The bigger picture

For the first time, three capabilities exist at once. We can record a human life in dense detail. We can build an agent that learns from that life under human constraints. And we can test whether the agent computes as the brain does. Their convergence changes what counts as a complete theory of the mind.

The program rests on one hypothesis: artificial and biological neural networks belong to the same family of computational solutions. The claim is strong and falsifiable. If alignment proves weak or coincidental, the framework needs revision. If it holds, an agent that learns and behaves as a person does becomes a working model of the mind and brain.

From here, each objective grows. The First 1,000 Days project can grow from fifteen homes toward a thousand, across cultures and languages, and follow people beyond the home, into the playground, the school, the university, and the workplace, with sensors that capture what each person sees and how their body responds. The agents can gain eyes, a body, and people who respond to them. And the models can take on more of the brain’s own machinery: loops in time, spiking neurons, a memory, a learning rule a real synapse could carry out.

Two things follow. Once an agent reproduces a typical child, we can ask what small change reproduces an atypical one, giving a new way to study development. And the lab does not disappear. A built agent can be brought into the lab like any participant, and unlike a participant it can be reset, rerun, lesioned, and inspected weight by weight.

No single lab can build this. A record of human lives from birth to old age must outlive everyone who starts it. We invite researchers across fields to join us: to record real lives, to build agents that learn from them, and to test those agents against human behavior and the brain.