The math lane: four results, two of them negative
We run a DePIN data business and a small formal-methods and number-theory lane. This is the honest ledger of the second one: an Erdős problem where we were right and second, a game-theory case we verified on our own hardware, a Riemann route that produced two measured laws and no progress on the conjecture, and a proof corpus where machine-checking refuted five of our own stated properties.