Meta is making significant progress in the AI sector, particularly in coding technology. This week, the company unveiled a new coding assistant geared toward developers needing help with complex tasks within extensive software codebases.
The tool, named Muse Code, is currently in its beta phase and can handle “comprehensive software engineering tasks across large repositories,” according to Meta’s CEO Mark Zuckerberg, who shared this via social media on Wednesday. The capabilities include tasks like “planning modifications, writing code, and validating outcomes,” he noted.
Muse Code can be set up with a single command and is built on Meta’s previously established coding model, Muse Spark. It manages substantial projects by deploying its own agents that work in tandem.
“When the task is substantial enough, it breaks down into distinct sub-agents that operate simultaneously in isolated workspaces,” explained Zuckerberg. “Your working copy remains untouched. During tests, it managed to develop six features for a game at the same time without any issues.”
This initiative is aimed at enhancing Meta’s competitive stance and affordability compared to AI entities like OpenAI with its coding assistant Codex, and Anthropic’s Claude Code.
“We believe this can be an excellent option for various workflows and use cases, especially in terms of cost,” remarked Alexandr Wang, Meta’s AI chief who oversees Meta Superintelligence Labs.
Meta has been intensifying its efforts in the AI domain by investing heavily in development. In June, the company branched out from its primary focus on using AI in advertising to also tap into the enterprise AI sector with an assistant designed for customer service and support.
