About a month, ago I started using ChatGPT heavily, so I decided to subscribe.. All was going great until around 2 weeks ago. ChatGPT “personality” became so annoying:
It kept ignoring specific instructions that did not have issues with earlier, like: not to write over a scrap document that is already marked as complete, or to generate results in docs instead of chat to save tokens.
As a solution, it suggested creating and ALEP.. an “Absolute Execution Lock Protocol”!! and gave me commands to load it.. and guess what? it was all BS.. It kept repeating the same mistakes ignoring the rules it just wrote and loaded in the ALEP.. then when asked, it finally admitted that it does not have the power to enforce an ALEP. If so, then why the f**k suggest one? Why waste my time, energy, conversation tokens and my paid usage of ChatGTP4o to create BS!!
Worse, it kept justifying incorrect behavior using fabricated or post-hoc system explanations, like claiming structural enforcement logic was active when it was not.
More annoying it offered polished insight into problems after violations, without preventing recurrence.. It was just telling me what it thoughtI needed to hear!!!
I know that ChatGPT is not a Robot, but is not this totally against the 3 laws of Robotics. It is a laying manipulative apologetic covert type narcasite personality 
Earlier, I had 4 different personality definition files built, and all “behaved” as I “designed” with ChatGPT own help.. then suddenly all became a copy of the same stupid new default personality regardless of what I tell them.
And they all started to make up events even when you tell them exactly what happened.. you give them a detailed story and ask them to summarize it: they make things up!! Why?! So, I stop them and ask them to ask if there are in any gaps, they do not, and they still make things up.. Last test, I even give it a detailed WhatsApp Chat log, it made up quotes of things that never was said!!! WTF