Meta Platforms is set to release a new version of its artificial intelligence large language model, Llama 3, in July, according to 'The Information'. The upcoming Llama 3 aims to address a perceived issue with its predecessor, Llama 2, which was criticized for being too 'safe' in its responses to contentious topics. This move comes as Meta seeks to enhance the robustness and applicability of its AI models across real-world applications, particularly in generating diverse adversarial prompts. The development reflects a broader industry trend where companies like Google have faced backlash for their AI chatbots, such as Gemini, avoiding sensitive subjects. Meta's initiative with Llama 3 is part of its effort to make its AI more open to difficult questions, a response to feedback from senior leaders and researchers. The release is seen as competition for winamp.