In addition to the chatbots, Meta is also developing an AI system that could potentially surpass the capabilities of OpenAI's GPT-4. The company plans to begin training this large language model in early 2024, with the aim of creating a more sophisticated AI chatbot. Meta has also purchased more Nvidia H100 AI-training chips and is working on improving its infrastructure to reduce reliance on Microsoft’s Azure cloud platform for training the chatbot.
Key takeaways:
- Meta is reportedly planning to release its AI chatbots as soon as this week, with the aim of driving user engagement and collecting more user data for targeted advertising.
- The chatbots will come with multiple personas, including a 'sassy robot' persona called Bob the Robot and an 'overly curious' persona called Alvin the Alien.
- Meta is also working on a large language model (LLM) AI system, which it hopes will be more powerful than OpenAI's GPT-4, with training set to begin in early 2024.
- The company has purchased more of Nvidia's H100 AI-training chips and is improving its infrastructure to reduce reliance on Microsoft’s Azure cloud platform for training its upcoming chatbot.