Back in 1996, Ishibashi, Kawai, Kitazawa and Tsuchiya proposed a supersymmetric matrix model as a nonperturbative formulation of superstring theory.
The beauty of this model is that 1) there is no parameter other than one scale parameter, 2) it has the maximal supersymmetry in ten dimensions, 3) the (3+1) D space-time is expected to emerge dynamically as the eigenvalue distribution of the 10 Hermitian matrices.
I will review the status of this proposal focusing on our attempts to explore the beginning of the universe by computer simulations.