19/11/2025
On AI Bubble
Currently AI is changing the world in many areas, including Go game, protein structure, and Generative AI areas.
That being said, there are some areas for AI to take many years to contribute.
For instance, superconductivity, called macroscopic quantum phenomena, will be very hard for AI to contribute, because superconductivity is due to the quantum mechanical interactions between Avogadro number (10^23) of electrons and Avogadro number of ions, which is impossible to solve exactly, unless we know the answer from experiment or ingenious insight. Specifically high Tc cuprates of Tc about 100K, discovered in 1986, still intrigue scientists for its superconducting mechanism. So, the superconducting mechanism of our room temperature ambient pressure superconductor, CES-2023, is out of the question right now, unless we find the clues from experiments.
For superconducting quantum computers, AI will have limited contributions, because we need to produce scalable superconducting qubits first, which is still on-going. Note that Dr. Kim revised the theory of the Josephson effect, guiding principle of the superconducting qubits, by inventing Cooper pair wavefunction approach [1]. For instance, he determined the maximum thickness of the insulating barrier from “Threshold Resistance”, found from experiment, which was not possible to calculate by previous theory.
The bottom line is that AI can be innovative and efficient in some areas, but in other areas it may need many years of learning and training to contribute.
[1].
