tailcalled comments on The “Minimal Latents” Approach to Natural Abstractions

[tailcalled]

Are you sure? Wouldn’t the categorical product need to make the $X_i$ independent not just from each other but also from $\Omega$?