Mark Zuckerberg unveils Muse Code AI, says it can do all software engineering tasks

Mark Zuckerberg unveils Muse Code AI, says it can do all software engineering tasks

Meta CEO Mark Zuckerberg has launched Muse Code, a new AI coding agent that he says can handle complete software engineering tasks from planning and coding to testing. The launch comes as AI leaders predict that traditional coding tools and even some software engineering work could soon be taken over by AI agents.Mark Zuckerberg launches Muse Code AI, says it can do all software engineering tasks(File photo: ITG)Meta CEO Mark Zuckerberg has introduced Muse Code, a new AI coding agent that he says can take care of complete software engineering tasks instead of simply helping developers write code. The tool is now available in beta and is powered by Muse Spark 1.2, Meta's latest coding-focused AI model. The launch comes as AI companies race to build smarter coding assistants that can work on entire software projects with little human intervention. Muse Code will compete with similar products such as Anthropic's Claude Code and OpenAI's Codex, which are also designed to automate software development.Announcing the beta launch on X, Zuckerberg wrote, "Releasing Muse Code in beta today. It's a terminal coding agent that takes on complete software engineering tasks across large repos: planning changes, writing code, validating the results. Powered by Muse Spark 1.2, a coding-focused model update."Unlike AI assistants that begin every task with no memory of previous work, Muse Code keeps specialised AI agents running throughout a developer's session. According to Zuckerberg, these background agents gradually understand the project better as they continue working, allowing them to handle larger and more complicated codebases. A screenshot of the announcement made by Mark Zuckerberg. Zuckerberg says this is just the beginning, more is comingFor bigger assignments, Muse Code can split work across multiple AI sub-agents. Each one works independently in its own isolated workspace, which means developers' existing code remains untouched while several tasks are completed at the same time. Zuckerberg said Meta tested the system by asking it to build six game features simultaneously without any collisions between the AI agents. Meta is also pitching Muse Code as a reliable tool for long-running projects. Every model request, tool call and code edit is recorded in a local event log before execution. If the system crashes halfway through a task, it resumes from the same point instead of requiring the developer to start over.Zuckerberg also shared results from an internal benchmark, saying Meta used Muse Spark 1.2 for a kernel optimisation task running on NVIDIA Hopper hardware. According to him, the model made more than 1,000 tool calls over 24 hours and continued finding performance improvements long after the initial exploration phase. The company says Muse Code is easy to install with a single command and is available through its contributor tier. Zuckerberg added that Muse Spark 1.2 is only the beginning, with larger and more capable AI models already in development.Claude Code creator says software engineering roles will be deadMuse Code arrives at a time when AI company leaders are making bigger claims about the future of software development.Earlier this year, Boris Cherny, the creator of Anthropic's Claude Code, said the software engineer role will begin changing rapidly as AI agents become capable of handling almost every part of the development process. He also suggested that developers may stop relying on traditional coding environments much sooner than many expect. He had already once said that all software engineering roles will become obsolete starting this year, and it now seems that Mark Zuckerberg is trying to make that a reality.Speaking during a talk at Anthropic, Cherny said, "There's a good chance by end of year people aren't using IDEs anymore." He was referring to integrated development environments (IDEs) such as Microsoft's VS Code and Apple's Xcode, which have been standard tools for developers for years.According to Cherny, Anthropic deliberately built Claude Code as a terminal-based tool instead of creating a graphical interface because the company believes AI models are improving so quickly that spending time building traditional coding interfaces may not make sense.He also said Claude Code is already capable of writing complete features, fixing bugs and navigating large codebases with minimal human guidance. Cherny even claimed that he has not manually changed a line of code since November 2025 as he is relying on AI agents to do the work instead.- EndsPublished On: Aug 6, 2026 12:30 IST

Original Source

Read the full article at Indiatoday →

KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.