Meta Accused of 'Bullying' Open-Source AI Community

• Meta's Llama models don't meet Open Source Initiative's new AI standards
• OSI requires full disclosure of training data, source code, and model weights
• Meta argues for 'spectrum of openness' due to LLM development complexity
• Llama models downloaded 300m+ times, setting Meta apart from proprietary LLMs
• Concerns raised about Meta 'abusing' open-source term for regulatory advantages

Source: economist