a few days ago, praetorian released Augustus, a penetration testing tool specifically targeting LLMs (your ChatGPTs and Claudes used in AI)
we are living in a new age where almost every support chat is run through an AI, allowing frontend users to potentially connect to backend systems
nobody can argue against the convenience AI gives us, but perhaps we should start thinking about the security we need to give it
https://t.co/C9lyeA6rwE
last chance to change your gpt4 models
on 13th feb, openai will be sunsetting their controversial gpt4 series, update your codes if you're still using them
do you think AI companies should just release something like a "gpt-latest" alias for us to use?
or is it too hands-off?