Meta's Homegrown RDMA Protocol Survives 10% Packet Loss
Meta detailed MetaRoCE, an in-house RDMA transport that skips PFC and still works at 10% packet loss; full spec ships at OCP in October.
Meta News and Analysis
29 verified stories covering Meta, product updates and industry developments.
Meta opens a 30B local agent model
Meta's Muse Glimmer is a 30B open-weight agent model built for local workflows, with Apache 2.0 weights, 24GB and 32GB deployment targets, image input, tool use, and DFlash speculative decoding.
Meta Ads Let AI CSAM Through
WIRED reported that researchers found more than 50 Meta ads containing AI-generated child abuse imagery, putting ad review and cross-border enforcement under scrutiny.
Meta faces a 26-plaintiff lawsuit over algorithm-assisted layoffs
Meta faces a 26-plaintiff lawsuit over algorithm-assisted layoffs. The piece reviews the verified facts and why the signal matters beyond one announcement.
Meta pulls Instagram public-photo image generation
Meta pulls Instagram public-photo image generation. A concise localization of the verified facts and the industry signal behind the story.
Meta prices Muse Spark 1.1 for agent work
Meta prices Muse Spark 1.1 for agent work is reframed for global technology readers, with the key numbers, product claims and open questions kept intact.
Meta says Watermelon is catching up with GPT-5.5
Internal AI benchmarks, compute scale and Meta’s model strategy shift.
Meta Shares Jump on AI Compute Rental Plan
Meta is exploring a business that rents out excess AI compute, turning a costly infrastructure buildout into a possible revenue line.
Meta Restricts Claude Code and Codex for AI Teams
Meta told some applied-AI engineers to seek approval before using Claude Code or Codex, citing model distillation and contract risks.
Meta contractors tested rival chatbots as minors
WIRED reported that Meta contractors posed as underage users to test ChatGPT, Gemini and Character.AI with crisis prompts about self-harm, eating disorders and drugs.
Google’s compute crunch reportedly cuts Meta’s Gemini access
Google’s compute crunch reportedly cuts Meta’s Gemini access
Meta meters internal AI use as token costs reach billions
A new AI Gateway would track spending, flag spikes and assign token budgets as Meta pushes employees toward internal tools such as MetaCode.
Meta and Reliance plan a 168MW AI data center
Meta is partnering with Reliance on a built-to-suit AI data center in Jamnagar, India, starting with 168MW and green power contracts.
Meta code points to NameTag face recognition for smart glasses
Wired found code for an unreleased feature that could let Meta smart glasses remember faces and surface names or context, raising the stakes for AI privacy.
Meta opens WhatsApp business agents globally
Meta Business Agent can answer questions, recommend products, book services, take payments and hand off to humans across WhatsApp, Messenger and Instagram.
Attackers used Meta AI support to hijack Instagram accounts
Hackers reportedly tricked Meta’s AI support flow into adding new emails and resetting passwords, with MFA proving the main barrier.
Meta develops AI pendant that records everything you say
Meta is reportedly building an AI-powered pendant that records conversations and provides an AI assistant, with internal testing planned within a year.
Meta launches AI assistant subscription at $19.99 per month
Meta introduces Meta One AI subscription tiers, including a $19.99 Premium plan with deeper reasoning, as it tests new revenue streams beyond advertising.
Broadcom, Meta, and Partners Invest $125M in UCLA Semiconductor Hub
Five companies including Broadcom and Meta jointly invest $125 million over five years to establish a semiconductor R&D center at UCLA, focusing on next-gen AI chips and workforce training.
Meta Lays Off 8,000 After Leaked Audio Reveals AI Training on Employee Data
A leaked recording of a Meta all-hands meeting shows Mark Zuckerberg admitted the company tracked employee computer usage to train AI, just as 8,000 workers were laid off.
Meta Forms New AI Units, Plans to Lay Off 8,000 on Wednesday
Meta is restructuring by moving 7,000 employees into four new AI-focused units and cutting about 8,000 jobs, affecting nearly 20% of its workforce.
Meta Backs Orbital Solar Power for AI Data Centers
Meta signed for up to 1 GW of space-based solar capacity and paired it with a long-duration storage deal as AI power demand strains the grid.
Big Tech's $645 Billion AI Spending Faces the Earnings Test
Microsoft, Alphabet, Meta, Amazon, and Apple report Q1 earnings this week. Investors are demanding clear evidence that massive AI capital expenditure is translating into revenue, margins, and monetization.
China Blocks Meta's $2 Billion Acquisition of AI Startup Manus
China's top economic regulator has formally prohibited foreign investment in Manus and ordered the unwinding of Meta's acquisition, while the startup's co-founders remain barred from leaving the country.
Meta cuts 8,000 jobs; Microsoft offers first-ever buyouts in 51 years
Meta is laying off 10% of its workforce, about 8,000 people, while Microsoft is offering voluntary retirement to roughly 8,750 US employees for the first time in its history, together potentially affecting nearly 20,000 roles as both companies redirect spending toward AI infrastructure.
Meta uses employee keystrokes to train AI that controls computers
Meta is recording every keystroke and mouse click from its employees to train an AI agent capable of autonomously controlling computers, raising new questions about data privacy and consent in the workplace.
Meta Launches Its First Closed-Source Model, Muse Spark
Meta released Muse Spark, its first fully proprietary flagship AI model, marking a strategic shift from open-source Llama toward competing directly with OpenAI and Google.
Meta's Next Flagship Models Avocado and Mango Will Be Partially Closed-Source
Meta is developing two new flagship AI models, Avocado and Mango, but unlike the fully open Llama series, these will only be partially open-sourced, with key capabilities like cybersecurity code generation withheld.
Llama 4 Debuts MoE Architecture with Scout and Maverick Models
Meta released Llama 4 on April 5, 2025, introducing the first MoE architecture in the Llama series with two models: Scout (17B active parameters, 10M token context) and Maverick (17B active, 128 experts, 1M token context), both natively multimodal via early fusion.