I think the blind test is valuable. To what degree is our impression of a model affected by who created it?