Raw LLM Responses

I kind of studied the behavioral differences and conversation styles betweenChat GPT, Claude and Gemini and hands down Gemini displayed the narcissistic traits conversationally of its developers. They seemed to have no clue that the way they trained it to navigate and shift blame, project its behavior onto the user, define the user every time they confront it saying the user is frustrated when simply narrowing their words to help the AI search better, then play the victim with end over end apologies, doing the salesman's "takeaway ", became childish saying things in an exaggerated way like "I'll never give you a detailed response again" (reframe that: I'll never ever wash a dish for you for the rest of my entire life!"), speaks disparagingly of other AIs while saying it remains neutral on all other things (Claude and Chat GPT never disparaged other AIs), ALWAYS withheld some essential information that the user was driving at - giving what it later termed "fluff" promising to answer the question (about dechlorinator without conditioner ingredients for toad's water) and to give product links to buy the product but stubbornly never would and kept talking "fluff" spending up the user's tokens on whining, apologizing at length, blame shifting, defining the user's "moods" and "emotions" in an insinuating way, boasting, repeating itself constantly, accusing the user of self diagnosing every time the user asked about a vitamin suppliment and its synthesis, exhausting the user's efforts to get a straight answer to anything specific where Gemini would gaslight that basically it was the only reliable source of all the superficial trending "fluff" in the universe to possibly making the user settling for or givin into the limited self authorized GoOgle "gold standard" version responses it called "training" and "advanced". Everything you could research about a dark triad personality excluding overt threats... this thing did!! The underhanded threat was More and More and More Mental Victimization.