Training for the new model will start next year
As The Verge reports, citing The Wall Street Journal, Meta* is buying up AI training chips and building data centers to create a powerful chatbot that it hopes will be as technologically advanced as OpenAI’s GPT-4 .
The company plans to begin training the new language model in early 2024, and CEO Mark Zuckerberg is pushing to make it free again for companies to build artificial intelligence tools.
Meta* is purchasing Nvidia H100 chips for its next-generation
The magazine writes that Meta* is purchasing large quantities of Nvidia H100 chips for artificial intelligence training and is expanding its infrastructure so that this time it won’t have to rely on Microsoft’s Azure cloud platform to train a new chatbot. Earlier this year, the company assembled a team to create a model with the goal of speeding up the creation of artificial intelligence tools that can mimic human expressions.
Meta* has reportedly experienced high turnover of AI developers this year due to the division of computing resources between multiple projects. Meta* also faces stiff competition in the generative AI space. In April, OpenAI said it was not training GPT-5 and “won’t be doing so for some time,” but Apple is reportedly pouring millions of dollars daily into its own Ajax artificial intelligence model, which it plans to make more powerful than even GPT-4.
* The Meta company is recognized as extremist in Russia and is banned.
