Urgent.News

the world's headlines, one feed

Editions

Tech

Meta’s new coding agent is cheap (but it’ll cost you your data).

Meta this week announced Muse Code, its first coding agent designed to tackle complex software engineering tasks, like planning changes, The post Meta’s new coding agent is cheap (but it’ll cost you your data). appeared first on The New Stack .

Meta’s new coding agent is cheap (but it’ll cost you your data).

Meta recently unveiled Muse Code, an AI-powered coding agent that aims to streamline complex software engineering tasks such as planning changes, writing code, and verifying results. Powered by Muse Spark 1.2, the agent has recently been released alongside an update, Muse Spark 1.1, which improves code generation, debugging, and end-to-end developer workflows.

Meta has positioned Muse Code as a more affordable alternative to its competitors, such as Anthropic's Claude Code and OpenAI's Codex. However, despite the lower price, many software leaders remain hesitant to adopt the tool, citing concerns about data privacy and security. Ken Ringdahl, CTO of Emburse and advisor to Snyk, expresses his concerns, stating that "Security, data privacy, [and] data security is paramount, and it’s not something we will trade off for a lower-cost option."

Similarly, Yunhao Jiao, co-founder and CEO of TestSprite, points out that the data-sharing requirement could expose more than just the code, as it includes developer prompts, agent responses, and corrections. Meta's pricing for Muse Code follows a similar model to Muse Spark 1.1, which charges $1.25 per million input tokens and $4.25 per million output tokens.

The company has also introduced a contributor tier for Muse Code, offering a significantly lower cost for users who opt-in to help improve the model by sharing their code. However, this data-sharing requirement has proven to be a significant deterrent for many software leaders, as they are unwilling to compromise on their intellectual property (IP).

Meta's foray into AI coding agents is part of its broader strategy to compete with OpenAI and Anthropic. Earlier this year, Meta released Muse Image, an image-generation model, followed by Muse Spark 1.1, its first paid AI model, indicating a shift towards proprietary AI models with aggressive pricing. Meta is reportedly seeking requests for zero-data retention to cater to companies seeking more control over their data.

Currently, Muse Code is available in beta, but its data-sharing requirement appears to be a significant hurdle for many potential users, who are wary of trading their valuable code for a lower price.

Written by urgent.news from The New Stack's reporting — not their text. Machine-written — it may contain errors, so check the original before relying on it.

Read the original at thenewstack.io →

More in Tech

Postgres Full-Text Search in Production: How to Load-Test the Index and Pin Down Relevance

Adding full-text search to Postgres is a two-line migration. Running it in production without surprises is not. The GIN index that makes search fast also adds write amplification and a background…

  • GIN index accelerates searches but increases write amplification in PostgreSQL.
  • Load testing required to validate search capabilities under real write rates.
  • Relevance contract measures configuration changes and defines expected query results.

Flock Accused of Reactivating Its Cameras Without Notifying a Town

A Massachusetts newspaper reports that officials from the town of Littleton (population: 10,141) "claim that five of Littletonâ(TM)s Flock Safety cameras were switched back on by the company without…

  • Five Flock Safety cameras reactivated in Littleton, Massachusetts.
  • Town officials canceled contract with Flock Safety after discovery.
  • Flock Safety considered installing dashcams in Uber, Lyft drivers' vehicles.