Well, yes and no. In its current state, if proven right, the paper is a mathematical breakthrough, although I'd award that to OpenAI's research regarding the disproval of Exponential Time Hypothesis (ETH) being false. Most of modern cryptography is strongly based on ETH as an assumption being true. While the research proves the time complexity is non exponential, it is still too high of a number to become an effective exploit tool. Well, since ETH has been disproved, if someone or something (swarm of agents) can find a more optimal algorithm than what OpenAI came up with, then we have more reasons to be worried about. And yes, if that ever happens, crypto coins / tokens including but not limited to XMR will be in deep trouble.