The Neverending Story of the Eternal Wormhole and the Noisy Sycamore

Abstract

There has been a great buzz surrounding Daniel Jafferis et al.'s latest Nature paper, "Traversable wormhole dynamics on a quantum processor". The Nature paper discusses an experiment in which Google's Sycamore quantum processor is used to simulate a sparse N = 7 SYK model with 5 terms (a learned Hamiltonian). The Nature paper shows that the learned Hamiltonian preserves the key gravitational characteristics of an N = 10 SYK model with 210 terms and is sufficient to produce a traversable wormhole behavior. I will examine the experiment and discuss some philosophical challenges concerning the experiment in memory of Ian Hacking. Recently, Norman Yao and two graduate students discovered multiple flaws in Jafferis et al.'s learned Hamiltonian and uploaded a comment on the Nature paper. As expected, Jafferis and his team found a simple way to clarify the misunderstanding. They found a physical justification that allowed them to avoid the problem. In this paper, I elucidate the main arguments Yao and his students raised and the way Jafferis et al. found to save their learned Hamiltonian. I will end this paper with a philosophical comment on this recent development in the context of the learned Hamiltonian.