Yann LeCun, the chief of Meta AI was recently on a podcast with Lex Fridman talking about Meta, AI, and AGI. He revealed that he recently met with an Infosys founder who is funding a project based on Llama 2.
“He’s funding a project to fine-tune Llama 2, the open source model produced by Meta so that it speaks all 22 official languages in India. It’s very important for people in India,” said LeCun. He did not reveal which co-founder he met.
LeCun was quite impressed by the recently released Kannada Llama. “I love this. This is why open source AI platforms will win: it’s the only way for AI to cater to highly diverse languages, cultures, values, and centers of interest.” he wrote on X, reposting Kannada Llama post.
He also spoke about Moustapha Cisse, who used to be a scientist at Meta FAIR who is also using Llama 2 to build language models in Africa, and catering it for medical use.
LeCun is a big proponent of open source AI models and recently also posted on X about the training infrastructure of Llama 3, the upcoming AI model of Meta which is expected to be open source as well.
Infosys was one of the first companies that had invested in OpenAI back in 2015. It donated around USD 1 billion to OpenAI, which was a non-profit back then. Infosys, along with Elon Musk, AWS, YC Research, among others, joined hands to make the USD 1 billion donation to OpenAI.
Apart from Infosys, a lot of Indian companies and startups have been leveraging Llama 2 and making Indic models such as Tamil Llama, Kannada Llama, Telugu Llama, and many more. IIT Bombay’s in work BharatGPT project is also focusing on catering to the 22 Indic languages and is pushing towards being open source.