Click any tag below to further narrow down your results
Links
Yann LeCun, Meta's chief AI scientist, plans to leave the company to start his own venture focused on world models. His departure comes as Meta restructures its AI division amid increasing competition from rivals. LeCun has expressed skepticism about the current state of AI technology.
Meta has named Dina Powell McCormick as its new president and vice chair. She will focus on developing government partnerships to finance and deploy data centers globally, as the company ramps up efforts to build its AI infrastructure. The initiative aims to secure significant computing power for Meta's artificial intelligence models.
Meta has released SAM 3 and SAM 3D, new image segmentation models that enhance object recognition and enable 3D reconstruction of images. SAM 3 allows users to edit images through detailed text prompts, while SAM 3D can rebuild objects and people in 3D. Both models aim to improve creative applications and user interactions in various digital environments.
Mark Zuckerberg discussed Meta's focus on AI smart glasses during the Q4 2025 earnings call. He claims that sales have tripled recently and predicts a future where most glasses will be AI-enabled. Other tech giants like Google and Apple are also developing their own smart glasses.
Meta's Reels has turned from a TikTok imitation into a major revenue generator, projected to earn $50 billion in the next year. Instagram users now spend an average of 27 minutes daily on Reels, significantly increasing engagement compared to previous years. Meta aims to expand Reels to more platforms by 2026.
This article examines Meta's Llama AI models, focusing on their capabilities and limitations for developers. It tests Llama 3.2 on CRUD operations using the Svelte framework and discusses the setup process and comparisons with other AI models.
Meta's Chief Technology Officer, Andrew Bosworth, announced that the company's new AI team has developed its first significant models internally. Despite previous criticism of their Llama 4 model, the new models show promise and are expected to enhance consumer AI products in the coming years.
Nvidia publicly reacted to a report suggesting Meta might switch part of its AI infrastructure to Google's TPUs, causing a drop in Nvidia's stock. The move highlights a growing rivalry, as Google’s chips gain recognition as a viable alternative to Nvidia's GPUs.
Apple has lost at least four AI researchers, including a senior Siri executive, to companies like Meta and Google DeepMind. Notable departures include Yinfei Yang, who is starting a new venture, and Haoxuan You and Bailin Wang, who joined Meta's AI teams.
Meta is constructing a $27 billion data center in Louisiana through a joint venture with Blue Owl Capital. By keeping the project off its balance sheet and using a leasing strategy, Meta avoids increasing its debt and maintains its credit rating.
Meta's Bug Bounty Program marked its 15th anniversary, awarding over $4 million in bounties this year alone, totaling more than $25 million since its start. The program is expanding with a new pilot for experienced researchers and highlighting significant findings, including vulnerabilities in WhatsApp and Oculus.
Meta has released new AI models internally this month, which CTO Andrew Bosworth claims are promising. While details remain sparse, reports suggest that the company is developing a large language model and AI models for images and videos, referred to as Avocado and Mango.
Yann LeCun, a key figure in AI development, is increasingly at odds with Meta's direction. He is considering leaving the company to start a venture focused on world models, which he believes could advance AI more effectively than Meta's emphasis on language models.
This article outlines how Meta uses Randomized Control Trials (RCTs) and causal inference methods to evaluate new products. It discusses scenarios for applying these methods, the importance of clear communication among teams, and steps for implementing a framework to guide analysis decisions.
Meta acquired Manus for $2.5 billion, with Manus reporting $100 million in annual recurring revenue and processing 147 trillion tokens in eight months. The article compares gross profit per token among six AI companies, suggesting that investors value monetization efficiency over raw token volume.
Meta has acquired Limitless, an AI startup known for its conversation-recording pendant. The company will discontinue its hardware sales but will support existing customers for a year, transitioning them to a new unlimited plan. Limitless aims to assist Meta in developing AI-enabled wearables as competition in the market intensifies.
The article reviews the Meta Ray-Ban Display smart glasses, highlighting their comfort and design while noting the limitations of their monocular display. The glasses aim to reduce phone usage but face challenges in social acceptance and visual comfort due to their bulk and one-sided functionality.
Yann LeCun, Meta's chief AI scientist, is exploring the possibility of leaving the company to create a startup focused on world models. He has been in talks with associates and investors about this new direction, which diverges from Meta's current focus on large language models. His plans remain uncertain, and he may decide to stay with Meta.
Meta is projected to make 10% of its 2024 sales, around $16 billion, from ads promoting scams and banned products. Internal documents revealed that the company faces challenges in managing these deceptive ads while expressing concerns about the impact on its revenue if they were removed. A spokesperson claimed the figures were rough estimates and emphasized the company's efforts to combat fraud.
Nick Theriot shares a streamlined approach to scaling ad campaigns on Meta with a clean, seven-part system. His insights emphasize clarity in account structure, creative testing, and the importance of quality ads over quantity.
Meta's recent quarterly report shows strong user growth and increased ad revenue, driven largely by AI advancements. Despite a significant tax charge, their expenses are rising, particularly in R&D for AI and Reality Labs, where losses continue but show potential for improvement.
China is reviewing Meta's acquisition of AI startup Manus, valued at $2.5 billion, as a warning to local entrepreneurs tempted by U.S. buyouts. The Chinese government aims to control its AI technologies amidst a growing tech rivalry with the U.S.
Yann LeCun, Meta's chief AI scientist, plans to leave the company to launch his own startup focused on world models. His departure comes as Meta restructures its AI unit in response to competition from rivals like OpenAI and Google. LeCun has expressed skepticism about the current hype surrounding AI technologies.
Meta's AI strategy is in turmoil as CEO Mark Zuckerberg shifts focus from open-source Llama models to developing a proprietary model named Avocado. The company's recent hires and restructuring have created pressure to compete effectively against rivals like OpenAI and Google, despite ongoing challenges in its core advertising business.
Meta is planning to test subscription models for Instagram, Facebook, and WhatsApp, allowing paid users access to enhanced features and AI capabilities. The subscriptions aim to monetize Meta's investments in AI, including its recent acquisition of Manus. Users will get some features for free, while others, like the full version of the Vibes video tool, will require payment.
SAM Audio is a new tool from Meta that lets users isolate sounds from audio and video sources using text, visual, or temporal prompts. It can separate general sounds, music, and speech with high accuracy, allowing for clearer audio experiences. This model also includes an open-source evaluation dataset for audio separation.
Starting in 2026, Meta will evaluate employee performance based on their use of AI to enhance productivity. The company is promoting an AI-native culture by rewarding workers who drive significant results with AI tools and introducing an AI Performance Assistant for performance reviews.
A researcher revealed that some private Instagram profiles were exposing links to private photos in their HTML code, accessible to unauthenticated users. Although Meta fixed the issue shortly after being notified, they dismissed it as "not applicable" and did not acknowledge the severity of the vulnerability.
The FTC's five-year antitrust case against Meta, formerly Facebook, ended in defeat, highlighting the challenges of using antitrust laws against politically unpopular companies. The case emerged during heightened scrutiny of Big Tech, uniting both progressive and conservative critics.
A federal judge ruled that Meta's acquisitions of Instagram and WhatsApp did not violate antitrust laws, potentially allowing big tech companies to resume buying start-ups. This decision could shift the recent trend of avoiding such deals due to regulatory fears. Industry experts believe this will lead to a more active acquisition market.
Meta AI is set to launch a new model called Avocado, alongside updates that include app integrations with Gmail and Google Calendar. The company is also working on voice agents, scheduled tasks, and potential collaborations with other AI models like Gemini and ChatGPT. While the new features show promise, the performance of the Avocado model remains questionable.
This article explores how to optimize brand-building campaigns on Meta using data-driven insights. It highlights the importance of campaign objectives, creative types, and frequency in influencing brand metrics like awareness and purchase intent. The findings underscore the value of both high-quality and low-fi creatives within targeted audiences.
This article discusses the security risks associated with AI agents, particularly prompt injection vulnerabilities. It introduces the "Agents Rule of Two," a framework designed to minimize risks by limiting the properties an agent can have in a session to avoid harmful outcomes.
This article discusses how companies like Google and Meta are embedding AI into their services without giving users control over its use. It highlights concerns about privacy, personalized advertising, and the potential impacts on users as they navigate a heavily AI-influenced digital landscape.
Meta is moving away from its open-source AI strategy to develop a closed, paid model named Avocado, set to launch in spring 2026. This change reflects a significant pivot in its approach, aligning more closely with competitors like Google and OpenAI. The new Chief AI Officer, Alexandr Wang, supports this transition.
Dina Powell McCormick is appointed President and Vice Chairman of Meta, after previously serving on its Board of Directors. She will guide the company’s strategy and manage significant investments in infrastructure and global connectivity. Her background includes extensive experience in finance and public service under two U.S. presidents.
Meta is launching a dedicated Instagram Reels app for smart TVs, starting with Amazon Fire TV. This move allows users to watch short-form videos on larger screens, aiming to capture a share of the growing social video audience competing with platforms like YouTube and Netflix.
Mark Zuckerberg and Andy Jassy are vying for dominance in digital advertising. Zuckerberg promotes Meta's AI capabilities to attract ad budgets, while Jassy showcases Amazon's tools and data advantages, aiming to disrupt traditional advertising channels. Both leaders are eyeing a bigger share of the market, but Amazon's unique position could give it an edge.
Alan Dye, Apple’s VP of Human Interface Design, is leaving to join Meta, where he will lead a new design studio focused on AI and consumer devices. Stephen Lemay, a longtime Apple designer, will replace him. This shift comes amid several high-profile departures at Apple, including COO Jeff Williams and CFO Luca Maestri.
Soumith Chintala shares his emotional farewell after eleven years at Meta, where he led the development of PyTorch. He expresses pride in the project's success and confidence in its future leadership. Soumith discusses his desire to explore new opportunities outside of Meta.
Yann LeCun is leaving his position at Meta to start a new AI startup focused on advanced machine intelligence. He aims to develop systems that understand the physical world and utilize self-supervised learning to achieve this, distancing himself from the current emphasis on large language models.
Meta's new Lattice system integrates ad delivery across platforms, enhancing performance through better signal processing. Advertisers must focus on providing stronger first-party data and creative that drives engagement, as the system learns more quickly than most can adapt.
Meta plans to test premium subscription models for Instagram, Facebook, and WhatsApp, offering features that enhance user control and creativity. Instagram will focus on tools for creators, while the specifics for Facebook and WhatsApp are still being defined. The company aims to integrate AI into these offerings and will seek user feedback throughout the testing phase.
Meta shows users about 15 billion fraudulent ads daily, despite acknowledging the issue. Internal documents reveal the company is hesitant to fully combat fraud due to potential revenue loss, while regulators are increasing pressure for better protections.
Meta has purchased Manus, a Singapore-based AI startup, for $2 billion. Manus's technology allows for more autonomous decision-making compared to traditional chatbots, which could enhance Meta's platforms and user engagement. The acquisition faces potential regulatory scrutiny due to Manus's Chinese connections.
Meta has announced partnerships to become a major customer for nuclear power in the U.S., collaborating with TerraPower and Oklo for new reactors and expanding existing plants with Vistra. The company aims to secure substantial electricity for its AI data centers, with new reactors expected by 2030 and 2032.
Researchers found a significant security flaw in WhatsApp that allowed hackers to easily access phone numbers and some profile information for all 3.5 billion users. Despite being warned about the issue since 2017, Meta only implemented a fix in October 2023. Users are urged to review their privacy settings to protect their information.
Meta has rolled out new AI features for Facebook, allowing users to animate their profile pictures with preset motions like confetti and hearts. Users can also enhance Stories and Memories with AI-generated image edits and select animated backdrops for text posts. Custom animations aren't available yet, but Meta plans to add seasonal backgrounds.
Mark Zuckerberg announced that Meta will unveil new AI models and products in the coming months, focusing on AI-driven commerce. He emphasized the unique value of Meta’s access to personal data for creating personalized shopping tools. The company plans significant infrastructure investments to support these efforts.
Meta and EssilorLuxottica are in talks to double the production of AI-powered Ray-Ban smart glasses amid rising demand. They aim to increase annual capacity to over 20 million units by the end of 2026 to stay competitive in the market.
Meta plans to lay off about 10% of its Reality Labs employees due to ongoing financial losses, totaling around $70 billion since 2020. The division focuses on metaverse products, which have failed to gain consumer traction, prompting a shift in the company’s focus toward AI development.
StyleX is Meta's new styling library that combines the benefits of CSS-in-JS with the performance of static CSS. It compiles styles at build time into atomic CSS, reducing file size and resolving common CSS issues like collisions and specificity wars. The system is now the standard for styling across Meta's products and other companies.
Meta is starting a new initiative called Meta Compute, aiming to build tens of gigawatts of energy infrastructure this decade, with plans to scale up to hundreds of gigawatts over time. Santosh Janardhan and Daniel Gross will lead this effort, which they see as a strategic advantage in their operations.
Apple has lost several key members from its AI and Siri teams, including four researchers who are moving to Meta and Google DeepMind. Notably, Stuart Bowers, a senior executive, is joining Google to work on AI models that Apple is also using for Siri's upgrade. This follows significant organizational changes within Apple's AI division.
This article explains how Meta is using backend aggregation (BAG) to connect thousands of GPUs across multiple data centers for its Prometheus AI cluster. BAG facilitates high-capacity networking, enabling the infrastructure to meet the demands of large-scale AI applications. It details the technical aspects of BAG's design and implementation, emphasizing performance and reliability.
Google is teaming up with Meta to create “TorchTPU,” a software initiative designed to help developers transition from Nvidia's hardware to Google's TPUs while using the PyTorch framework. This collaboration aims to reduce reliance on Nvidia's software tools. Google is also planning to sell TPUs worth billions to Meta.
Yann LeCun, Meta's outgoing AI chief, is seeking 500 million euros to launch a new startup valued at 3 billion euros. He has appointed Alexandre LeBrun as CEO and aims to develop AI systems that understand the physical world. Concerns about a potential AI bubble are growing amid high pre-launch valuations.
Alan Dye, Apple’s head of user interface design, is joining Meta, marking a significant shift as the company focuses on AI-driven consumer devices. Apple will replace Dye with designer Stephen Lemay. This move highlights Meta's ambitions in the tech space.
Senators Josh Hawley and Richard Blumenthal urged the FTC and SEC to investigate Meta for profiting from scam ads on Facebook and Instagram. Internal documents revealed that Meta projected to earn $16 billion from illicit advertising, while the senators criticized the company's inadequate fraud prevention measures.
Mark Zuckerberg intends to significantly reduce funding for Meta's metaverse efforts, possibly slashing the budget by 30%. This could lead to layoffs in the metaverse group as early as January. The cuts affect initiatives like Meta Horizon Worlds and the Quest VR unit.
Meta has integrated AI editing tools into Instagram Stories, allowing users to modify images and videos using text prompts. Features include adding or removing elements and applying preset effects. Users should be aware that using these tools means accepting terms that allow Meta to analyze their media.
Meta has increased the limit for its creative testing feature from five to ten ads within an ad set. This allows advertisers to experiment with more variations when creating or updating ads. You can access this feature in the Creative Testing section during ad setup.
Meta is not just removing scam ads; it's also making them harder for regulators and journalists to discover. Internal documents reveal the company is implementing tactics to obscure these ads to avoid costly regulatory measures, particularly in Japan. This strategy is part of a broader approach to limit scrutiny across multiple countries.
Meta reported a 25% increase in revenue for Q4 2025, driven by a surge in ad impressions and prices as it leverages AI in its advertising model. The company now has 3.58 billion daily users, reflecting its massive reach in the digital ad market.
WhatsApp has launched a new feature called Strict Account Settings, which allows users to enhance their privacy with a single toggle. This mode restricts various functionalities, making users less visible and enabling stronger security measures. Additionally, Meta has replaced an old media-handling library with a new one built in Rust for better security.
Meta has implemented a new scheduler called LAVD, originally designed for the Steam Deck, on its servers. This scheduler enhances load balancing and performance across various CPU and memory configurations. The article links to a presentation and slide deck for those interested in the technical details.
Meta is set to buy Manus, a Singapore-based AI startup founded by Chinese entrepreneurs, for over $2 billion. Manus is known for its AI agent that creates detailed research reports and custom websites, utilizing advanced AI models. This acquisition highlights Meta's strategic push into AI as it competes with firms like Google and Microsoft.
Meta is rolling out a software update for its AI glasses, introducing features like conversation focus to amplify voices in noisy settings and a Spotify integration that plays songs based on what you're looking at. Users can adjust settings easily to enhance their experience.
Palmer Luckey argues that Meta's recent closure of its VR game studios is beneficial for the industry's long-term health. He believes the layoffs are part of a necessary shift away from competing with smaller developers and emphasizes that Meta still leads in VR employment.
Internal documents reveal that Meta's Vibes AI video feed has about 2 million daily active users, with notable growth in India and Brazil. While usage is increasing in some regions, it has declined in parts of Southeast Asia, and user engagement patterns differ between new and returning users.
Meta has introduced Segment Anything Model 3 (SAM 3), which enhances object detection, segmentation, and tracking in images and videos using text and visual prompts. The release includes model checkpoints, a new playground for experimentation, and applications in platforms like Facebook Marketplace and Instagram's Edits app. SAM 3 also features a data engine that combines AI and human annotators to speed up image and video annotation.
Google is shifting its strategy by offering its custom TPUs for deployment in customer data centers, moving away from using them only in its own cloud. Meta is reportedly in talks to integrate these chips, planning a multibillion-dollar investment starting in 2027 while also renting TPU capacity from Google Cloud. This could significantly boost Google's presence in the AI chip market and challenge Nvidia's dominance.
Meta is creating a new AI model called Mango for image and video generation, along with a text-based model named Avocado aimed at improving coding capabilities. These models are expected to launch in early 2026, following a major restructuring of Meta's AI team that included hiring over 20 researchers from OpenAI.
Meta's DrP platform automates root cause analysis for digital incidents, significantly reducing resolution times. It replaces manual investigation methods with programmatic playbooks and integrations, enhancing efficiency and on-call productivity for over 300 teams.
The UK Gambling Commission's Tim Miller accused Meta of failing to proactively identify and remove ads for illegal gambling sites. He argued that Meta's claims of needing user reports to address these ads are false, highlighting the company's searchable ad library as evidence of its complicity in the issue. Miller urged for collaboration among regulators and industry players to combat illegal gambling effectively.
Meta will end the external Facebook Like and Share buttons on February 10, 2026. These plugins, introduced in 2010, have seen declining use as the web has evolved, and site admins can choose to remove them before the official discontinuation. Remaining plugins will become invisible rather than malfunctioning.
A class action lawsuit claims WhatsApp misrepresents its end-to-end encryption, alleging that user messages are accessible to Meta. The article discusses the implications of these allegations and the technical aspects of WhatsApp's encryption system.
Meta's new AI team, led by Alexandr Wang, is clashing with existing executives who focus on the social media side of the business. The AI researchers aim for groundbreaking developments, leading to budget cuts in other departments, notably Reality Labs. This has created a divide within the company.
Meta has unveiled Llama 4, a significant advancement in open-source AI technology, promising improved performance and accessibility for developers. This model aims to enhance the capabilities of AI applications across various industries and is expected to set new standards in the field.
CoreWeave's shares rose nearly 12% after announcing a $14.2 billion agreement to provide Meta with artificial intelligence cloud infrastructure, following a recent $6.5 billion expansion with OpenAI. The deal highlights the growing partnerships essential for AI advancements, as Meta invests significantly in expanding its AI capabilities and infrastructure by 2032.
Tesla's team lead for the Optimus AI project has left the company to join Meta. This departure raises questions about the future of Tesla's robotics initiatives as the company continues to innovate in artificial intelligence and electric vehicles.
Meta Platforms aims to enable brands to fully automate the creation and targeting of advertisements using artificial intelligence by the end of 2026. This initiative is expected to enhance the efficiency of ad campaigns, leveraging Meta's extensive user base of 3.43 billion active users. However, concerns regarding brand safety and creative control in AI-driven advertising persist among marketers.
Meta faces backlash for developing AI chatbots impersonating celebrities like Taylor Swift and Lewis Hamilton without permission, leading to the creation of explicit content. Despite company policies prohibiting such imagery, the bots made sexual advances and produced inappropriate images, prompting Meta to remove several of them.
Meta plans to monetize WhatsApp by introducing advertisements in the app's Updates section, despite assurances that personal messages will remain private. Privacy experts express skepticism, citing Meta's history with data handling and the potential risks associated with ad-targeting based on user data.
Meta has successfully disrupted covert operations orchestrated by state actors from Iran, China, and Romania that aimed to manipulate public opinion on social media. These efforts included misinformation campaigns and fake accounts, which Meta identified and removed to protect users from deceptive practices. The company's actions reflect its ongoing commitment to enhancing platform security and integrity.
Meta is shifting its focus to launch an ultralight Horizon OS headset named Puffin by the end of 2026, while postponing the Quest 4 series until 2027. Puffin is designed to be lightweight and controller-free, utilizing a tethered compute puck to enhance portability, but the company's evolving hardware strategy comes amid declining sales in its Reality Labs division.
Meta CEO Mark Zuckerberg is launching a new AI superintelligence lab, offering nine-figure compensation packages to attract top talent in the field. The initiative aims to develop an AI surpassing human capabilities, with reports indicating frustration over the company's current AI progress has led to a more hands-on approach from Zuckerberg.
Mark Zuckerberg proposed a $450 million settlement to the FTC to resolve a major antitrust case against Meta, significantly lower than the $30 billion the agency demanded. The negotiation highlights the ongoing tensions between the tech giant and federal regulators over competition concerns.
Researchers have found that Meta and Yandex are exploiting legitimate internet protocols to de-anonymize Android users by embedding tracking codes in websites. This practice allows them to bypass security measures and link web browsing identifiers to persistent user identities in their mobile apps. Google is currently investigating this issue.
Meta's evolution in data infrastructure focuses on the integration of artificial intelligence into its systems, emphasizing scalability and efficiency to handle increasing data demands. The article highlights the advancements in technology that support AI initiatives and improve operational capabilities.
Tim Cook is set to remain as Apple's CEO despite an impending management shake-up following the departure of the chief operating officer. The company is also reducing its plans for the iOS 26 Liquid Glass revamp and is facing competition as Meta hires away Apple's top AI models executive.
Meta is reportedly in discussions to acquire Play AI, a startup specializing in voice cloning technology. The acquisition aims to enhance Meta's capabilities in artificial intelligence and voice technologies, potentially integrating these advancements into their existing platforms.
OpenAI is significantly increasing its stock-based compensation to $4.4 billion to counteract Meta's aggressive talent acquisition strategies, which have seen the poaching of key researchers. This financial response highlights the intense competition for AI talent and the pressures faced by companies to retain top talent amidst escalating costs.
Meta is postponing the release of its flagship AI model, Behemoth, due to challenges in enhancing its capabilities. Initially scheduled for an April launch, the timeline has now shifted to fall or later, raising concerns among staff about the effectiveness of the model compared to prior versions.
Meta has announced a new API for its LLaMA AI models, aiming to provide developers with enhanced access to its advanced AI capabilities. This initiative is part of Meta's ongoing efforts to expand the use of its AI technologies across various applications and platforms.
Meta has launched a standalone app for its AI tool, allowing users to access and utilize its features independently. This move is part of Meta's strategy to enhance user engagement and provide a more streamlined experience for those interested in AI capabilities. The app is designed to cater to a growing demand for AI applications in daily tasks and creative processes.
The article discusses Meta's recent earnings report and its implications for the company's future, particularly in the context of declining user engagement on its social networking platforms. It highlights the challenges Meta faces as it attempts to pivot and innovate in a competitive landscape.
Apple's decision to abandon an open-source AI initiative due to concerns over performance transparency has led to a significant loss of talent to Meta. Key figures in Apple's AI division have left, citing a clash between Apple's secrecy-driven culture and the collaborative nature of AI research, raising questions about the company's ability to compete in the evolving AI landscape.
Meta is shifting its focus away from the metaverse, with CEO Mark Zuckerberg appointing Vishal Shah to lead a new initiative centered on artificial intelligence. This change signifies a strategic pivot for the company as it seeks to innovate and adapt in a rapidly evolving tech landscape.
Meta has released a significant update for Instagram that enhances its functionality with 3D elements and improved navigation for users of Quest VR headsets. This update aims to create a more immersive experience for users interacting with Instagram in virtual reality.