(FILES) A photo shows US technology Meta group logo during the Vivatech technology startups and innovation fair at the Porte de Versailles exhibition centre in Paris, on June 14, 2023. Facebook parent Meta on July 26, 2023 beat market expectations for quarterly earnings powered by a reviving digital ad business. Meta reported a profit of $7.8 billion on $32 billion in revenue during the recently ended quarter, as the number of people using Facebook monthly rose to 3.03 billion. (Photo by ALAIN JOCARD / AFP)
Share on:

Shiva Joshi, Pune

When the top cloud computing providers resell their artificial intelligence large language model, Llama 2, Meta Platforms Inc. expects to profit. 

“If you’re someone like Microsoft, Amazon, or Google, and you’re going to basically be reselling the services, that’s something that we think we should get some portion of the revenue for,” Meta Chief Executive Officer Mark Zuckerberg said Wednesday on a quarterly earnings call. “I don’t think that that’s going to be a large amount of revenue in the near term, but over the long term, hopefully that can be something.”

The company’s tone has shifted since it declared last week that it would give the technology away for free. On July 18, Meta announced a partnership with Microsoft Corporation and Amazon.com Inc. to make its vast language model available for free commercial usage on its partners’ cloud services.

Meta, on the other hand, maintained that making the technology openly available and publishing its source code will result in breakthroughs from a broader developer community.

On Wednesday’s call, Zuckerberg emphasized the advantages of Llama 2, which he continues to refer to as open. Zuckerberg noted that although Meta made the technology open-source, it also stipulated that the biggest cloud providers would need to enter into a business agreement with Meta in order to use it rather than receive a free license.

Large language models (LLMs) underpin text-based chatbots, however, LLMs have additional uses besides supporting AI in question-answering. To enhance a technique for producing replies, the models are given massive amounts of data, generally from the internet.