The discussion hinges on the premise that if one can perform (2^b) grinding steps for each signature, it becomes possible to leak (b) bits per signature. Consequently, to fully leak a 128-bit seed, one would need (128 / b) signatures. This mathematical relationship underlines the direct proportionality between the number of signatures and the ability to leak bits of the seed, emphasizing the potential vulnerability and the computational effort required to exploit such cryptographic systems.