Sorry if this comes across as a rant, I just came off a frustrating session with my LLM, who tries to be "smart" by assuming that their mode of thinking is "sufficient" for my requirement. I recalled in 2024/2025, which new model brought a new excitement to the users than the previous version - "you mean the model can do this now?" Now, it is the inverse - "you mean the models are trying to optimise itself?" Flexible thinking on the pretext of saving t...