Means nothing at all. These people confuse formal verification with bug free results. They think they can fully translate ambiguous language software requirements into formal mathematical propositions. They think we have a formal language for software specifications. They think that formal verification of large complex systems with multiple external dependencies is possible and maintainable across software updates and upgrades. They think we don't have already formal mechanisms to check for implementation errors and domain errors in the form of syntactic analysis tooling and unit tests, and comprehensive runtime analysis of software in the form of feature tests, end-to-end tests, deterministic test systems, ... They think formal verification doesn't require formal validation and no peer review of the mathematical proof is required by external parties. They think formal verification is only possible now because of AI, that we haven't done that before, that we haven't realized the difficulty is translating real-world requirements, that we haven't measured results, that we haven't understood the cost-benefit relationships... In short, that AI eliminates all of it and that we can reach to that conclusion without any single evidence so far of a single large system being successfully verified and maintained. It's pea-sized brain discourse.