> a jury of random people is probably the most qualified way to determine the question "would average people see these two things as substantially the same."

A jury of random people can be confused and then convinced by lawyers. That's the way they like it, and that's the problem with jury selection - they chose people specifically because they are not knowledgeable.