New lower values for p get discovered all the time (maybe once a year). It is conjectured that they will approach 2.0 without ever getting quite to it. Somehow Quanta Mag heard about the new result ...
Keep in mind all the agent "sees" is a sequence of tokens and unless the model is specifically trained to be able to interpret that specific kind of spatial reasoning -- to line those up in columns -- ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback