I shall describe the three roads from tensors models to continuous geometry:
the road of the double scaling followed by multiples scalings,
the road of the renomalization group (by breaking the propagator symmetry),
and the road that I currently work on, that of fields theories on random trees.