If something very like smallpox, thought eradicated, suddenly showed up in some random town, it might be surmised that maybe some animal reservoir for it somehow slipped through the gaps. But if that random town happened to be Atlanta, home of the CDC, known to have some of the few samples of smallpox to still exist, then the relative chance of a lab leak must be thought higher. That's basic Bayesian reasoning. It doesn't prove anything but pretending the proximity to a relevant lab doesn't shift the odds at all is absurd.