No they wouldn’t seek out the new interface. They just bounce.
When they first launched the new interface some new users got the new one and some new users got the old one. The bounce rate for the new one was lower and the signup rate was higher.
I think this thread should be an important lesson for HN devs:
You are not normal. Even the fact that you are likely consuming this website on a computer with its own keyboard, is not as common as it might seem if you surround yourself with other software developers. What you want, and what the majority wants are not the same thing.
How was this measured? Because with the new layout then it’s incredibly easy to accidentally exit a thread (without leaving the site) and you have various dark patterns to force users to sign-up. For instance, whenever I open a reddit link through from a google search then I’m unable to view all comments without signing up, and I accidentally close the thread all the time. I see the exact same thing happen when a twitch streamer I follow visit reddit at the start of every stream (despite visiting reddit every single day using the new layout then he keep accidentally closing threads and getting lost on the subreddit page).
I wouldn’t be surprised if me reading a thread with all its 200 comments on the old version and then leaving the site is considered bouncing, whereas me reading a thread (or rather the 20 comments on the thread you make available) and then accidentally closing the thread to return to the subreddit on the new version is considered successful retention.
I wouldn't say that, and my comment wasn't an accusation.
If you want to be guilty of something, I'd say it was approaching the conversation adversarially rather than charitably. "You were asked for a source" treats the other commenter as an inattentive schoolboy, and coming from a place of curiosity with a question like "But how do you know that's actually what happens?" might have done the job just as well.
When they first launched the new interface some new users got the new one and some new users got the old one. The bounce rate for the new one was lower and the signup rate was higher.