The whole advantage of threads is that they introduce slop back into the frame, allowing you to essentially align them on installation of the BB.
Pretty much every non-bike application of bearings has them pressed right into the related structure like BB30. This is the way to do it if you can achieve the required precision. That said, I highly doubt this is why you liked those bikes better, and I doubt you could feel a difference in BB stiffness at the BB thread-to-shell interface. That said, it's possible the ER: One is noticeably more flexible under pedaling than other bikes due to broader design choices, and that would be a sign that something has gone wrong with the design. While I doubt that, they did get some criticism for the AR: One being too stiff.
Overall, people need to chill on the ER: One. As far as I know, there is one rider on aero test that shows it's faster than an S5 with a slightly different body position. This is basically meaningless except to that one rider. I have a lower CDA on my round tube Ti "progressive geo" bike than my acquaintance, who is the same height, on his Noah Fast using the same bars and wheels. Does that mean my custom bike is "more aero"? Of course not, though the bikes have similar stack and reach, position matters a lot when it comes to aero and a few mm here or there can make the difference.