Or someone simply noticed the lid does more than keep the liquid in.

I have made more progress by simply eliminating parts that it could not be, logistical analysis of key space and just some plain hands-on common sense. Real-world problem solving applied to reducing the key space by logical means. All keys with repeating prefixes, all with more than two repeating characters to list some already mentioned before. I am actively seeking puzzle solutions personally.

I have only been working on these puzzles for perhaps a year with direction and structured aim. I can only imagine what stage most of you all are at if far beyond my new presence is stumbling blindly through.

I do work with ML, tinker with special custom models in attempts to bridge the rational logic beyond the simple prediction, transformers or Berts.

Someone with the knowledge and means to code bug free model training and compile datasets upon which to aim these models.... now that, seems more likely as those able and with enough understanding could be leveraging this emerging technology which is gaining ground and establishing a stable foundation anyone dedicating the time can begin doing in some garage setup.

I'm sure after removing all those false positives. The reduced key space hardly takes a "hidden mathematical genius".

Though all my work is done, on a single graphics card on a laptop. I believe any solutions happening at this moment is due to the collimation of over four years of continual persistence and well mapped logic in isolated sectors, in calculated sub-ranges.