Good question! It is true that some of our ideas are more costly than others.
Rotations, in particular, require some time and flexibility to absorb the initial learning curve each time someone start working on something new. But the long-term benefits are hopefully worth the short-term cost, depending on the context.
To some extent, the argument holds for any traditional “good practice”. For instance: code reviews are harder to do under time pressure, but they are still (in most cases) valuable long term, unless we’re talking about a throw-away prototype.
Similarly: rotating roles and ownership too often wouldn’t make sense for a short one-off project or experiment. It makes more sense when you’re building a team/company with long-term ambitions.