13 – Introducing: A Window on Intelligence

Transcript

This transcript was generated automatically and may contain mistakes.

Last summer, I started writing a book about what I think is our best understanding of creativity and artificial general intelligence. So today, I have an exciting announcement to make. I have now completed and published the book, and I wanted my listeners to be the first to know. It’s called A Window on Intelligence. If you like this podcast, I think you’re going to love the book. I have made some progress, answered some open questions about creativity, and changed my mind about some things too. You can read all about that in the book.

I want to share an audio excerpt with you here. I hope you enjoy reading the book as much as I enjoyed writing it. You can find it on platforms such as Amazon as paperback and ebook. And now I give you an excerpt from A Window on Intelligence.

The universe is dark, cold, and mostly empty, interspersed with the occasional galaxy. Each galaxy, in turn, is mostly empty too. Its stars are spread out thinly, but they seethe with violent activity on the inside. There, transmutation turns lighter elements into heavier ones and releases energy that makes stars shine brightly. Over large time scales in this cosmic theater, stars are born and die, planets form and perish. We do not know why the universe contains something rather than nothing, but we do know that all of these processes follow simple, elegant laws of physics.

The few places in the universe that are neither violent nor empty are mostly planets and moons with barren, bleak landscapes. Earth’s moon is one such desolate place. And yet, here on Earth, just over one light second away from the moon, we see more around us than just gray dust and craters. Here, we find complexity. We see grass, trees, and animals.

The creation of stars, elements, and planets does not require knowledge. It happens spontaneously. Therefore, it only represents a tiny part of all possible physical processes. The more complex ones we find here on Earth, such as, say, trees, are rare. They do not occur spontaneously. Something has to create them. What is this mysterious thing? It is a replicator. The few mistakes it makes during replication add up over time. What starts as a simple replicator can turn into vastly elaborate contraptions.

A replicator’s activity can be described as an open-ended search problem. It inadvertently searches the space of possible configurations of matter to create instances of itself that replicate even better. That is how extremely unusual, even queer structures can emerge. Some of these structures happen to be trees, others bacteria or dinosaurs. In at least one lucky instance, people emerged. These structures are the result of biological replication, but they still do not account for all of the complexity we see around us. Cities, for example, carry the imprint of a different kind of replicator.

While some animals influence their environment, such as beavers building dams, it is doubtful that biological adaptations could result in something as complex as cities or something as far-reaching as space shuttles. Maybe it is possible in principle to genetically evolve the ability to fly to the moon, but if it is, it cannot happen as rapidly and with as much reach as through this other kind of replicator. Therefore, even the complexity we observe in the Earth’s biosphere can only account for a small part of all possible physical transformations.

So, where do complex structures such as cities, airplanes, and computers come from? From the replicator inside people’s minds. Its effects on the physical world begin in a brain’s memory storage and reach out to create the complex configurations of matter we see all around us. In at least that one lucky instance in our evolution, a genetic mutation was the spark that lit the mind’s creative fire and gave us a searchlight to understand the universe.

Behind every physical instance of a replicator of any kind is an abstraction, a function. Functions are programs we can run on computers. They are also explanations. They can evolve more rapidly and develop more reach when they are instantiated not biologically in DNA molecules, but people’s minds. Given enough time and memory capacity, people’s minds can evolve any computable function. Since functions can reflect any physical process and can, in principle, perform any side effect, their creators, people, have infinite reach to transform the universe in any way they desire.

That is how we account for the complexity we see around us. And it is how we know that every change in the physical world is born out of some function. Every such change is the result of software, and all people are software engineers. Cities, airplanes, computers, and other inventions are part of an intricate life support system designed by people to thrive in an otherwise bleak and hostile world. Using their creativity, people can turn even the barren landscapes on other planets into life support systems bustling with activity.

Most possible physical transformations do not occur spontaneously because they require knowledge with sufficient reach that only people could create. Although we cannot predict any particular such transformation, an understanding of what transformations might happen non-spontaneously anywhere in the universe depends in large part on an understanding of people. Therefore, no science and no philosophy can even be close to complete without such an understanding.

The research behind artificial general intelligence, AGI, is the study of people and develops this understanding. Its goal is to simulate the wonders occurring inside our minds, thereby creating a person on a computer. Since nothing possible is outside the reach of people, they can influence everything in the universe, and so an explanation of anything mentions them at least implicitly. Therefore, the study of AGI underlies everything else and is of fundamental cosmic significance.

We know how to make progress by writing good function implementations, mentally or otherwise. There are forces trying to stop us from succeeding. They want us to tumble into stagnation, depression, and eventually death, and they compel us to avoid the very things that help us make progress. They permeate even Silicon Valley, the technological hub one would expect to birth AGI. Alas, it appears they are trying to prevent that birth, but the solution to this problem is as much within our reach as everything else. We can move mountains, planets, stars, and maybe even black holes, if we choose to. Our future is open.

We should work on AGI because we want to, not for the benefit of society, but our self-generated reasons. It is a project of momentous reach because it solves a great many problems. I think simulating a person on a computer would be about the coolest thing anyone has ever done.

But it goes deeper than that. Those of us who work on AGI do so because we are lucky to be alive at this critical moment in the history of the universe. Without the development of AGI, our species may never leave its limited origins on Earth. It may eventually go extinct, just like 99% of all species that have ever lived. We can do better than that. We can be the one species that rises to the challenge. If we develop AGI, not only can our species survive, we can spread throughout the universe and live among the stars.

Would you not like to one day look up at the moon and see lights of cities shining back down on us, like in those photographs of nighttime on Earth? And would you not like to look up at the stars and know that there are people out there thriving indefinitely?

A Window on Intelligence, out now.