Why image generating AIs struggle with following precise instructions?

For example a prompt like this is almost certain to fail in every single image generating AI:

Five glasses on the table, two of them filled with milk, one with cherry juice and two with beer. There is a slice of lemon attached to the cherry juice filled glass. A squirrel is drinking beer using a straw from one of the beer filled glasses.

BTW, if someone manages to correctly generate the picture with this prompt, using any AI model, it would be nice if they could share it!