This reads like a lunatic thesis, but in reality it is not far-fetched. With AI all things are possible. Major change is so rapid recently that even the timelines are not unrealistic. Co-author Tsimerman was recently hired by OpenAI to work on "AI safety."
formatting link