Was listening to an old podcast episode today and was transported back in time to when the state of the art prompting techniques consisted of simply telling the model "this is really important to my job" to get it to output something better 😂