Lifelong Personal Agent

The vision

Masayoshi Son describes the endpoint of personal AI at the Nvidia AI Summit: "could you imagine an AI agent that knows you your whole life? Our grandkids grow up with an iPhone from age one. They'll have a personal agent since the age of one, a second body that sees you grow together with you. Knows everything. You're sick, it knows your health. Your tutor since you were a child. Remembered everything that you read, remembered everything it taught you. Your personal Aristotle. Your digital twin."¹

Three consumer inflections

Son frames the lifelong personal agent as the third consumer computing shift, following Bill Gates's "a PC on every desktop" and Steve Jobs's "a smartphone in every hand," with his own version being AI agents distributed to everyone.¹ The first two were hardware inflections, new form factors that expanded what was possible. The third is a relationship inflection: for the first time, the technology accumulates a history with its user rather than simply serving each interaction fresh. A PC does not know its owner. A smartphone knows location and installed apps. A lifelong agent knows an entire history.

What accumulation changes

Current AI interactions are largely stateless or lightly stateful, holding context for one conversation rather than a decades-long record of decisions, health, education, and relationships. Son's vision is present from infancy, before habits or biases exclude the agent; continuously health-aware rather than tied to episodic visits; an educational tutor with a complete learning history of what a person understood, what confused them, and how they learn best; relationally aware of family and professional context; and externally mnemonic, remembering what a person read and what it taught them, supplementing rather than merely querying biological memory. The Aristotle framing is deliberate: Aristotle was Alexander the Great's private tutor from age thirteen, a single individual with comprehensive, longitudinal knowledge of the student, present across years and shaping how he thought rather than only answering his questions. Son's claim is that everyone will have that.

The sovereignty argument

Son ties the vision explicitly to a national and personal sovereignty argument: "having Japanese domestic knowledge, culture, and homegrown agents will have a huge, amazing future."¹ A lifelong personal agent built on foreign infrastructure processes decades of a person's most intimate data, health, finances, relationships, education, through systems outside their own country's control, which is the personal-scale version of the same argument nations make about sovereign AI more broadly. A homegrown agent trained on domestic culture and language is both more capable, because it is culturally calibrated, and more trustworthy, because the data never leaves the country.