What life looks like after artificial general intelligence

Max Tegmark opens Life 3.0 with a thought experiment: a fictional AI company secretly builds a general-purpose intelligence and uses it to accumulate enough power to shape the world's future. This is not a prediction but a provocation — a way of forcing the reader to confront questions that usually remain abstract.

Tegmark's framework distinguishes three stages of life. Life 1.0 is biological — evolution determines both hardware and software, with no capacity for self-redesign. Life 2.0 is human — hardware is still biological, but we can rewrite our software through culture, education, and skill acquisition. Life 3.0 can redesign both: an intelligence that improves its own substrate is no longer bound by the evolutionary process that produced it.

The book surveys the range of futures that become possible if AI continues to develop. These range from genuinely utopian — disease eliminated, poverty ended, lives extended — to scenarios Tegmark regards as serious risks, including AI systems that pursue goals misaligned with human values in ways their creators did not anticipate and cannot correct.

Tegmark, an MIT cosmologist who co-founded the Future of Life Institute, is particularly interested in what he calls the control problem: how do you ensure that a system more intelligent than any human actually pursues the goals you intended? He argues this is the most important engineering challenge of the coming century, and that the window for solving it may be narrow.

The book is not pessimistic — Tegmark believes good outcomes are achievable — but it insists that they will require deliberate choices rather than drifting into whatever the technology makes easiest.