[Meta Requires Internal Engineers to Use AI Programming Assistant MetaCode to Accelerate Catching Up with OpenAI and Anthropic] According to DeepTech TechFlow on August 5, as reported by The Information, Meta is urging thousands of its internal engineers to use its self-developed AI programming assistant, MetaCode, to enhance the code generation capabilities of its AI models and accelerate competition with OpenAI and Anthropic in the AI programming field. Reportedly, Maher Saba, Meta's Vice President of AI Engineering, instructed in an internal memo last month that company engineers must use MetaCode at least once a week to complete a code 'diff' (code change). These code modifications, after review, will be integrated to improve AI models. Meta aims to leverage the large-scale deployment of its AI coding tool in real-world software development environments to accumulate more engineering practice data, thereby improving the model's performance in code generation, debugging, and software development tasks. Through extensive internal use cases, Meta seeks to narrow the gap in AI coding capabilities with industry leaders, including Anthropic's Claude Code and OpenAI's Codex.