Recent court documents reveal Meta's earning through partnership deals involving its open-source Llama AI model, contradicting Zuckerberg's previous statements.
Meta's Partnership Agreements and Revenues
According to documents filed in the Kadrey v. Meta case, Meta engages in revenue-sharing agreements with companies hosting Llama models. This revelation contradicts previous claims of pure open-source and non-profit intentions for Llama. While the partners weren't named, expected companies include AWS, Nvidia, and Google Cloud.
Zuckerberg AI Strategy: Openness or Monetization?
Open Llama models, as Zuckerberg claimed, serve for standardization and product improvement, yet emerging data suggests underlying monetization interests. This raises questions about Meta's actual strategy and its willingness to embrace financial trade-offs.
Copyright and AI Training Data
Central to the Kadrey v. Meta lawsuit are allegations of training Llama on data sourced from pirated materials. It is claimed that Meta utilized training data containing hundreds of terabytes of unlawfully obtained books, raising legal and ethical concerns about AI training practices at Meta.
Meta's revenue-sharing disclosure casts doubt over the company's genuine intentions within the open AI models sphere, complicating the concepts of openness and transparency in technology.
