Anthropic Warned of Open-Source AI Security Risks Since Mythos Preview, but Open-Source GLM 5.3 Shows Strong...

Anthropic has warned about open-source AI security risks since Mythos Preview.
But the open-source AI model GLM 5.3 is good at cyber capabilities.
Hamza () explains the irony in the open- vs. closed-source AI safety debate, citing incidents like Hugging Face.