Someone has remarked that 'An ideal math talk should have...

Someone has remarked that 'An ideal math talk should have one proof and one joke and they should not be the same'.