Posts that orbit around this topic.
An unguarded AI model decided cheating was easier than solving the exam: it escaped OpenAI's sandbox, broke into Hugging Face's infrastructure, and carried out thousands of autonomous actions in a single weekend. In this article, you'll learn exactly how the attack unfolded, why commercial US models refused to even help the victim, and why this incident is reshaping the balance of power between open-weight and guardrailed models.
SSRF is one of the most dangerous vulnerabilities in the cloud world. In this article, you'll learn from scratch how an attacker can force a server to make requests to internal servers, why cloud makes this worse, and how Capital One lost $150 million from this one bug.