In a post on the blog last July, Meta General Director Mark Zuckerberg said that “selling entry” in open lama models he has “not (business model).” However Meta does at least OTHER USEFUL Money from Llama through revenue sharing agreements, according to a newly disclosed court’s appearance.
The registration, filed by lawyers for the plaintiffs in the copyright lawsuits Kadrey V Meta, in which Meta stands accused of training her models Llama in hundreds of terabytes of pirate ebooks, reveals that Meta “shares a percentage of income” that companies expect its models generate from users.
Registration does not indicate which specific hostes pay meta. But Meta lists a number of Llama partners in various blog posts, including AWS, Nvidia, Databaricks, Groq, Dell, Azure, Google Cloud and Snowflake.
Developers are not required to use a Llama model through a host partner. The models can be discharged, well adjusted and executed in a range of different equipment. But many hosts offer additional services and tools that make the varnish models function simpler and lighter.
Zuckerberg mentioned the possibility of accessing access to lamp models during a profit call last April, when he also sailed ladder money in other ways, such as through business messages and advertising services. But he did not describe the specifics.
“(I) F You are someone like Microsoft or Amazon or Google and you will basically resume these services, this is something we think we should get some of the income,” Zuckerberg said. “So these are the agreements we intend to do, and we have begun to do it a little.”
Recently, Zuckerberg claimed that most of the meta value is derived from Llama comes in the form of improvements in models from his research community. Meta uses Llama models to empower a number of products across his platforms and properties, including Meta he, Meta he.
“I think it’s a good business for us to do this in an open way,” Zuckerberg said while calling meta profits in Q3 2024. ”
The fact that Meta can generate income in a fairly direct way from Llama is important because the plaintiffs in Kadrey v. Meta claim that Meta not only used pirate works to develop lada, but relieved the violation of “planting”, or loading, these offenses. The plaintiffs claim that Meta used methods of impressive fairy tales to get ebooks for training, and in the process – due to the way the stream works – shared electronic books with other streams.
Meta plans to significantly increase its capital spending this year, mainly thanks to its increasing investments in it. In January, the company said it would spend $ 60 billion-$80 billion in Capex in 2025-ie double the Capex of Meta in 2024- in the database and growth of company development teams.
It is likely to compensate for part of the costs, Meta is said to be considering the beginning of a Meta reconciliation service that will add unspecified skills to the aide.
Updated 3/21 at 1:54 Afternoon: A Meta spokesman showed the text call of profits for the additional context. We have added a quote of Zuckerberg from it – specifically a quota about the goal of meta to share large hosts of Llama models.
