Yes think the replication crisis makes it pretty clear that even with all the fancy software, we still can't remove confounders today (and probably never will).

Yes think the replication crisis makes it pretty clear that even with all the fancy software, we still can't remove confounders today (and probably never will). You can pretty safely translate we removed confounders to: we slightly cooked the data and got this significant result.