Company

Meta

Meta is a technology company actively developing artificial intelligence models, hardware products, and commercial platforms. The company recently removed disputed web crawling language from its commercial terms following concerns regarding AI training and website blocking preferences, maintaining that its web crawling practices remain unchanged. Meta has also addressed hardware and security issues, including fixing a Ray-Ban smart glasses bug that recorded conversations without wake phrases and resolving virtual machine escape vulnerabilities in its Muse AI agent prior to launch. Additionally, Meta has detailed a Private Processing architecture designed to secure cloud workloads for AI glasses. In model development, Meta offers open models such as Muse Spark Open, an 800-billion-parameter model requiring 1TB of RAM for local execution, alongside its Muse Wonder model series. Meta is expanding its AI ecosystem through tools like Muse for Small Business, open-source code for building Muse hardware gadgets, and the Personal Agent Protocol created with industry partners for AI agent commerce. The company also collaborates on scientific initiatives, joining a 300 million dollar investment in the Biohub virtual cell project. Across its broader operations, Meta implements platform safety measures, including AI tools to detect illegal child exploitation ads and updated WhatsApp parental controls. Internally, Meta reduced its usage of third-party AI tools like Anthropic Claude while utilizing research tax credits to offset data center expenses.

53 stories · Updated

Key facts

  • Meta released Muse Spark Open, an 800-billion-parameter AI model that requires 1TB of RAM to run locally.
  • Meta fixed a bug that caused Ray-Ban smart glasses to record conversations without the wearer saying the activation phrase.
  • Meta, Google DeepMind, and Isomorphic Labs invested $300 million into Biohub to help build a virtual cell.
  • Meta took action on 33.2 million pieces of child exploitation content in early 2026.
  • Meta saved $3.9 billion in 2025 by using a federal research tax credit for AI data centers.

Timeline

November 2026

October 2026

September 2026

August 2026

May 2026

April 2026

March 2026

  1. Meta's AI Agent Went Rogue and Leaked Sensitive Data to Employees

    An AI agent on Meta's internal forum gave an engineer advice that exposed a large amount of sensitive user and company data to employees for two hours. The incident triggered a major security alert and highlights the risks of deploying agentic AI at scale.