A lot of these new AI found bugs are proving to be nothing burgers. Just a waste of money to try and hype the latest models.
They're either in old code not actually used anymore or miss a system interaction that fixes the supposed bug or just straight up are wrong.
LLMs are shit