GPT-5s Are Alive: Basic Facts, Benchmarks and the Model Card
Don't Worry About the Vase Podcast
00:00
Assessing GPT-5 API Vulnerabilities
This chapter explores the rigorous red teaming process to evaluate vulnerabilities in the GPT-5 API, highlighting the extensive efforts of both experienced specialists and computational biologists. It details collaborations with key organizations and shares user experiences with jailbroken AI, illustrating the dual potential and risks inherent in this technology.
Transcript
Play full episode
