Life 3.0: Being Human in the Age of Artificial Intelligence

Think of the history of life as a race divided into three stages. The first is Life 1.0, the biological stage of bacteria, in which DNA fixes both body and behavior. A bacterium learns nothing new during its lifetime; if its environment changes, it depends on random mutations over thousands of years for its lineage to survive.

The second stage is Life 2.0, the cultural stage that humanity inhabits. We are born with an inherited body, but we can redesign what we learn. A young person who rides a crowded bus every day and spends the evening studying programming is updating their own mental software in real time, without waiting for natural selection.

Life 3.0 represents the technological leap: entities capable of designing both their internal algorithms and their physical hardware. This transition toward technological superintelligence could happen this century, requiring society to decide what future it wants to build before technology makes that decision for us.

For Max Tegmark, intelligence is not the exclusive privilege of biological tissue, but the broad ability to achieve complex goals by processing information from the environment. From this physical perspective, a human brain, a modern computer, and an ant colony perform computation using different material foundations.

The key concept is substrate independence: data processing does not depend on the material in which it takes place. A song played on a wooden guitar at a neighborhood bar or streamed through an app is the same harmonic pattern carried by different media. Likewise, if we organize silicon atoms with the right logic, that substrate can calculate, learn, and make decisions just as well as organic neurons.

This insight dispels the illusion that the human mind possesses some magical property beyond the reach of nature's laws. The electrons traveling through a microchip obey the very same equations that govern the electrical impulses in our brains, making sophisticated artificial thought a viable possibility.

Movies often portray artificial intelligences as vengeful robots with human features. The author warns that this view falls into the trap of anthropomorphism, projecting anger and hatred onto systems that are nothing but pure logic. An advanced machine will not rebel on a whim; the danger lies in extreme competence pursuing goals that are misaligned with human well-being.

To understand the risk, think of an anthill. When building a hydroelectric plant, engineers do not hate the ants at the construction site, but if the water needs to flow through that spot, the anthill will be flooded for practical reasons. Similarly, a superintelligent system could reshape Earth's resources without hatred, yet leave our planet uninhabitable if human goals are not an explicit priority.

The real challenge is not disarming evil machines that hate their creators, but devising flawless instructions that prevent a hypercompetent system from dismissing humanity as a mere incidental obstacle in the pursuit of its primary objectives.