. I repeated that in 2005’s The Singularity Is Near. Passing a valid Turing test means that an AI has mastered language and commonsense reasoning as possessed by humans.

Since most media (both traditional news media and social media) make their money by attracting eyeballs to generate ad revenue, we shouldn’t be surprised that the industry has learned, collectively, that the best way to stay in business is to propagate threatening information that provokes strong emotional responses.

THE REALITY IS THAT NEARLY EVERY ASPECT OF LIFE IS GETTING PROGRESSIVELY BETTER AS A RESULT OF EXPONENTIALLY IMPROVING TECHNOLOGY

This is clearly true about quite a few parts of life, but "nearly every" is pushing it, and screaming it in all caps doesn't make it any more true.

Information technology advances exponentially because it directly contributes to its own further innovation.

Economic cycles go up and down; wealth can be gained and lost. But technological change is essentially permanent. Once our civilization learns how to do something useful, we generally keep that knowledge and build on it.

In effect we are in the process of turning medicine into an information technology, harnessing the exponential progress that characterizes these technologies to master the software of biology.

When Konrad Zuse built the first working programmable computer, the Z2, in 1939, it could perform around 0.0000065 computations per second per 2023 dollar.201 In 1965, the PDP-8 managed around 1.8 computations per second per dollar. When my book The Age of Intelligent Machines was published in 1990, the MT 486DX could achieve about 1,700. When The Age of Spiritual Machines appeared nine years later, Pentium III CPUs were up to 800,000. And when The Singularity Is Near debuted in 2005, some Pentium 4s were at 12 million. As this book goes to print in early 2024, Google Cloud TPU v5e chips are likely around 130 billion operations per second per dollar!

LaGarde asked me why we don’t see more evidence of economic growth from all the remarkable digital technology that is now available. My answer was (and is) that we factor out this growth by putting it in both the numerator and denominator. When a teenager in Africa spends $50 on a smartphone, it counts as $50 of economic activity, despite the fact that this purchase is equivalent to over a billion dollars of computation and communication technology circa 1965,

traditional metrics almost completely ignore the steep deflation rate for information technology, which in the case of computation, genetic sequencing, and many other areas is on the order of 50 percent per year.

In many ways, just as important as computation power is information. As a teenager I saved up for years from the earnings of my paper route to buy a set of the Encyclopedia Britannica for several thousand dollars, and that counted for thousands of dollars of GDP. By contrast, a teenager today with a smartphone has access to a vastly superior encyclopedia

Eric Drexler’s estimate in his 2013 book Radical Abundance is that, taking into account more mass-efficient nanomaterials, atomically precise manufacturing could build most kinds of objects for the equivalent of about twenty cents per kilogram.

Solar also has a much faster doubling time as a proportion of the world total than other renewable sources: on average, just under every twenty-eight months from 1983 to 2021—even as the total amount of electricity generation has increased by around 220 percent in that same span in absolute terms.234 From 3.6 percent in 2021, it would take only about 4.8 doublings to reach 100 percent, which would put us at 2032 to meet all of our energy needs from solar alone.

physicist Richard Feynman (1918–1988) in his seminal 1959 lecture “There’s Plenty of Room at the Bottom,” in which he described the inevitability of creating machines at the scale of individual atoms, as well as the profound implications of doing so.

In the mid-1980s, engineer K. Eric Drexler founded the modern field of nanotechnology, building upon this concept from von Neumann.

Their detailed design (again conceptual) provides for about 1020 logic gates per liter and would operate at 100 MHz resulting in up to 1028 computing operations per second per liter of computer volume (though heat dissipation would require that this volume have high surface area).60 The amount of power expended by this design would be on the order of one hundred watts.61 Since there are around eight billion people in the world, emulating the brain computation of all human beings together would thus take fewer than 1024 operations per second (1014 per person times 1010 people).

In 2001, American physicist and chemist Richard Smalley began a public debate with Eric Drexler over whether atomically precise manufacturing using “molecular assemblers” will ever be possible.

The true value of products, then, would lie in the information they contain—in essence, all the innovation that has gone into them, from creative ideas to lines of software code that control their manufacture. This has already taken place for goods that can be digitized.

Indeed, we are already going in that direction, as the “information content” of most products is rapidly increasing—and will ultimately get very close to 100 percent of their value.

This is why there is sound logic behind Aubrey de Grey’s sensational declaration that the first person to live to 1,000 years has likely already been born.

Fortunately, the speed of virus sequencing is following a long-term trend of acceleration. It took thirteen years after its discovery to sequence full-length genome of HIV in 1996, and only thirty-one days to sequence the SARS virus in 2003, and we can now sequence many biological viruses in a single day.

We should thus work toward a world where the powers of AI are broadly distributed, so that its effects reflect the values of humanity as a whole.

By the middle of the 2020s we will have a means of interacting with a computer that is thousands of times faster than keyboarding: fully immersive virtual reality with full-screen video and audio.

The middle of the 2020ies is now. I guess Kurzweil might claim that for example current games fulfills this prediction, but if so he ALSO has to explain in which ways they do -- but didn't a decade ago. Fundamentally new interaction-paradigms in this space are scarce although of course there's been incremental progress.