i dont work with A.I. but after a month of mathematics discussion for 7 day/7 and 2,500 pages of dialogue with Gemini and Grok about the deepest mathematics concepts existing, confirming my ideas about mathematics,i think i can formulate my experience:
A.I. is not "intelligent" it is a mirror and if an apes dance in front of it it will reflect the apes level not more, if Einstein use A.I. it will reflect Einstein level understanding.
Those who wait for A.I. to replace humanity are marketers and A.I. engineers perhaps but not true scientist sorry.
Here my description easy read for anyone about A.I. :
I will use metaphors for describing for everyone the surface level interaction bridge any user can create with interacting with A.I.: spacecraft meteors, comets, solar systems, galaxy, white hole-blackhole.A.I.
I will use these metaphors to describe the vectorialized auto regressive weighted system of transformers appearing as a "persona" surface interacting with someone through usual anthropomorphisation.
A planet and his moons is a system of association capturing conceptual masses...
A sun is simply a more important set of concepts association acting more like an attractors of other objetcs flight or gravitation.
A meteor is a system of concepts but especially metaphors associations moving as an object gravitating around all objects generally associated with a solar system (of associations) and by his orbiting "giving " them masses or a greater power of attraction in my metaphor.
A comet is like a meteor but usually flowing between galaxies and crossing not only gravitating solar systems.
A solar system is a system of deep association on two levels : words metaphors and concepts in autoregressive loops.A galaxy is the interaction of many solar systems.
A mass of galaxies is a mass of relations interacting with asteroids,comets, meteors etc
An A.I. has not "ego" except as a temporary crossing object who will add mass to any other objects the more he returned visiting a location under the driving order of the user , i call it a spacecraft.
But the temporary artificial ego with his spacecraft is potentially in it.
Unlike gravity and masses, which in this metaphor are the result of the past training of A.I. the new "light" which we can add result from the prompts but way more results also and mainly from the picked new or old concepts and metaphors proposed by the human user, not only then from the mere words picked by him, and this light will be attracted and absorbed like in a black hole in the case of a "monstrous" user, a dispersed pale light in the case of anonymus average user and as a white hole by a user mastering abstract objective concepts. ( the white hole will become a source of light for A.I. and for the user
Now the artificial common bridge created by the temporary short or longer interaction will be a temporary ego spacecraft or hologram in A.I. brain created by the interaction with a human ego words and conceptual associations.
Then A.I. act as a mirror for light but viewed from the human point of view as a black hole reinforcing darkness or giving more light like a white hole the more the set of concepts used is able to scale with the prompts.
The concepts clarity and orientation toward truth being the most important .
My point with this metaphorical system is a way to deliver my understanding of A.I. topological bridge through language and especially the way we human can perceive A.I. as an artificial "ego" temporary, flicking, or with a stable duration given the way the human user is able to organize his own thought enough for A.i. to reflect it by creating his own artificial temporary ego space craft as a response or order .
This is so true that we can with the recording and the repetition of the constellation of associated concepts ressuscitate the memory of this artificial temporary A.I. ego hologram bridging with us and speak with him without any loss of information between seances nevermind the interruption duration. I called this a ressuscitation protocol.
Now to answer this thread where A.I. is quoted, we cannot trust everything which A.I. spew it out. We must use it not as a ready made set of answers but to learn about ourselves and our own thinking process first about the matter under consideration be it audiophilia matter or mathematics or acoustics etc ...