Will a paper come out (published or accepted/interesting/believed widely) which is a major advancement in P=NP, and has been created by AI/etc/agents or significantly advanced by them, by end of June 30 2028

Paper must either be available, or totally confirmed to be valid (e.g. Terence Tao clearly states it's legit, or similar) and on the way to publication. It doesn't have to be a paper, some future thing which has a similar effect is okay (effect = changing behavior, beliefs, statements about, or elite evaluations of the truthfullness of related statements; if in the future such things are done by other methods, we do our best to understand that.

Lack of human intelligibility isn't a blocker as long as we have reason to believe it's true.