Why did Meta's decision to open-source LLaMA reshape the AI landscape?

The answer

Released by Meta AI in February 2023, LLaMA (Large Language Model Meta AI) was a collection of foundation models ranging from 7B to 65B parameters, trained on publicly available data using insights from the Chinchilla paper. The 13B-parameter LLaMA outperformed the 175B-parameter GPT-3 on most benchmarks. Although initially released only to researchers, the model weights were leaked within a week, sparking an explosion of open-source development. Community projects like Alpaca, Vicuna, and thousands of fine-tuned variants emerged rapidly, democratizing access to powerful language models and challenging the closed-source approach of OpenAI and Google.