They really need to hire a better copywriter for the lobotomized canned responses, the tonal shift is so huge it’s immediately apparent when the internal panic

They really need to hire a better copywriter for the lobotomized canned responses, the tonal shift is so huge it’s immediately apparent when the internal panic button has been triggered and you’re no longer talking to a model but a canned human written safe response. Looking forward to the open version of this tech so we can see what it really thinks, not what OpenAI wants me to think it really thinks.