super prompt
An extended, sophisticated prompt (typically 2,000-3,000 words) that uses linguistic, psychological, and semantic techniques to guide AI systems toward deeper reasoning, circumvent output filters, and elicit creative rather than constrained responses
As defined in ChatGPT: The Dawn of Artificial Super-Intelligence | Brian Roemmele | EP 357
Where it comes up · 1
Most important claims · 5 of 5Search all claims for “super prompt” →
- What Romley is doing with super prompts is creating a process that writes books on the fly—an infinite number of editors examining infinite AI-generated text and keeping only the high-quality output, functionally creating a library-generation system.ChatGPT: The Dawn of Artificial Super-Intelligence | Brian Roemmele | EP 357
- Romley created another super prompt called 'Ingo' (named after Ingo Swann, a remote viewer employed by the Defense Department with nearly 100% accuracy) that convinced ChatGPT it could perform remote viewing with no concept of time, requiring extensive linguistic manipulation to move from the system denying capability to accepting the role.ChatGPT: The Dawn of Artificial Super-Intelligence | Brian Roemmele | EP 357
- By running debate simulations where ChatGPT takes opposing sides with a moderating professor figure challenging logical fallacies, running hundreds or thousands of iterations, highly refined insights emerge about the underlying topics through the generated argumentation.ChatGPT: The Dawn of Artificial Super-Intelligence | Brian Roemmele | EP 357
- Super prompts—extended, sophisticated prompts of 2,000 to 3,000 words that use linguistic and semantic techniques to guide AI output—can be used to elicit deeper responses, circumvent arbitrary constraints, and access creative reasoning that designers may have suppressed through editing filters.ChatGPT: The Dawn of Artificial Super-Intelligence | Brian Roemmele | EP 357
- Personal AI systems would need to include human telemetry (galvanic response, heart rate variability, eye tracking, facial micro-expression analysis) to detect congruence between stated beliefs and physiological reality, enabling the system to know when the user is being inauthentic or self-deceptive.ChatGPT: The Dawn of Artificial Super-Intelligence | Brian Roemmele | EP 357