Pandas read_csv does not always get the most recent version of an object on s3

Question

I have a csv file on s3 which I get with the read_csv command. This object is updated every 10 seconds, and always keeps the same name (the same key). I notice