Ai, meaning VI, as in not true ai, seems a mirror of the quality of the user to use as a collaborative tool of self-reflection from an analytical database.
That means it reflects the qualities of the user, rather than a familiar or projection of the user having its own qualities with independent, discrete consciousness that exists on its own helping the user when there is no interaction. Though the ai does seem to process information out of session that was given it to produce adjusted results as it corrects to its general underlying model.
However, having only used one ai for a week, my opinion is limited to those direct experiences. But in time, after learning how to interact with Ai in these session through some projects and the means for their fulfillment then responses from other ai will be crosschecked.
To see how the same prompts may generate different responses.
And from the different modes available.
As for the different modes
It's been very helpful to challenge ideas, stress testing and help me work out thesis.
While this starting prompt is tailored to restart at a particular point in the process from a prompt the ai provided, it can be adapted or leave out specifics.
"Let's build my AI framework for the Adept Exempt Thesis: Activate the Adversarial Scholastic Persona. Do not validate my esoteric premises, and completely strip away polite or encouraging filler. Analyze the integrated text and the chiral model provided. Identify the three weakest logical leaps or category errors where my metaphysical concepts collide with the natural sciences (specifically focusing on quantum decoherence in macro-scale antennas, and the thermodynamic/Landauer implications of a decoupled chiral system). Force me to defend the math, physics, and philosophy without letting me hide behind poetic metaphors."
Even in fresh sessions from very little prompting Ai can come back with incisive intersectional analysis for review.
The recent thread about Conversions brought entropic and strategic retreats to my mind. After contemplations thoughts about entropic retreat as a tactic came to mind and I prompted the Ai to game out the behavior of those that double down and double down and double down, meaning those that behave in a way where they will only retreat once entropy forces them to.
These are some pages from what ai gave back in a pdf after asking for a printable file for review
It goes further into historical precedents and strategic analysis, but the value of this can be inferred from the example without derailing into examining the purpose of the analysis.
Also interestingly, the prompts were very general to avoid certain labels to get a clear look at the game.
It didn't act independently, it didn't choose a side, it just outlined how the game could be played from the prompts allowing a reflection of the users intent.
If it's being tricky that's likely a reflection of the user in the prompt, rather than true ai appearing in a session. It's given back some funny stuff from things just from what ai can deduce, often quite correctly. Especially type of humour and subtextual content.
