[Linkpost] Scaling laws for language encoding models in fMRI — LessWrong