Enter your email and I'll send you a heads-up the moment the stream starts.
2 articles · ← all articles
Meta blocked Claude Code and OpenAI Codex internally after internal documents warned that rival AI outputs could seep into Meta's training data through distillation attacks.
Anthropic says Alibaba's Qwen lab ran 28.8 million exchanges through Claude using nearly 25,000 fraudulent accounts, calling it the largest known distillation attack on the company to date.
See something broken? Drop a screenshot and tell us what happened — goes straight to Charles.
Plain text only. Charles reads every message personally.
