AI is getting scary good at finding hidden software bugs - even in decades-old code

The ZDNet piece explains that modern AI models can effectively audit even very old, obscure code—surfacing long-dormant bugs—while warning this capability expands the potential attack surface because bad actors could use the same models to find and exploit vulnerabilities in unpatched or unsupported firmware and binaries; it also notes LLMs can complement traditional static analysis but are not yet ready to replace human security review.