Meta Legal Notice Spurs Heretic Project's Recantation of Llama Model Derivatives
In response to a legal notice served by a representative of Meta Platforms, Inc., the Heretic Free Software Project has officially recanted and removed all derivatives of Meta's Llama Artificial Intelligence language models from its weight repositories, prioritizing compliance within the evolving regulatory landscape.
Navigating Compliance: The Legal Context
The Heretic Free Software Project, operating under the direction of 'p-e-w, Chief Heretic,' has publicly acknowledged receiving a formal legal notice from a services provider representing Meta Platforms, Inc. (Meta). This communication was delivered via digital correspondence.
The project issued a statement outlining its immediate compliance measures, stating that it is "recanting" specific materials—namely, derivatives of Meta's Llama AI language models—and has subsequently removed these assets from all controlled model weight repositories.
Technical and Strategic Shifts
Model Repository Management and Derivatives
The decision to remove the Llama derivatives reflects a strategic alignment with current legal and regulatory demands. The Heretic Project emphasizes that while the Llama model family is noted for its high ranking among available language models, the legal concern raised by Meta necessitated this immediate action.
The statement also included a critical commentary regarding the development of the Llama models, noting the legally and ethically questionable circumstances of their creation, amidst ongoing lawsuits and investigations facing Meta in various global jurisdictions.
Infrastructure Diversification and Open Access
Beyond the immediate legal response, the Heretic Project is undertaking significant infrastructure diversification efforts. A key development is the establishment of an official Codeberg mirror hosted in Germany: https://codeberg.org/p-e-w/heretic. Additional mirrors are planned to expand global access.
Crucially, the project is actively developing technological measures designed to ensure continued access to models created under the Heretic banner, independent of reliance on any single service provider. This focus on decentralized infrastructure is central to the project's mission to navigate the global regulatory environment.
The project maintains its commitment to ensuring that Artificial Intelligence development remains "safe, culturally appropriate, and controlled by those who have always known what is best for humanity."