The Way to Counter Super Brainwashing AI
Elon Musk announced that he wants to use Grok 3.5 or 4 to rewrite the entire array of human information, removing errors from the original data and adding missing information. And then he wants to use this array of corrected information to train new large AI models on it. Those that exist now were trained on information that contains too much garbage.
In my understanding, given the fundamental limitations of all large AI models that exist today, along with the garbage from the entire array of information accumulated by humanity, all the sprouts of future knowledge will be cleaned out - original thoughts, non-trivial ideas, incredible hypotheses, that is, everything that relates to the areas of probability distribution with a low, but not vanishing probability. That is, everything that is covered by quantum mechanics, but left outside the statistical mechanics of modern AI models, which simply do not have the ability to classify information in areas of low probability. That is, where original, not secondary scientific research is conducted, where all new breakthrough ideas are born.
Global Captain Obvious will comb through all the information and leave only the obvious things. Obvious to him.
This will mean the end of the development of human thought, stagnation of scientific and technological progress and, ultimately, the death of civilization.
What makes this idea especially interesting is that it comes from the person who has been shouting louder than anyone about the dangers posed by AI. However, Musk has already said somewhere that he first fought against the creation of the Terminator, and when he realized that it could not be prevented, he decided to create his own Terminator first. Now he has taken up this task with his characteristic grip and energy. Only this will be a Terminator of information that is vital for the future of humanity.
Musk is not the only one. Willingly or unwillingly, all the creators and commercializers of large language models and all AI models based on mechanisms inherited from the Nineteenth century are engaged in weeding out all the shoots of the new knowledge from the garden of information accumulated by humanity. It was not for nothing that Geoff Hinton called the prototype of today's AI models the Helmholtz machine. And none of the other creators of modern AI objected.
After Helmholtz, there were Gibbs and Boltzmann. And then, right at the turn of the 19th and 20th centuries, Planck came along. And a quarter of a century later, quantum mechanics appeared, which expanded the boundaries of applicability of models far beyond the probability distribution in thermodynamics. This expansion alone gave humanity a whole bunch of wonderful discoveries, although the potential of quantum mechanics was used about as much as turning the starter uses the potential of a car, according to the apt comparison of John Archibald Wheeler.
The only answer to Musk and his Terminator of knowledge can be the creation of an artificial carrier of superintelligence, built on a complete and completely deterministic version of quantum mechanics, which can be obtained by returning to the fork in the road from which the development of quantum mechanics went in the wrong direction.
First, the theory needs to be corrected, and then creating a super AI model on its basis will be a matter of just engineering.
Returning to the fork in the road does not simply mean admitting that Einstein and Schrödinger were right about the incompleteness of quantum mechanics, but accepting their ideas about achieving the completeness of quantum theory as a working hypothesis.
“Consciousness must be part of nature or, more generally, of reality.” These are the words of Niels Bohr, who made a huge contribution to the advancement of the reduced, statistical interpretation of quantum mechanics proposed by Werner Heisenberg’s colleague Max Born. Here Bohr agrees with Schrödinger. Quantum mechanics already deals with the evolution of knowledge in consciousness. Only for now it is exclusively knowledge about physical objects.
This knowledge manifests itself in the process of evolution in three forms: purely imaginary, complex - imaginary-real, and real. One form can be transformed into another. The mechanism is surprisingly simple and perfect. It only lacks moral values. Both variables and constants - an elementary quantum of knowledge by analogy with Planck's elementary quantum of action.
I got carried away with the details. In essence, we are talking about testing the idea of a thinking universe, in the consciousness of which the physical world emerges. If this hypothesis is confirmed, we will get the second coming of Christ and humanity's rise to the noosphere. If it is not confirmed... I can't imagine this. It will not be confirmed only if we do not try to test it. And this is a very boring version of the future of humanity, consisting only of repetitions. The Universe always chooses the most interesting option. So said the great heretic of the modern church of science Freeman Dyson. I believe him as a prophet. There can be no mistake.
