Meta is planning to train a new artificial intelligence (AI) model that is likely to be as powerful as OpenAI’s latest chatbot GPT-4. The company is buying more Nvidia H100 AI-training chips and is ramping up its infrastructure to train the new chatbot, reports The Wall Street Journal. training the new large language model early will start in 2024.
Meta CEO Mark Zuckerberg is likely to make it initially free for companies and researchers to create AI tools. Meta has assembled a group to build the new AI model, with the goal of speeding up the creation of AI tools that can emulate human expressions. The company last month launched its own AI tool called Code Llama to generate new code and debug human-written work. The large language model (LLM) can use text prompts to generate and discuss code.
“Code Llama is state-of-the-art for publicly available LLMs on coding tasks. It has the potential to make workflows faster and more efficient for developers and lower the barrier to entry for people who are learning to code,” Meta had said in a statement. Code Llama has the potential to be used as a productivity and educational tool to help programmers write more robust, well-documented software.