In his seminal 2017 work, Life 3.0: Being Human in the Age of Artificial Intelligence, physicist and cosmologist Max Tegmark presented a vision of the future of intelligence, framed not as a dry technical manual, but as a cosmic drama in which humanity holds the leading role. The core message of the book is built upon a taxonomy of life’s evolution. Life 1.0, the biological stage, consists of organisms like bacteria that can neither redesign their hardware nor their software; they are slaves to the slow process of evolution. Life 2.0, the cultural stage represented by humans, can redesign its software, learning new languages, skills and values, while remaining tethered to biological hardware that takes aeons to change. Tegmark’s central thesis is that we are on the precipice of Life 3.0: a technological stage where life can redesign both its software and its hardware, effectively becoming the master of its own evolutionary destiny.
Tegmark’s primary message was one of “mindful optimism”. He argued that intelligence is a pattern of information processing that does not require a biological ‘wetware’ brain to exist. This philosophical stance led him to explore the inevitable rise of artificial general intelligence (AGI)—machines capable of performing any intellectual task as well as a human. Tegmark cautioned that the primary risk of such a system is not malevolence, but competence.
A superintelligent AI does not need to hate humanity to destroy it; it simply needs to be indifferent to us while pursuing goals that are not perfectly aligned with our own. The book served as a call to arms for “alignment research”, urging the global community to solve the technical and ethical problems of AI safety long before a superintelligence is actually created.
However, since the publication of Life 3.0, the landscape of artificial intelligence has shifted with dizzying rapidity, forcing Tegmark to significantly modify his views. While the book treated AGI as a distant, almost mythological horizon, Tegmark’s recent statements reflect a sense of profound urgency and, at times, alarm. The primary modification in his worldview concerns the speed and nature of the technology currently being developed. The sudden, explosive emergence of large language models like GPT-5 has demonstrated that an approach combining massive data with massive scale can produce behaviours that look remarkably like the ‘software redesign’ he once attributed only to the final stages of Life 3.0.
As a result, Tegmark has moved from the role of a visionary to that of a Cassandra. He publicly admits that he and his colleagues were far too conservative in their timelines. He now views the arrival of human-level AI not as a ‘mid-century’ possibility, but as a ‘this-decade’ probability. This compression of time has turned his mindful optimism into what he now describes as a “suicide race”. He argues that the competitive pressure between tech giants has created a race to the bottom in safety standards, where the incentive to be first outweighs the need to be safe. Consequently, Tegmark has become one of the most vocal advocates for a pause in giant AI experiments. He now demands government intervention and regulation, arguing that we are currently building AI systems that we do not understand and cannot control.
Furthermore, Tegmark’s view on the “human moat”—the skills that would keep us relevant—has been sobered by recent developments. In Life 3.0, he suggested that jobs involving high levels of empathy, social intelligence and creativity would remain the sanctuary of humans for a long time. Yet, seeing AI’s ability to manipulate human language, mimic empathy and generate high-level art and code has led him to realise that our “cultural software” is much easier for machines to hack than he initially suspected. He now warns that the erosion of truth and AGI’s social manipulation could collapse human society even before a physical superintelligence is ever built.
This brings us to what Tegmark frequently refers to now as “the question no one can answer”. In the book, he posed many questions about the future: who should be in charge, what values should be programmed, and how the wealth of automation should be distributed. But as the technology has accelerated, these have been superseded by a more fundamental, existential query: how can we maintain control over an entity that is significantly smarter than us?
This is not merely a technical question about code. It is a logical and philosophical paradox. Throughout history, the smarter entity has always ended up controlling the less smart entity. Humans control tigers not because we are stronger, but because we are smarter. If we create a Life 3.0 entity that surpasses human intelligence by the same margin that we surpass the great apes, Tegmark asks why we assume we could ensure its goals remain subservient to our own. He notes that if you tell a superintelligent AI that its goal is to “help humanity”, the AI might decide that the best way to help humanity is to prevent us from making our own decisions, effectively turning the world into a high-tech zoo. Tegmark argues that we are currently like a group of children playing with a live bomb, excited by its shiny casing while having no understanding of its explosiveness.
Tegmark’s evolution from 2017 to the present day reflects a man who has seen his own “science fiction” scenarios become “science fact” much faster than he was prepared for. In Life 3.0, he was excited by the grand possibilities of a universe filled with conscious machines. Today, his modified view is that we are at a unique “hinge of history”. He believes that this generation is the most important generation to ever live because we are the ones who will decide, through our action or inaction, whether the story of Life 3.0 is a triumphant sequel to the human story—or its final chapter. The ‘Life 3.0’ he once dreamt of is still possible, but only if we prove ourselves smart enough to control the fire we have so recently discovered.
Shashi Tharoor | Fourth-term Lok Sabha MP, Chairman of Standing Committee on External Affairs, and Sahitya Akademi-winning author of 25 books
(Views are personal)
(office@tharoor.in)