It sounded like dataflow was doing a Fermi estimate, which usually need to be within one order of magnitude of the real answer to be considered decent. I think

It sounded like dataflow was doing a Fermi estimate, which usually need to be within one order of magnitude of the real answer to be considered decent. I think the reason why they didn’t like the headline was because it seemed like it was several orders of magnitude off.