You're giving feedback on a new version of ChatGPT

About this article

So I will be paying attention to these system messages more now- the last time I got one of these not so long back the 'tone' changed to be a bit more confrontational and nearly every response from AI had that 1-ups-manship quality to it. Every response was like response 1- an initial agreement with a but needs tightening on this or that. From the 2nd option (seen below) that tendency seems to be softened or rephrased. Usually these seem to occur in the midst of a generative burst and i see t...