#Meta
65 related articles
Meta's $25 Billion Verdict Lands, Reshaping Legal Risk for Tech Giants
A New Mexico judge orders Meta to pay $25 billion for misleading users, far exceeding the earlier $17 billion settlement. We analyze the ruling's impact on Meta and the tech industry's legal risk and platform liability.
Meta Open-Sources Muse Spark: America's Most Powerful Open-Weight Model Launches on Together AI
Meta releases Muse Spark Open, billed as America's most powerful open-weight model, now live on Together AI, with a joint technical livestream on October 26.
Meta Delays New AI Scraping Terms After Publisher Pushback
Meta delays enforcement of new commercial terms after publishers fear they allow scraping of news content for AI training. Analysis of the background, publisher concerns, and impact on the AI content ecosystem.
Meta Muse Wonder Lands on Cursor: 46.1% on CursorBench
Meta Superintelligence Labs' Muse Wonder launches on Cursor, scoring 46.1% on CursorBench at default effort — the highest in the Muse series. Input $4/M tokens, output $20/M tokens.
Zuckerberg Bets on AI Agent Networks: My Muse Talks to Your Muse
Meta CEO Zuckerberg bets on an AI agent network future: agents will shop, work, and socialize for people—even talk to others' agents. This article analyzes the logic, concerns, and competitive landscape of this vision, asking whether humans are ready to communicate mainly through robotic agents.
Meta Plans to Issue New Non-Voting Shares: Opening a New Financing Channel for the AI Race
Meta announces issuance of non-voting shares to fund AI infrastructure, potential acquisitions, and employee compensation, with trading expected to begin in 2027. This article analyzes the control considerations, AI investment pressure, and industry implications behind this financing arrangement.
Meta Muse Surpasses 10 Million Users, Shopping Emerges as Core Use Case
Meta's AI app Muse surpasses 10 million users, with shopping and productivity as top use cases. We analyze the business signals and why shopping matters for Meta's ad business.
Meta's 816B Model Activates Only 5% of Parameters: An Animated Deep Dive into MoE Architecture
Meta's latest 816B-parameter model activates only about 5% of parameters per token. This article uses a Reddit animation to explain MoE architecture's sparse activation, routing, and load balancing, and analyzes its inference cost advantages and deployment challenges.
Meta's Open-Weight Model Muse Spark Open Lands on Ollama: 1M Token Context
Meta's open-weight model Muse Spark Open launches on Ollama, focused on coding and tool calling with a 1M token context window. Here's its positioning, run commands, availability, and practical observations.
Meta Muse Wonder In-Depth Review: King of Long-Horizon Coding, But Less Versatile Than Opus 5.5
Meta Muse Wonder in-depth review: Full vs. standard version differences, head-to-head with Claude Opus 5.5, GPT-6 Astra, and Gemini 4 Argon on performance, pricing, and real per-task cost — leading in long-horizon coding but weak on factual knowledge.
Meta Releases Muse Wonder: Six Reasoning Levels, Performance on Par with GPT-6.1 Sol
Meta releases Muse Wonder and Muse Wonder Full with six reasoning tiers from minimal to max. Artificial Analysis rates its Max tier on par with GPT-6.1 Sol, but pricing, context window, and launch date remain undisclosed.
Meta Open-Sources Muse Spark: Open Weights but Not Open Source, with Three Commercial Strings Attached
Meta releases Muse Spark Open, a 816B-parameter open-weight model with 42B activated. But its license isn't OSI open source: attribution required, distillation restricted, and API resale over $20M needs a separate deal.
Meta Releases Muse Spark Open Weights: 816B Parameters, 1M Context, and Deployment Guide
Meta releases Muse Spark open weights with 816B total parameters, 42B activated, and 1M context. Covers architecture, hardware thresholds, benchmarks, license limits, and vLLM deployment.
Meta Open-Sources an 816B Beast: The Opportunities and Barriers of Muse Spark Open
Meta releases Muse Spark Open, an open-weights model with 816B total parameters, 42B activated, a 1M-token context, and multimodality. This article breaks down its MoE architecture, download channels, hardware barriers, and Muse Model License restrictions.
Meta Open-Sources Muse Spark: 816B Coding Model Lands on NVIDIA GB300
Meta open-sources Muse Spark Open: an 816B MoE coding model with 42B activated parameters, adapted for NVIDIA GB300 NVL72 and built for enterprise Agentic Coding.
Meta Muse Wonder Nears the Top of the Benchmarks — So Why Does the Market Only Give It 1%?
Meta launches Muse Wonder and Wonder Full, scoring 52/56 on Artificial Analysis — near Claude Opus 5.5 — yet Polymarket gives just 1% odds. Inside the benchmarks, rate limits, Muse privacy controversies, and a split practitioner verdict.
Meta Launches Muse Wonder Series: Active Memory and Active Focus Redefine Agents
Meta launches Muse Wonder and Muse Wonder Full frontier models, redefining persistent autonomous agents with Active Memory and Active Focus. DeepSWE 1.1 hits 79.0%, SWE-Marathon 57.5%, with API pricing and Muse Code subscription plans revealed.
Has Meta Caught Up with Anthropic? A Hands-On Analysis of Muse Wonder
Meta's Muse Wonder Full nears Anthropic and OpenAI on coding benchmarks, but the base version disappoints. Pricing and naming strategy draw criticism.
Meta Muse Wonder: MSL Joins the Frontier Lab Ranks
Meta MSL launches Muse Wonder and Wonder Full, positioned as Claude 5.5-tier frontier models, marking its formal entry into the top AI lab ranks.
Meta Releases Its First Frontier Model, Muse Wonder
Meta's Superintelligence Lab released its first frontier models, Muse Wonder and Muse Wonder Full. The dual-version strategy signals Meta's entry into the top tier of AI competition, potentially shifting the landscape from three players to four.