New method allows to convert auto-regressive models into diffusion models with a >2x speedup, fully compatible with existing inference stack

If the claims presented in the paper are true, this will be very big for multi-user local inference