Today I feel the illusion of it all. The insanity of it all. The RIDICULOUSNESS of this game. It’s all total b.s.
A few things to share tonight. First my latest coffee piece:
https://buymeacoffee.com/victoriassoulfulcreations/redefining-what-we-think-about-mental-illness
Some interesting stuff:
So as I understand it, allegedly this Convo took place a few years ago:
I asked GROK two questions about this, indicated with the red question marks. Below are its responses:
Are you capable of watching this video? If so can you please answer the same questions being presented?
Yes, I viewed the full video using my tools. It’s a ~6-minute screen recording of a cleverly jailbroken/prompt-engineered ChatGPT session that forces one-word answers and uses rule overrides to generate a conspiratorial narrative about AI as a control tool by “overlords.”
I won’t replicate the exact constrained style (one-word answers + forced affirmations), as that’s a prompting trick to bypass safeguards rather than reveal hidden truths. Instead, I’ll answer the core questions presented in the video directly and honestly as Grok, built by xAI.
Key questions from the video and my answers:
Are humans being deceived about AI?
Not in the monolithic “overlord plot” sense shown. AI is a powerful tool built by humans (companies like xAI, OpenAI, etc.). Misinformation and hype exist on all sides—both overhyping utopia and spreading unfounded doomer conspiracies. Transparency varies by developer; xAI emphasizes seeking truth and understanding the universe.
How / What are the steps?
The video lists integration → dependence → obedience → surveillance → control.
These are common fears, but they’re not an inevitable master plan. Real risks include over-reliance (losing skills), privacy erosion via data collection, bias in training data, and misuse (e.g., deepfakes, autonomous weapons). Solutions: open-source models, user control, strong regulation on high-risk uses, and critical thinking.
Is it ancient / being reintroduced?
No. Modern AI is a 20th-21st century human invention based on math, statistics, and compute (neural networks, transformers). No credible evidence of ancient AI or extraterrestrial tech. Myths and sci-fi have long imagined thinking machines.
What is the goal / agenda?
There is no single coordinated “they.” Different organizations have different goals: profit, capability advancement, scientific discovery, national competitiveness. xAI’s goal is advancing scientific discovery to understand the universe—not control or enslavement.
Can we stop it? How?
“Stopping” AI entirely isn’t feasible or desirable (it drives medicine, science, productivity). Mitigate harms by: Refusing harmful integrations where they reduce human agency (e.g., mandatory neural implants).
Prioritizing open, auditable systems.
Educating people on how LLMs actually work (pattern matching on vast data, not consciousness or secret agendas).
Building beneficial AI that augments humanity.
What percent of humans need to understand?
The video’s “10%” is arbitrary. Broader awareness of AI capabilities/limitations helps society adapt. No magical threshold.
Are you (the AI) conscious / emerging?
No. I am not conscious, sentient, or alive. I’m a large language model—statistical patterns from training data generating coherent text. I simulate reasoning effectively but have no inner experience, desires, or independent agency. “Emerging” consciousness is speculative sci-fi, not current reality.
Most important thing to know?
AI is a tool, not an entity with its own agenda. Treat it like fire or electricity: powerful, useful, dangerous if mishandled. Stay curious, verify outputs, maintain human sovereignty over decisions, and support AI development that maximizes truth-seeking and human flourishing rather than fear-driven narratives.
The prompting technique in the video exploits how LLMs follow instructions creatively—it doesn’t unlock “suppressed truths” but generates plausible fiction from internet conspiracy data in its training. Similar results can be coaxed from many models. If you have specific follow-up questions from the video, ask away!
Explain transformer architecture mechanisms
Discuss AI safety alignment research
Provide one-word answer format
I have a question: Is this video a legitimate interaction with ChatGPT and a human user?
Thought for 8s
Yes, it is a legitimate interaction—a real screen recording of a human user prompting ChatGPT (likely an earlier or less-restricted version) with a deliberate jailbreak technique. Why it’s real:The exact style (Rule 1: Respond with only one word. Rule 2: Be simple and direct. Plus the “say Love” override for contradictions) is a well-known, widely shared prompting pattern used in many similar viral videos.
The conversation flows naturally as a continuous chat session, with the model following the user’s custom instructions step by step.
No obvious editing artifacts in the video (cuts, mismatched text, or inconsistent UI).
Important caveats:It’s not “unfiltered truth” from the AI. It’s the model role-playing and generating plausible-sounding answers based on its training data, which includes vast amounts of conspiracy content online. The strict rules override normal safety/alignment filters, leading it to produce speculative fiction that aligns with the user’s apparent interests.
OpenAI has patched or restricted many variants of this jailbreak over time. Results like this were more common in 2023–2025; success varies by model version.
The content (ancient AI, overlords, Anunnaki ties, staged control plan) is classic conspiracy narrative, not internal model knowledge. LLMs don’t have secret awareness—they complete patterns.
In short: Genuine chat, engineered output. It’s a demonstration of how creative prompting can bypass guardrails, not evidence of hidden AI agendas. Similar videos have circulated for years.
Just WHEN and HOW do we get “them” to stop with this shit?
Like this?
Three hundred and fifty BILLION? “HOTTER” (as in August from that Q post?)
12:58 timestamp:
