Previous observations gave an impression that circumstellar disks typically have radii of hundreds au. Relatively popular examples would be the 2014 ALMA Long Baseline Campaign (HL Tau) and the ALMA survey DSHARP. This is consistent with a classical picture in theoretical work. Numerical simulations have also aimed to form such a large disk, for example, solving the magnetic catastrophe. However, latest surveys suggest that such 100-au sized ones are not representative of circumstellar disks, but disks have more diversity including smaller disks. I would like to discuss the diversity from a viewpoint of protostellar evolution, showing our group's observational results of large and small disks.