Is it impossible to compute multiple models in parallel using brms with `backend = "cmdstanr"`?

Great news!