Ray Kurzwell revisits AI singularity prediction 20 years on | Technology | sfexaminer.com – San Francisco Examiner

Left: Kurzweil with Stevie Wonder, as advisor, shown here in the 1980s. Kurzweil Music Systems made the Kurzweil Keyboard K250, the first music synthesizer capable of re-creating the grand piano and other orchestral instruments.

Preparing for my interview with author, inventor and futurist Ray Kurzweil, I did something Ive never done before. I opened the ChatGPT app on my phone and began to type: What should I ask Ray Kurzweil when I interview him?

Within seconds, eight questions appeared on my screen. What normally takes a couple of hours of preparation was done in a nanosecond. I must admit, it kind of felt like cheating.

Now in his 70s, Kurzweil is upping the ante in his newest book, The Singularity Is Nearer: When We Merge with AI,revisiting his prediction of the melding of human and machine, with 20 additional years of data showing the exponential rate of technological advancement. Its a fascinating look at the future and the hope for a better world.

If I had a magic wand, I would accelerate our ability to add more time to our lives.

The biggest risk Ive ever taken is starting a company. Your idea could be very good, it could lead to great things, but if you cant pay your bills, and you cant pay your rent, and you cant pay peoples salary, the whole thing will go away.

Im happiest when well, Freud said were motivated by work and play. And work has definitely been my contribution to coming out with acceleration of knowledge. In terms of play, its really just being with people I love. With my wife, my kids, my grandkids, and so on.

The biggest regret I have is I cant really think of anything that I regret. If I had something, some idea, Ive generally pursued it.

Kurzweil has long been recognized as a great thinker. The son of a musician father and visual artist mother, he grew up in New York City and at a young age became enamored with computers, writing his first computer program at the age of 15.

While at MIT, earning a degree in computer science and literature, Kurzweil started a company that created a computer program to match high school students with colleges. In the ensuing years, he went on to found (and sell) multiple technology-fueled companies and inventions, including the first reading machine for the blind and the first music synthesizer capable of recreating the grand piano and other orchestral instruments (inspired by meeting Stevie Wonder). He has authored 11 books.

On a recent afternoon, I met with Kurzweil at the Merchants Exchange Building in downtown San Francisco. Wearing his signature hand-painted suspenders, he sat down with me to talk about large language models, the singularity and the future of humanity with AI.

Ray, for many of us laypeople, artificial intelligence is a fairly new concept, but not for you. Ive actually been in artificial intelligence for 61 years, and thats a record. Nobody else has that longevity. When I started, no one had heard of artificial intelligence. Artificial intelligence? Whats that? Now everybodys heard of it, and people are very nervous about it.

Yes, they are. But theyre also very intrigued. This added intelligence its really coming from people, and its going to make us smarter. People are nervous about that, but thats really the goal of humanity. Were trying to increase our intelligence. If you were to ask a mouse, Would you like to become as intelligent as a human? Well, they wouldnt even understand the question, and theyd have no idea how to answer. But we are intelligent humans, and were glad about that. If we make ourselves even smarter, thatll be even more beneficial, and thats what this is all about. And its going very, very quickly at an exponential rate a 20 quadrillionfold increase over 80 years.

I cant even fathom what that number means! So, how did you get interested in artificial intelligence? It seemed pretty amazing to me that we could actually replicate the human brain. My father was a musician, so the very first project I did was to write music using a computer, and you could feed in particular melodies. It would try to replicate the kind of melodies that Mozart created. It was not quite as good, but you could recognize it as Mozart or as Chopin. And I worked with my father on it and showed it on [the game show] Ive Got a Secret.

Below: A born inventor, Kurzweil at age 6 with his father, a conductor, composer and musician, and his mother, a visual artist, in New York City.

That was very exciting to have a computer compose music. No one had ever done that before. Now automatic composition is a major field, but that was 61 years ago. Ive always been very excited about artificial intelligence because thats the secret of humanity, the fact that we have intelligence. If we can recreate that which were getting very close to within a few years, well be able to emulate the intelligence that any human being can have.

Thats unbelievable. Large language models [computer programs that generate human-like text] are actually better than humans. You could ask it anything, and it will answer you with an intelligent prompt. And if you dont like that [answer], you could ask it again, and youll get something else that means something similar. It can do that for any topic. No human being can do that. Einstein knew something about physics, but he didnt know philosophy and so on.

In what areas do you see the most promising benefits of AI, and by areas I mean education, health care, climate change Well, health care I think is the most significant because I think well reach longevity and escape velocity in five years, by 2029. Right now you go through a year, you use up to a year of your longevity. But we also have scientific progress, which is moving ahead. So you actually get back some of that loss. Right now youre getting back about four months. However, scientific progress is progressing exponentially. By 2029, youre going to get back a full year. So you lose a year, but youre going to get back a full year from scientific progress. Past 2029 youll get back more than a year. As far as longevity is concerned, youll actually go backwards in time.

One of your boldest predictions is that we as humans may one day live forever. Do you still believe that? Theres a little bit of a misquote [with laughter].

City leaders grapple with how to bring street vendors back to Mission Street while continuing a clampdown on illegal fencing operations

UCSF infectious disease experts weigh in on the impacts of climate changetemperatures and other events change disease patterns

Ask Dennis Hancock what its like to work in youth-dominated Silicon Valley, and he will quote Will Rogers

OK, its good to have you here to get it right. So its not a guarantee, but your longevity is actually going to increase, not decrease. And as we go further, well be able to overcome more and more possibilities.

For example, were going to combine with large language models. Its going to become part of us, and its going to increase our intelligence. So thats part of our intelligence that we can get back because we back up everything thats computational. The part of our intelligence that comes from machines is going to be backed up. So if that gets destroyed, you can recreate it. Its like you lose your phone, but you can get it back because its all backed up. But the intelligence thats real, that is part of our body. If you lose that, you cant get it back. We will be able to do that. Well be able to access stuff thats going on in our real brain and back that up, just like the computational part.

In 2015, record producer Mike Clink presented Kurzweil with a Grammy Award for outstanding achievement in music technology.

So the coexistence of man and machine, is that what singularity is? Yeah. Singularity comes from physics. If you have a black hole, you cant actually look inside the black hole. Nothing can get out because gravity keeps it inside. Thats really what singularity means that we really cant access it, we cant back it up, and we cant really fully understand it because we cant access it. So thats what I mean here. Were going to merge with computation, and were going to become so intelligent we really cant describe what its like. ... We will understand it when we get there, but right now we cant really predict it.

When you say that, Im thinking of a chip in our brain. Well, we will have chips in our brain, but we wont be able to describe what its going to be like past the singularity. Thats very hard to imagine. But we can say some things. Our longevity will be way past what it is today. We really can go on forever, and were not actually going to get older. Well continue to look the way we do now.

Youve talked about some of the benefits of AI. What do you fear about artificial intelligence? Well, what do we fear now about humans even before AI? When I was a small child, they invented atomic weapons. We exploded two nuclear weapons. If you ask people, We just exploded two weapons within one week in anger; whats the likelihood thats going to happen again? People say, Well, absolutely thats going to happen. And here we are 80 years later. The people who have atomic weapons are not the nicest people in the world, and yet it hasnt happened.

But human beings, even without AI, can cause a lot of destruction. Thats actually getting better as weve gotten more intelligent. We have two wars in the world now that people are paying attention to, and thousands of people are dying, and thats terrible, but not as bad as it used to be. Eighty years ago, we had 100 million people die in Europe and Asia through World War II. But intelligence can make us what we would consider evil and cause a tremendous amount of destruction. And thats just from natural intelligence. If we can make ourselves more intelligent with AI, people are afraid of that.

They also see it as not part of humanity. Right now, AI can, for example, do coding. Not quite as good as humans, but within five years itll be as good as humans. People are afraid of that because theyve got a job doing that, and AI can do that instead.

A Price-Performance of Computation chart by Kurzweil, who began tracking the pace of technological progress 40 years ago and discovered that it was advancing at an exponential rate. He has used this data to make many accurate predictions, including his 1999 prediction that computers would reach human level intelligence by 2029. While controversial at the time, AI is now right on track to hit this milestone by 2029, if not sooner.

Speaking of jobs, one of the fears of AI is that machines will replace jobs and even further widen the inequality gap. What do you think? It hasnt so far. The Luddites came along 200 years ago because they created machines that could create clothing and so on, and people say, Well, look at these machines, and theres more coming out every day, and all jobs are going to go away. But it didnt happen. We actually have more jobs today than we did then, despite the fact that machines can do jobs.

This is a revolution of sorts, like the Industrial Revolution. But there was a period after the Industrial Revolution that was pretty awful with child labor, dangerous workplaces, environmental harm. But it was worse before that. If you go back, say 300 years, life was horrible. Life expectancy was like 28. You go back 100 years compared to today, we make 10 times as much in constant dollars as we did 100 years ago. And of course, we didnt even have things like electronics, for example. So despite the fact that we get rid of jobs, were actually in better shape, and thats going to continue. Large language models will create the ability of people to get enough money to live, and well be doing things that we enjoy doing compared to today.

So you dont anticipate a dark period during this transition? There may be some difficulty if you have a certain job, and then they can use large language models instead of yourself; you may be unhappy about that. But eventually well be far better off than we are today. Thats been true of every type of automation.

Lets talk about your new book,The Singularity Is Nearer, the sequel to 2005sThe Singularity Is Near, and which comes out this summer. In 1999, I wrote The Age of Spiritual Machines, and I predicted that the Turing test, which is where a computer can emulate a human being, would happen within 30 years, by 2029. Stanford was so alarmed, they held an international conference, and they had hundreds of people come from all over the world who were experts in AI. They agreed with me that this would happen, but not in 30 years. They felt it would take 100 years. And right now we have large language models. Not quite the Turing test, but very, very close. A lot of people like Elon Musk think its going to happen next year. Im saying five years. So one large language model will know everything that any human being knows in five years, and people are now accepting that. This was a major controversy back in 1999. Now its just about to happen, so I wrote this book telling what life will be like. And it basically will be better. But there are also some problems because human beings can create problems. If we can add to our intelligence, well be creating more problems and also solving more problems.

Well, thats very optimistic. I like that. I read recently that Mark Zuckerberg/Meta intended to spend upwards of $30 billion annually on AI development. Of course, Google, with which you are affiliated, and Microsoft and others are investing heavily as well. This is an enormous amount of money. Right. But millions of people are using it, and it is making us smarter. If we can use these kinds of tools, we have much more capability than we had before. Everybody does well. Could you imagine not having these phones?

No. Exactly. Five years ago I used to ask after a speech, who here has their phone? Id say maybe two-thirds had it, but a third didnt have it. Ten years ago, almost nobody had it. Its getting adopted very quickly, but it makes us smarter. No one goes out without it.

Its very true. Sometimes I have to stop myself from googling things I cant remember because I feel like I need to use my brain, test my memory. Its just too easy to find the information on my phone. But that is your brain. Theres no purpose to memorizing these things. You have it in your hand; you might as well use it.

What do you think is the biggest technological discovery in the recent past that is going to profoundly impact the way that we live 10 years from now? Well, the ability to extend our longevity. Thats going to radically change what humanity is about. All of religion, for example, is about this issue. What do you do as you get older? What happens when you die? You have to get used to the fact that time goes by, and you have less and less time. Thats going to completely change. Ultimately, well have more and more ways of keeping our longevity going. And then people say, Well, I dont want to live past 90 or 100 or 110. If you say, I dont want to live past 100, but when youre 99.9 years old, you want to continue anyway. The only time that people want to take their lives is if theyre in unbearable pain. Physical pain, emotional pain, spiritual pain. Then people dont want to go on. But if youre feeling fairly good about life, nobody wants to end their lives. And were continually changing who we are. Life expectancy was 20 a thousand years ago. It was 35 in 1800. Its 48 in 1900. Its now pushing 80. But as it continues to expand, were actually happy about that, and nobody really complains about it.

You are a computer scientist, an inventor, an author, a world-class thinker. Are you religious at all? I believe in innovation. Thats really what my life has been about and what Ive tried to contribute to. I think religion is something weve made up because we didnt have any alternative.

This interview has been condensed and edited for clarity.

Read the rest here:

Ray Kurzwell revisits AI singularity prediction 20 years on | Technology | sfexaminer.com - San Francisco Examiner