@rdviii @dabacon @jenseisert You mean providing a quantum speedup asymptotically? It remains unproven but partial evidence is usually attributed to the presence of entanglement. I am not sure a single author can be attributed this but the first papers on classical simulation of non-entangled circuits that come to mind are Jozsa - Linden, Vidal and Markov - Shi. This maybe was known as well by many people working on tensor networks.