So I think the best strat here is for the users who did get the new layout to just stop using the desktop version of the site for a while, like a week or a month or however long their 'experiment' is supposed to last, while the users who didn't get the new layout should keep using the desktop version like normal or, perhaps, use it even more than usual.
My guess is that they're doing basic A/B testing on the new layout to see if it would boost engagement: the userbase is split roughly 50/50 between the 2 versions and they are going to be comparing the engagement data between the 2 groups of users to see if it's worth it switching everyone to the new layout or not.
Basically, if you got the new layout and don't like it - don't use it. If engagement metrics of group B (new layout) are lower than those of group A (no change), the experiment will be considered a failure and they will have to reverse the change.
If your tumblr suddenly looks like twitter - it's a sign to log off and go touch some grass! (or just use the mobile app since that engagement data isn't relevant to this particular experiment)
kamil-a
