Meta's Muse appears to use an OpenAI model labeled muse-special

A recent investigation into Meta's Muse project has uncovered evidence suggesting the integration of an OpenAI model, specifically identified in code as 'muse-special'. The discovery, which originated from technical analysis shared on Hacker News, points to potential cross-pollination between major AI research labs. While Meta has positioned Muse as a proprietary generative model, the presence of these specific labels has sparked significant debate within the developer community regarding the provenance of the underlying architecture and training data. Observers note that such findings raise questions about the transparency of model development and the extent to which proprietary systems may rely on external APIs or pre-trained weights from competitors. As of now, Meta has not issued a formal statement clarifying the nature of the 'muse-special' designation or its relationship with OpenAI's infrastructure, leaving the tech community to speculate on the implications for future AI development and intellectual property standards.