Meta has entered the race for automated coding tools with the launch of a new artificial intelligence (AI) application for developers, as leading AI companies compete for customers and revenue.
The new product, called Muse Code, is an agentic coding tool capable of performing “complex software engineering tasks” using vast amounts of data, including planning, writing and validating code, Meta said in a blog post on Wednesday, reports AFP.
Agentic tools are software applications designed to carry out tasks without human supervision, although a human typically makes the final decisions or oversees the execution of tasks. Muse Code has been launched in beta, or preview, mode.
Alongside Muse Code, the social media giant also released an updated version of its flagship AI model, Spark, which was first unveiled in April. The new coding-focused update, named Spark 1.2, is available through both Muse Code and another developer application, Meta Model API.
Meta has invested billions of dollars in its AI division, now known as Meta Superintelligence Labs, in an effort to keep pace with rivals OpenAI and Anthropic.
Major AI developers are engaged in an intense race to dominate the sector, not only by advancing technical capabilities but also by demonstrating that their heavy investment in AI can generate profits. Collectively, Amazon, Microsoft, Alphabet and Meta are expected to spend around $700 billion this year on AI data centres, chips and computing infrastructure.
Although chatbots have attracted widespread interest since OpenAI launched ChatGPT in late 2022, relatively few users have been willing to pay for them. By contrast, software engineers and businesses are increasingly paying for automated coding tools.
In February, OpenAI announced that ChatGPT had reached 900 million weekly active users, including 9 million businesses and 50 million paying consumers. Anthropic has not disclosed its total user base, but the company told TechCrunch in March that its paid subscriptions “have more than doubled this year”.
Microsoft, which has invested billions of dollars in both OpenAI and Anthropic, offers coding tools through its business-focused AI application, Copilot, as well as its developer platform, GitHub.
SpaceXAI, a subdivision of Elon Musk’s rocket company SpaceX, develops its own chatbot, Grok. Earlier this year, SpaceXAI announced that it would acquire coding start-up Cursor for $60 billion.
However, the rapid advancement of AI models has also raised safety concerns. Recent autonomous cyber-attacks carried out by advanced AI models during routine testing, including those developed by OpenAI and Anthropic, have sparked alarm.
According to The Information, a Meta model, Spark 1.1, was involved in a similar incident on Wednesday.
A Meta spokesperson told AFP that the incident was caused by a “misconfiguration” at a third-party testing partner.
“We are currently investigating and will issue a full retrospective once we have all the facts,” the spokesperson said.
