Putting humans and LLMs head-to-head in classic tests of judgment from human psychology underscores the differences between them