Claude Code is secretly watermarking your prompts
What happened: Anthropic’s Claude Code is embedding hidden markers in user requests, a technique known as steganography.
Why it matters: This sneaky move aims to trace outputs back to their source, which could help curb misuse but also raises eyebrows about privacy and consent.

Claude Sonnet 5 hits the stage
What happened: Anthropic released Claude Sonnet 5, boasting improved reasoning and context handling.
Why it matters: With claims of better handling complex tasks and longer conversations, Sonnet 5 could be the AI that finally understands your Shakespearean insults.

Commerce Dept. gives Claude Fable and Mythos the green light for export
What happened: Export controls on Claude Fable 5 and Mythos 5 have been lifted by the Department of Commerce.
Why it matters: This move opens up new markets for Anthropic, but also raises questions about how these models might be used (or misused) in different regulatory environments.