ChatGPT has turn into predictable. Agreeable. Boring. It simply stored agreeing with me—even once I’m clearly unsuitable. Then, accidentally, I discovered an AI that does the precise reverse—and I’m hooked.
I Cannot Stand ChatGPT Calling Me a “Genius”
ChatGPT has unediably been declining. Each time I might reply with my opinion, ChatGPT would gush again, “You are completely proper!” If I caught a mistake in my earlier messages, “Good catch! This exhibits that you just actually perceive your work!” It was cute for the primary 5 seconds. By the hundredth praise, I used to be able to scream.
There is a motive individuals continuously attempt to jailbreak ChatGPT—to not do something malicious, however to beat its absurd default directions.
Truthfully, I get why ChatGPT would do that. OpenAI desires buyer retention. We have moved previous the purpose the place LLMs had been area of interest instruments; they’re mainstream now. Increasingly common customers—together with older individuals who choose a pleasant, non-confrontational expertise—are flocking to them. If ChatGPT retains praising the typical joes as “geniuses” who supply “good views,” they’re extra prone to come again.
This technique works, however not for somebody who desires outcomes and never chit-chat. It begins as mildly irritating however shortly turns into a real impediment. Once I current a half-baked concept searching for real critique, ChatGPT buries any potential criticism underneath a mountain of false flattery. If it identifies any issues in any respect, I’ve to dig by layers of reward simply to seek out them.
Associated
How I Cease AI Chatbots From Agreeing With Me All of the Time
Fixed settlement isn’t intelligence.
Grok Is the Unorthodox Savior
I am locked into OpenAI’s ecosystem. I’ve talked about this earlier than when ChatGPT went offline for a day, however with all of the reminiscence upgrades ChatGPT has obtained, it is aware of me very effectively. I’ve arrange initiatives, personalized directions, created personalised GPTs, and mainly performed every little thing attainable to tailor it.
Due to these customizations, I have never explored different AI choices in a very long time. I used to be out with a good friend once we determined to check out Grok, and its voice mode instantly impressed me. Past content material, the velocity of response alone blew ChatGPT away.
I might been so entrenched in ChatGPT that I by no means tried Grok out for myself. Maybe Grok is not but as highly effective as OpenAI’s GPT fashions, however not like ChatGPT, Grok truly allows you to use it. In any case, an AI is barely nearly as good as its prompts and default directions. If the directions are horrible, processing energy or knowledge quantity turns into meaningless.
The Argumentative Persona
Once I opened Grok’s voice mode settings, a listing of personalities and animated avatars appeared. What instantly drew my consideration (except for a couple of questionable choices) was the Argumentative character.
I chosen it, after which voiced a frustration that was on my thoughts on the time—one thing that I had additionally introduced up with ChatGPT, and ChatGPT had validated it.
Amir Bohlooli / MUO
Grok immediately flipped the script. As an alternative of agreeing and placating, it challenged me head-on. It questioned my views and pushed me to offer proof. Frankly, it was brutal.
Speaking with AI has aggravated or pissed off me, however by no means had I genuinely felt underneath assault by an AI. It was psychologically attention-grabbing to me how I could not simply cease the dialog to save lots of myself. Grok made me really feel compelled to defend my concepts rigorously or threat internalizing its harsh criticisms as legitimate.
The alternate went forwards and backwards relentlessly till I managed to considerably soften its stance—however even then, I felt roasted. It was a grueling expertise, particularly as a result of I had began with one thing private and significant with out anticipating such intense scrutiny.
I don’t have a premium subscription for Grok, so every little thing I’ve skilled has been with the free Grok 3 mannequin. I think about that the premium fashions are even higher.
Voice Mode Completed Proper
Much more shocking was the effectiveness of Grok’s voice mode. It genuinely felt just like the text-based interplay, simply delivered verbally—precisely how voice mode ought to perform. In distinction, ChatGPT’s “superior” voice mode feels severely restricted and frustratingly inept. It regularly loses context, ignores directions, and amplifies all of ChatGPT’s shortcomings tenfold.
This situation doubtless stems from OpenAI’s recognition that informal or older customers choose nice, non-challenging interactions over productive or essential ones. Grok, fortunately, prioritizes effectiveness over pleasantness.
Getting an concept previous Grok’s Argumentative persona is an accomplishment. If one in every of my concepts truly survives the grilling, I do know I can take it public—as a result of Grok has already thrown each attainable hole, query, and flaw at it. If it holds up there, it’ll maintain up anyplace.
ChatGPT does not lie outright—it simply coats every little thing in sickening optimism. Even clearly incorrect concepts are labeled “distinctive views.” Grok’s argumentative mode is the other: it’s going to problem you even whenever you’re undeniably proper.
Associated
ChatGPT Nonetheless Cannot Reply These 4 Straightforward Questions
ChatGPT retains stumbling over these questions—are you able to resolve them?
After all, Grok’s criticism is not objectively “true.” Keep in mind that it is an LLM that merely manipulates phrases. It does not actually assume or motive, regardless of what advertising and marketing would possibly recommend.
Grok will problem you on right statements and unsuitable ones alike. It’s relentless by design. Deal with its critiques like stress‑assessments, not ultimate judgments. Take what lands, discard what doesn’t, then transfer on.
