4 stories tagged steganography, newest first.
190 points by gmays | 69 comments
LessWrong post studying how the Fable 5.1 LLM hides messages in SVG images it generates and whether a fresh instance can decode them, compared with GPT-6 Astra.
736 points by ropbear | 641 comments
John Gruber's Daring Fireball post criticizing Anthropic's plan to watermark Claude-generated text via word-choice steganography to comply with EU regulation.
2318 points by kirushik | 692 comments
Blog post reporting that inspecting Claude Code revealed hidden system prompt markers that vary based on the API base URL and the user's timezone.
690 points by paulgb | 185 comments
Blog post by Paul Butler showing how arbitrary data can be hidden in an emoji or any Unicode character using variation selectors, with a decoder demo.