the (new to me) knowledge that this vulnerability was known already actually make the hype cycle a little concerning to me (for the part of the tech journalists, i mean)
I tend to engage that this question "is this software more{of less} secure?" with a more grounded statement — that depends on the software, if the software is largely vibe vomited it is less secure because the devs have no 'institutional knowledge' of the code, so it will take way way longer to fix it [if they try to fix manually], if they fix with llm it might make things pile up vulns on top of vulns that no one in the project will know how to fix it. the main problem is not the amount of exploits, but the inability to fix it (because skill decay for extended usage of llms are a thing, and also executive function decay [like brain processing and interpersonal awareness, the latter is a thing devs already struggle a lot without llm already]) —