Right, it's not solvable. But as I stated in the other reply, it's also unnecessary in the vast majority of systems - either failure is not catastrophic, or you

Right, it's not solvable. But as I stated in the other reply, it's also unnecessary in the vast majority of systems - either failure is not catastrophic, or you only need consensus about the past, or the decision is not executed in a shared synchronised matter, or some other property that means you don't actually need to solve the actual Two Generals Problem. I really think we need to stop citing it as a "thing", it's like the Halting Problem, stupidly strong and not necessary most of the time. Saying that something "solves" Two Generals is diverting the conversation into pointless impossibilities.