how unreasonable is it to think that during the ai assited claude bruteforce process, it bruteforced also the kernel and built on a randomly found logical inconsistency
Highly unlikely. The lean kernel has been heavily scrutinized
unreasonable
very, due to the precautions explained in the link. For example, they used multiple kernels with different implementations etc.
As unreasonable as assuming all human's peer reviewed proofs to be incorrect due to an error collectively missed by all humans.
This is covered in point 4 of the linked post. The exported raw proof has been checked by several different kernels, not just one.