So for a directed graph with positive weights, this beats Dijkstra!? This would be an amazing result, right?

Or is this one of those Algorithms that is only meaningful with impractically small, large, or weird data?