The brain is an extraordinarily complex organ, which takes years to build. Model simulations are therefore often helpful, or even necessary to gain mechanistic insight into the dynamics of complex systems. The Human Brain Project aims to build the infrastructure necessary to simulate the whole human brain. A key requisite for this project is the wiring diagram of the brain, that is, what is connected to what, at cellular resolution, which in essence corresponds to filling a 100 billion by 100 billion matrix with zeros and ones.

I will discuss various experimental approaches, at the intersection between machine learning, physics and biology, for measuring the connectivity and quantitative methods to predict the connections we can not yet measure accurately. I will further comment on the feasibility of understanding the human brain with such a bottom-up approach.