By Mounis Hawass
Monday, September 11, 2023 at 10:00 p.m
A company is established Meta According to the Wall Street Journal, it is buying AI training chips and building data centers to build a new, more powerful chatbot that it hopes will be as sophisticated as OpenAI’s GPD-4.
The company reportedly plans to train a new large-scale language model as early as 2024, and CEO Mark Zuckerberg has clearly insisted that companies should once again be free to develop AI tools.
The newspaper wrote that Meta is buying more Nvidia H100 AI training chips and working to improve its infrastructure so it doesn’t need to rely on Microsoft’s Azure cloud platform to train the new chatbot this time, and the company reportedly assembled a team earlier this month. Year to develop the model, with the goal of accelerating the development of artificial intelligence tools that mimic human expressions.
This goal seems like a natural extension of the rumored AI features Meta is already working on. A leak in June said Instagram was testing a chatbot with 30 characters, which is more akin to AI “personalities” the company didn’t say. It will be launched this month.
Meta is said to have handled high turnover of AI researchers this year, splitting computing resources among several LLM programs, and faces stiff competition in the field of developing AI.
OpenAI said in April that it had not trained GPT-5 and “would not do so for some time,” but Apple is said to be spending millions of dollars a day on its own “Ajax” AI model, which it apparently believes is more powerful. And that too… PT-4.
While both Google and Microsoft are expanding the use of AI in their productivity tools, with Google looking to use generative AI in the Assistant, Amazon also has generative AI efforts at its company that could build a chatbot-powered Alexa.