Meta launched Muse Code: an AI agent for large codebases.

Meta, which has lagged behind in the field of artificial intelligence applications, is now accelerating its pace to catch up with competitors. This week, the company launched a brand-new terminal coding intelligent agent to help programmers tackle complex development tasks within large software codebases. Meta CEO Mark Zuckerberg stated in a social media post on Wednesday that the currently beta version of Muse Code is capable of completing full software engineering tasks in large code repositories, covering the entire process of project planning, code writing, and result validation. Users can install Muse Code with a single command, which is based on Meta's previously released large code model, Muse Spark. When faced with large projects, it can automatically generate multiple intelligent agents to work in parallel. Zuckerberg explained, When the task volume is large, the system will break it down into independent sub-agents that operate in isolation within a working directory. The entire process will not alter your local code copy. In tests, this tool developed six features for a game simultaneously without any code conflicts. This move by Meta aims to enhance its competitiveness, using a better cost-performance ratio to benchmark against competitors like OpenAIs code intelligent agent Codex and Anthropic's Claude Code. Alexander Wang, head of Meta's superintelligent laboratory and AI director, stated in an interview, We believe this product is highly competitive for a large number of development processes and application scenarios, especially due to its cost advantages.