I wish these posts that talk about non-human mistakes that agents make would post some examples. They would be interesting to see.