In a move to catch up with rivals, Meta is focusing on open-sourcing its AI models, despite potential concerns about misuse. The company has been investing in H100 chips needed for training large language models and is expected to rank 2nd globally next year in terms of its number of H100s. Despite a "significant gap" in AI research and investment, Meta is hoping that open-sourcing its AI models will help reduce the lead built up by competitors like OpenAI and Microsoft.
Key takeaways:
- Meta is reportedly developing an AI model that aims to compete with OpenAI's GPT-4, according to unnamed sources cited by the Wall Street Journal.
- The new AI model is expected to be significantly more powerful than Meta's recently released Llama 2 AI and is primarily aimed at businesses for producing text and analysis.
- Meta's CEO, Mark Zuckerberg, is pushing for the upcoming AI to be open source, despite concerns about potential misuse of the AI model.
- Meta has been rapidly purchasing H100 chips needed to train large language models and is expected to rank 2nd in the world next year in terms of its number of H100s.