Hacker News (AI, 50+ points)Article
Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)
This is a real security issue in a production inference service. Multi-tenant systems that share KV cache are vulnerable to context leakage, which means user data or prompts from one request can bleed into another's response. If you're running inference at scale, especially on shared infrastructure, audit your cache isolation now. This isn't a theoretical risk anymore.