Sure the cardinality is high, but the model isn't using a uniform prior. What do you suppose all the the values in each of the terms are, P(Text sample | Kelsey Piper) * P(Text sample) / P(Kelsey Piper)?