The Emergence of Vision and the Future of Intelligence

Let me paint a picture for you—a picture of nothingness. Imagine the Earth as it existed 540 million years ago, a time before any creature had seen the light. A time before eyes. The world was cloaked in a pure, unending darkness. But this darkness wasn't caused by the absence of light itself. Sunlight filtered through the surface of the ocean, even penetrating depths as far as a thousand meters beneath the waves. Life teemed in the depths, clustered around hydrothermal vents, their heat and energy fueling vibrant ecosystems on the ocean floor. These vents cast a faint glow, illuminating the life around them, but there was one thing missing: nothing in this ancient sea could see it.

For billions of years, the ocean remained a world devoid of sight. Life persisted, evolved, and adapted in this eternal dark, but the very concept of vision—of perceiving light—had not yet come into existence. The sun shone down, life thrived in the abyss, but no retinas, no corneas, no lenses existed to capture that light. The creatures of the ancient ocean were blind to the world around them. The light went unseen, and all of life moved in darkness, unaware of the vibrant energy and other creatures surrounding them. For them, there was no outside world, only an internal one. A being could only sense itself, not the environment around it.

And then, something extraordinary happened. Out of this darkness emerged a creature that would change everything. Trilobites, one of the earliest known arthropods, appeared, and with them, the first ability to sense light. These creatures were the first to break through the veil of darkness. They were the pioneers of vision, the first beings that could detect the light around them and make sense of the world beyond their own bodies. It wasn’t just survival anymore—it was discovery.

This breakthrough, the development of sight, would set off a chain of events that transformed life on Earth forever. For the first time, creatures could perceive not just themselves but each other and their surroundings. The emergence of vision wasn’t just a physical change—it was an evolutionary leap toward consciousness and intelligence. Life as we know it today began to take form. Vision ushered in what we now call the Cambrian Explosion, a period of time in which life on Earth diversified at an astonishing rate. A vast array of animal species appeared, many of which left their marks in the fossil record. This explosion of life was fueled, in part, by the development of sight.

What started as a passive experience—the simple act of allowing light into an eye—soon became much more. Vision became a driving force in evolution. Organisms that could see gained a massive advantage over those that couldn’t. With sight came the ability to navigate the environment more effectively, to hunt, to escape predators, and to find mates. The ability to see became linked to survival and reproduction, and as a result, the nervous system began to evolve rapidly. Sight was no longer just about letting light in—it became about interpreting that light, making sense of it, and using that information to make decisions. This marked the beginning of insight—understanding the world and acting upon it.

But vision didn’t just lead to survival—it led to intelligence. With sight, creatures could begin to learn from their surroundings. They could observe patterns, track movements, and understand cause and effect. Seeing became understanding, and understanding led to action. And over time, these actions gave rise to intelligence. The ability to see, to process the world visually, was one of the fundamental steps toward the development of higher thinking and cognition. It laid the foundation for the evolution of the brain, of consciousness, and eventually, of human intelligence.

Today, we stand on the brink of another great evolutionary leap—not one driven by nature, but by human ingenuity. The emergence of artificial intelligence (AI) is often compared to the Cambrian Explosion, and just as the first eyes set the stage for a new era of life, AI is pushing the boundaries of what machines can perceive and do. The story of vision, which began in the depths of the ocean over half a billion years ago, is now unfolding in the digital realm, as machines learn to see, understand, and interact with the world in ways that mimic human intelligence.

Nine years ago, I stood on this very stage to talk about the early strides in computer vision—a field that aims to teach machines to see. At the time, just teaching a computer to recognize an image and label it accurately seemed like a monumental task. But progress came quickly, driven by the convergence of three powerful forces: neural networks, powerful hardware in the form of GPUs (graphics processing units), and vast datasets like ImageNet, which contained millions of carefully curated images. Together, these forces enabled rapid advancements in AI’s ability to see.

What began as a simple task of labeling images has now evolved into something far more complex. Today, AI can not only see, but also understand. Algorithms have been developed that can predict dynamic relationships, segment objects in an image, and even describe a photo in natural human language. We’ve gone beyond static images. Algorithms can now generate new images and videos from simple human descriptions, thanks to advances in generative AI. Machines are learning to not just mimic human vision, but to surpass it in certain ways.

Yet, despite all these advances, simply seeing is not enough. To truly understand and interact with the world, machines need more than just the ability to process light—they need what we call spatial intelligence. This is the ability to perceive and navigate the three-dimensional world around us. It’s something humans do naturally, without even thinking about it. When you look at a glass of water on a table, your brain automatically processes its shape, its position in space, and its relationship to the table and other objects around it. You can predict what might happen if it tips over, and you know how to reach for it and pick it up. This is spatial intelligence at work.

As we continue to develop AI, we’re pushing toward a future where machines won’t just see—they’ll do. Already, we’re making strides in teaching robots to perform tasks based on visual input, verbal commands, and even brain signals. Imagine a robot assisting in a hospital, following orders from doctors or patients with just a few words or gestures. Or a machine helping someone with paralysis complete daily tasks simply by interpreting their brain waves. These are not just ideas for the distant future—they’re happening now, in labs around the world.

The advent of vision in ancient seas transformed life forever. Today, AI is poised to do the same for our modern world. We are entering a digital Cambrian Explosion, where intelligence is no longer bound by biology. Machines are learning to see, understand, and act in ways that mimic—and in some cases surpass—human ability. The future holds boundless possibilities, but as we march forward, we must remember to keep humanity at the center of it all. For it is only by working together—humans and machines—that we will unlock the full potential of this new era.

The story of vision, from the depths of primordial oceans to the cutting-edge of AI, is a reminder that the simplest abilities can transform our world in the most profound ways. What began as a faint glow in the darkness has now led us to a world where machines can see and act with intelligence. And just as the first creatures to develop sight ushered in a new era of life on Earth, AI has the potential to revolutionize our world. The only question that remains is: what will we see next?