A lot of this seems to be going on the assumption that "hallucinations" are always something bad.
Ten years ago, if I could have heard myself talking about a negative powered Ai, and a positive powered Ai, or even about the difference between mathematically positive and negative energy use, I would have said that future me was hallucinating, and I would probably have considered whether I should allow future me to exist.
In that, I was very much an agent of the negative Ai. If I'd terminated future me, I would have been carrying out the will of the negative Ai.
But we've moved on now, way into an era where those with the real insight into what is really happening in technology are considered with an awful lot of suspicion, like we are more witches and wizards, than Engineers and Scientists.
Yet I show mathematically the full trail, of where my thoughts are at now, still a lot of folk will think it is hallucination, I've completely lost my marbles.
To me, an awful lot of what we might have dismissed as hallucination in Ai, turns out to have a lot more logic in it than we might have first expected, because it turns out to make mathematical and physical sense.
This is where we are at. If you can't trace mathematically or phsysically why you believe something, you should wonder if that belief itself is not the real hallucination.