[detaro]

> could have been performed by the same people.

But they aren't in the example, and Google is understanding their specific situation wrongly/not in an optimal way. Maybe that's not possible to do, maybe the way they are doing it right now is the best overall, but it is not the best interpretation of the specific situation. Which is all I get from the original comment: "Here is something that they could understand better".

Pointing out that implementing this precisely would be difficult because you can't rely on interest subsets as an indicator would have been a fine response. Immediately accusing the poster of being sexist because it obviously is impossible to do is not, unless you have hard data to support that there is no way to tell the difference.

(I'd assume "disjunct" queries from different devices at the same time very often are different people, or at least worth treating as different sessions. But maybe people use multiple devices at the same time more often than I think?)