Not having cognitive load motivates something like the LLAMA work: https://research.google/pubs/pub49008/