TLDR: I think we have too many examples, in particular in

• clustering
• ensemble
• generalized-linear-models This is the worst.
• model-selection
• nearest-neighbors
• support-vector-machines

On top, several examples don't render a plot in the gallery, e.g. Release Highlights 1.0.0.

I would like to discuss how we can improve this situation. My main worry is that, currently, a user does not find what she/he is looking for but gets overwhelmed by the sheer amount of examples.

Edit: See task list in #27151 (comment).