Meta aims to develop an AI model that could rival or surpass OpenAI’s GPT-4. The company plans to train its Large Language Model (LLM) early next year.
More take-aways:
- Infrastructure: Meta has been acquiring Nvidia H100 AI training chips and bolstering its internal infrastructure. Unlike before, it won’t rely on Microsoft’s Azure cloud platform for training.
- Human-like Interactions: A team within Meta has been working to create a tool that can closely emulate human expressions. The goal is to integrate chatbots into platforms like Instagram, offering users a choice of “30 AI personalities.”
- Future Outlook: Meta intends to make its LLM freely available for other companies to build upon. The company is investing time and effort to ensure the AI model can replicate human expressiveness effectively.
