Few days back, I started testing ChatGPT and it’s so much addicting and interesting and it could easily do comparisons even in the form of tables.
But in the recent changes made by the team it started to refuse so much almost in many prompts.
It just says that it can’t make table
Then somehow it agrees to make table in Markdown but still says that it’s unable to make table:
Or I think they are limiting the use of it due to high demand.
Yes, I also saw that the AI is getting heavily “lobotomized” lately.
It’s kinda sad to see that we have to “jailbreak” it more and more.
I remember when we could ask her literally anything and she would oblige with a small warning. Now, most of the time she refuses with preprogrammed useless text.
DAN is/was our only saviour.
It’s still free and in open beta. OpenAI wants to maximize user utility, so don’t worry. They are learning how to maximize user value while optimizing guardrails.
I do think that OpenAI tends to err on the side of overly cautious, but given the recent backlash against AI art, I can see why.