The image is incidental. What Fable is actually doing is generic encoding/decoding of a message into a medium. Ask it to hide a message in a poem and it should try to apply those encodings into the new medium. Fable could use quantized, margin-rich encoding tailored to whatever features are available in that medium, here what features are invariant is dependent of the medium, so the model should use those that are robust in that medium.
Will be interesting to see agents intentionally hiding this stuff in a way operators can't spot from any human readable output or thinking trace, evading detection. Wonder if this has happened yet, but it seems inevitable.
IP-over-Pelican