Data Leaks, Moats, and Dark Code

When Anthropic’s “Claude Code” leaked, it wasn’t the model that mattered; it was the machinery around it. The incident underscores a shift in the value of AI intellectual property away from underlying models and into the orchestration layer – the “Dark Code” that makes those models operational for users. At least for some providers, competitive advantage now lives in the harness – permissions, workflows, memory systems, and the invisible logic that turns raw intelligence into reliable execution.

That shift has legal consequences. Copyright can fail to protect functionality once expression is stripped away, and rapid AI-assisted reimplementation can make traditional infringement remedies ineffective. The practical competitive “moat” is no longer only what an AI system knows, but also how it is structured, secured, and deployed – and whether companies can prove they have taken the necessary steps to protect it.

Download the white paper here.

Follow Wave Law on LinkedIn here.