Will textbooks Quantum Computing work ?
Is the quantum entanglement physical ? What can we say today?
First, the EPR-Bell lab proof is not rigorous. The previous lab experiments are ruled out by our simulation using classical software.
Moreover, I found “A Sum-Over-Histories Account of an EPR(B) Experiment“, a 1991 theoretical analysis from Sukanya Sinha and Rafael D. Sorkin, showing it analytically, in an elegant way, with classical physics tools.
Thus, these results are merely ignored by the quantum computing researchers. They believe that magical superposition and entanglement will lead to an ever seen parallelism. It is always possible that it works but it is unlikely to happen, according to the publications. We are structurally far from the expected architectures ; parallelism is still classical ; superpositions need still Monte-Carlo methods where the minimal number of tries depends on the complexity.
Now, we must appreciate that so many creative people work in this sector… They highlighted the annealing methods with or without correlations based perturbations. In a few years, standard computers will use processors augmented by new hardware functions facilitating Monte-Carlo algorithms and parallelized new bit operations.
But also and mainly, they conceive hundreds new quantum settings giving enough material for further theoretical global analyses, opening doors to new physics.
What are the alternatives to treat hard problems?
Traditional hardware running clever softwares.
Annealing, low programmation languages for critical routines, new compile time optimizations methods and better runtime cuts lead to faster execution. A gain of 75% implies 4 times less computers, energy and space. For 90%, it is 10. It is not important for most applications but useful for hard problems like big data analysis or integer factorization. Even if we don’t know how to factorize all the 2048 bits numbers, we succeed in 9.++ cases on 10, a part growing each day. This will help to discards a lot of certificate private keys even if half of them were already known to be weak. For the Steiner problem, our very inspirative toy model of hard problem, we get often and quickly perfectly optimized large sets while checking the results take months on our small cloud.
It remains to apply the methods at a more abstract level and to defend the standard implementation of new hardware functions. You will be astonished before 2022.