Everything tagged meta, newest first. Tags come from the classifier reading each item's summary; 394 tags used 25 times or more have their own page.
Meta’s new personal AI agent Muse represents its biggest consumer AI push yet, but by seeking access to emails, calendars, payments, health services, and other sensitive data, it becomes a major test of whether users still trust Meta to handle their information responsibly.
Meta was slow to remove ads that promoted tools for digitally undressing young girls’ Instagram photos, raising concerns about child safety, platform moderation, and the company’s responsiveness to harmful content.
This video satirically reviews a chaotic week of major AI model launches from Anthropic, Meta, and OpenAI, highlighting technical capabilities, pricing strategies, outages, embargo drama, and the hype-driven status games surrounding early access and AGI claims.
Meta’s new Muse Spark model for coding and other agents offers a steep average 95% discount to users who share prompts and model outputs, framing the data contribution as a way to help improve future models.
Meta introduces ZGateway, a proxy layer in front of ZippyDB that unifies traffic and adds admission control, load balancing, cross-region resilience, and improved operational capabilities for its most widely used key-value store serving billions of requests.
Mark Zuckerberg’s vision of broadly distributed personal AI, framed as “superintelligence,” is presented as a way to boost prosperity and prevent AI power from concentrating in a few government or corporate hands, while a roundtable questions whether wider access truly means wider control.
Meta describes an AI “secondary expert” designed as an organizational second brain, combining a structured, auditable knowledge architecture with expert learning so specialized knowledge can be preserved, shared, and reused across teams.
The company is experimenting with robots to handle tasks that technicians can perform, suggesting a move toward automation of routine technical work.
Meta updated its AI glasses so recording stops whenever the safety light is covered, addressing a privacy loophole that could let users secretly capture video by blocking the indicator meant to signal active recording.
Sandhya Devanathan has taken on a new role overseeing selected OpenAI operations across Southeast Asia and Australia, expanding the company’s regional leadership and operational focus in those markets.
Meta’s settlement with 29 states permits the company to keep some data from children under 13 for training and testing age-detection models, underscoring a privacy compromise embedded in the agreement.
Meta is finding it difficult to replace human workers with AI agents, as a report highlights the limits, challenges, and practical gaps in automating tasks that people currently handle effectively.
Meta has agreed with nearly every U.S. state to impose daily limits on children’s social media use, signaling a broad regulatory compromise aimed at addressing concerns about youth screen time and online safety.
Meta plans to introduce a daily two-hour time limit for teens, which can only be turned off with parental permission, signaling a stronger approach to youth screen-time management.
Andrew Brown introduces a course on MetaMuse Code Essentials, focusing on Meta’s vertically integrated AI offerings—Muse Spark and Muse Glimmer—highlighting their multimodal capabilities, benchmark strength, open-weight nature, and token-based pricing as practical options for building AI applications.
Meta was accused in a lawsuit of intentionally designing Instagram and Facebook to be addictive for children while allegedly knowing the platforms could harm young users.
MTIA 300 is Meta’s first in-house training and inference accelerator for ranking and recommendation models, featuring built-in NIC chiplets and communication-offloading engines that, together with a co-designed HCCL library, improve training communication performance beyond general-purpose GPUs.
Ars reviews Zuckoff, a new free app designed to detect Meta AI glasses, addressing privacy concerns associated with wearable technology.
Mark Zuckerberg has recently purchased a comfortable residence near Meta's international headquarters in Ireland.
Meta is launching Pocket, an AI-driven app for creating and sharing interactive games, to U.S. users following a test phase in Brazil.
Meta introduced Glimmer, an open-weight AI model for public use, promoting AI accessibility, contrasting with their proprietary Muse Spark model.
Meta has released Muse Glimmer, a 30 billion parameter open-source AI model, aiming to regain credibility by enabling local execution on consumer hardware, following criticism of its previous closed-source approaches and strategic acquisitions.
Meta's Muse Glimmer 30 billion parameter model excels in agentic tasks and multimodal capabilities, offering a competitive edge for local deployment, despite lagging in pure coding benchmarks compared to Quen 3.627B.
Judges have clarified that Section 230 offers a legal defense for Meta against lawsuits, but it does not grant complete immunity.
Meta is emerging as a significant player in the AI industry with its latest frontier-level AI models, including Meta Muse Spark 1.1, which outperform competitors like Opus 4.8 and GPT-5.5 in real-world task efficiency.
Meta's Muse Glimmer model reflects Mark Zuckerberg's vision for personal superintelligence and highlights the growing divide in AI accessibility and ownership.
Meta's release of Muse Spark 1.2 and Muse Code marks a significant AI advancement, offering competitive pricing and innovative features for complex software engineering tasks, while leveraging real-world UI references through Mobin's MCP for enhanced app development.
Meta's Generative Ads Recommendation Model (GEM) now trains at LLM scale, achieving doubled training efficiency and quadrupled FLOPs using thousands of advanced GPUs.
Meta is restoring access to WhatsApp accounts that were mistakenly flagged and placed under review, following user reports of unexpected lockouts.
Researchers claim that TikTok, X, and Meta are failing to comply with legal obligations to provide necessary data, raising concerns about transparency and accountability.
Meta is researching Hierarchical Interest Representation to create a unified embedding layer that connects users' inferred interests with advertisers' offerings in deep funnel ads.
Meta has refuted allegations that it employs artificial intelligence to dismiss employees who have disabilities or medical issues.
Muse Spark 1.1, Meta's latest multimodal reasoning model, excels in Gench workflows, coding, and multimodal reasoning, outperforming competitors like Opus 4.8 at a lower cost, while advancing Meta's vision of personal super intelligence.
Meta addressed potential latency issues in its ad serving fleet caused by a Linux kernel upgrade by implementing a customized scheduling policy using the BPF-based sched_ext framework to optimize ad delivery performance.
Meta's AI image generator utilizes public Instagram photos for training, but users can opt out to prevent this.
Meta has developed a non-invasive AI system that translates brain activity into text, offering potential communication breakthroughs while raising privacy concerns about brain data collection.
Recent AI advancements include a full-body waifu, a Chinese model trained without Nvidia GPUs, global availability of Claude Fable 5, efficient Google models, real-time video editing, thought-to-text conversion, and a new sheet music reading AI called Musev VIT.
Meta celebrates a decade of sponsoring the Python Software Foundation, highlighting Python's critical role in its engineering stack and commitment to open-source development.
Patronus AI, a startup founded by ex-Meta AI researchers, is witnessing overwhelming demand for its agent-testing services, according to its investor.
Meta has engineered ultra-narrow batteries to fit within the temple arms of smart glasses, enabling them to power advanced features like cameras, speakers, and AI workloads.
Attackers exploited Meta's AI by politely requesting access, resulting in the unauthorized takeover of 20,000 Instagram accounts, highlighting a potential widespread vulnerability in AI systems.
Meta is beginning to dismantle its $2 billion acquisition of Manus following a directive from Beijing to reverse the deal.
A recent report indicates that a unit with 6,500 employees is nearing a state of rebellion.
Meta has not provided any information regarding the reasons for its decision or whether it plans to reverse it.
Meta is introducing global paid subscription plans for Instagram, Facebook, and WhatsApp, alongside testing new AI, creator, and business services under the "Meta One" brand.
Apple, Meta, and Google provide specialized security modes to enhance device protection against targeted spyware attacks, detailing their functions and activation processes.
Critics have pointed out that a lawsuit filed by a US Senate candidate is lacking in factual support, raising concerns about its validity.
The Friend Bubbles feature on Facebook highlights Reels watched by friends, requiring complex engineering to scale for billions of users.
Meta has revamped its data ingestion system to improve reliability at scale, involving a large-scale migration to a new architecture.
Meta stated that Kenyan workers failed to meet the company's standards, leading to their dismissal.
