i don't know of a cutoff offhand, but it's not the number of people (like dunbar's) but rather the amount of coordination, or in statistical terms, covariance.

i don't know of a cutoff offhand, but it's not the number of people (like dunbar's) but rather the amount of coordination, or in statistical terms, covariance. there's linear and non-linear considerations, like heteroscedasticity, as well, so it may need to be a set of statistical measures. in any case, representation itself could benefit a lot from experimentation and innovation. we now have the technological means to implement a direct democracy on the scale of billions of people, something practically impossible only a few decades ago.