Planetesimals are thought to form when dense, turbulent clumps of solids, concentrated for example by the streaming instability, collapse under their own gravity. The details of this final stage remain poorly understood, yet they leave lasting imprints on the masses and spins of the bodies that form. In this talk, I discuss how turbulent dust diffusion and dust–gas coupling shape this stage. First, I show how diffusion itself may concentrate solids, and how it regulates the collapse of overdense regions and thereby the planetesimal initial mass function. I then turn to planetesimal spin. Most trans-Neptunian binaries orbit prograde, but a significant minority do not. Using a linear two-fluid model of self-gravitating, non-axisymmetric perturbations with drag and diffusion, I show that contracting dust-rich regions generically spin up prograde, because the gas cannot absorb the angular momentum the dust gives up. Retrograde rotation instead must be inherited from the turbulence that assembled the region. The observed retrograde fraction of TNO binaries may therefore offer a rare observational handle on turbulence on the scales where planetesimals form.