CTO likes what he sees, but craves more memory and compute - The third version of text-to-image model Stable Diffusion will be released into public preview on June 12.