News/Meta
Meta announces cautious Q1 revenue forecast citing pricey bets on AI
Facebook-parent Meta reported strong Q4 2024 revenue of $48.4 billion while forecasting significant AI investments and organizational changes for 2025. Financial performance: Meta exceeded analyst expectations for Q4 2024 with revenue of $48.4 billion against estimates of $47.0 billion, though it provided cautious Q1 2025 guidance. Q1 2025 revenue is projected between $39.5 billion and $41.8 billion, potentially falling short of Wall Street's $41.72 billion forecast The company's Reality Labs division, focused on metaverse development, posted a $5 billion loss in Q4 despite beating sales targets Daily active users across Meta's family of apps grew 5% year-over-year to 3.35 billion...
read Jan 30, 2025Meta smashes Q4 expectations, touts ambitious AI infrastructure plans
Meta reported record-breaking fourth quarter results for 2024, with revenue reaching $48.39 billion and net income of $20.84 billion, significantly exceeding Wall Street expectations. Financial Performance: Meta's Q4 results demonstrate substantial growth with revenue up 21% and net income increasing by 49% year-over-year. The company achieved earnings of $8.02 per share, surpassing analyst expectations of $6.76 Daily users across Meta's apps reached 3.35 billion in December 2024, showing continued user base expansion For Q1 2025, Meta projects revenue between $39.5 billion and $41.8 billion, representing 8-15% year-over-year growth AI Investment Strategy: Mark Zuckerberg announced ambitious plans for AI infrastructure expansion...
read Jan 28, 2025Meta AI wants to be your all-knowing, all-remembering digital companion
Meta AI's conversational capabilities have expanded with new features that let the AI remember personal details shared in chats and integrate user data from across Meta's platforms. Key developments: Meta is updating its AI assistant across Messenger, WhatsApp, Facebook, and Instagram in the U.S. and Canada to provide more personalized interactions. The new Memory Boost feature enables Meta AI to retain and recall information shared during one-on-one conversations Users can delete stored memories at any time The feature does not apply to group chats The system operates across Meta's platforms, including Messenger, WhatsApp, Facebook, and Instagram How it works: Meta...
read Jan 27, 2025Meta AI is rolling out enhanced ‘memory’ features for its chatbot on Facebook, IG and WhatsApp
Meta's AI chatbot is expanding its personalization capabilities by remembering user preferences and incorporating data from Facebook and Instagram profiles. Core functionality: Meta is rolling out enhanced memory features for its AI chatbot across Facebook, Messenger, and WhatsApp platforms in the US and Canada. The AI can now remember specific details about users, such as dietary preferences and interests, through both direct input and contextual learning The system will adjust future responses based on user interactions, like automatically accounting for dietary restrictions once mentioned Memory features are limited to one-on-one conversations and can be deleted by users at any time...
read Jan 25, 2025Zuckerberg unveils Meta’s massive AI data center plans
Facebook founder and CEO Mark Zuckerberg announced plans to invest up to $65 billion in artificial intelligence initiatives in 2025, including a massive data center in Louisiana. Major investment details: Meta's AI infrastructure expansion represents one of the largest technology investments in recent history, centered around a Louisiana data center of unprecedented scale. The Louisiana facility's footprint would cover "a significant part of Manhattan," according to Zuckerberg's statement on Threads Construction of the $10 billion data center began in December 2024 and will continue through 2030 Meta plans to expand its GPU (specialized AI processing chips) capacity to over 1.3...
read Jan 24, 2025Meta aims to double its GPU count to 1.3M for AI development
Meta CEO Mark Zuckerberg plans to significantly expand the company's AI computing capabilities by doubling its GPU count to 1.3 million units by the end of the year. Strategic objectives: Meta aims to develop cutting-edge AI assistants and launch its Llama 4 model to compete with industry leaders ChatGPT and Google Gemini. Zuckerberg expects Meta AI to serve over 1 billion users in 2025 The company plans to create an "AI engineer" capable of contributing to Meta's R&D efforts This expansion follows recent layoffs of 3,600 Meta employees Infrastructure investment: Meta is planning a massive data center expansion that will...
read Jan 23, 2025Meta’s top AI lawyer explains why he’s leaving the company
A prominent lawyer and Stanford professor has cut ties with Meta over concerns about the company's policy changes and leadership direction. The breaking point: Mark Lemley, a leading copyright attorney, announced his departure from Meta's defense team in the Kadrey v. Meta lawsuit, citing the company's controversial policy shifts and leadership decisions. Lemley explicitly stated he "fired" Meta as a client due to what he characterized as the company's "descent into toxic masculinity and Neo-Nazi madness" under Mark Zuckerberg's leadership A key factor in his decision was Meta's recent policy change allowing users to refer to gay and transgender individuals...
read Jan 20, 2025Authors demand Meta’s AI training data in copyright lawsuit
Meta faces allegations of using BitTorrent to download and distribute pirated books for AI training, leading to new developments in an ongoing copyright lawsuit filed by authors. Core allegations: Authors including Richard Kadrey, Sarah Silverman, and Christopher Golden have filed a class action lawsuit against Meta for using their works without permission in AI training. Meta previously acknowledged using unofficial sources containing pirated content for AI training The company maintains that such use falls under fair use protection Meta denies that these allegations warrant updates to the original complaint New legal developments: United States District Judge Vince Chhabria has allowed...
read Jan 20, 2025Zuckerberg was reportedly aware that Meta trained its AI model on pirated works
The core revelation: Meta CEO Mark Zuckerberg approved the use of Library Genesis (LibGen), a known pirated content repository, to train the company's Llama 3 AI model, according to newly unsealed court documents. Key details of the disclosure: Internal communications revealed through a class-action lawsuit show Meta executives discussing the company's deliberate use of unauthorized copyrighted material. Sony Theakanath, Meta's director of product management, confirmed in an email that Zuckerberg approved LibGen's use for AI training The company explicitly planned to keep its use of LibGen confidential Meta employees discussed methods to remove copyright indicators from the pirated content Internal...
read Jan 19, 2025How Meta’s Segment Anything model is advancing the future of digital fashion
In a recent blog post on Meta, digital artist Josephine Miller demonstrated how Meta's Segment Anything 2 model is enabling real-time virtual fashion transformations. The innovation: Using Meta's Segment Anything 2 (SAM 2) model and other AI tools, London-based XR creative designer Josephine Miller creates videos where clothing appears to change colors and patterns instantly. Miller showcased the technology in an Instagram post featuring a gold evening gown that transforms through various designs and colors The project aims to demonstrate how digital fashion can reduce reliance on fast fashion while promoting sustainability The process combines ComfyUI, an open source stable...
read Jan 15, 2025Meta’s new AI model can translate speech from 100+ languages
Meta has unveiled SeamlessM4T, a new AI model capable of translating speech across 101 languages, marking significant progress toward real-time language interpretation technology. Key innovation: Meta's SeamlessM4T model enables more direct speech-to-speech translation, improving upon traditional multi-step approaches that convert speech to text, translate the text, and then convert it back to speech. The model demonstrates 23% higher accuracy in text translation compared to leading existing systems While Google's AudioPaLM can handle 113 languages, it only translates into English, whereas SeamlessM4T can translate into 36 different languages The technology leverages parallel data mining to match audio with subtitles from web...
read Jan 15, 2025WhatsApp’s next feature may be letting you create your own AI chatbot
Meta's WhatsApp is developing a new feature that will let users create and customize AI-powered chatbots within the messaging app. What's new: WhatsApp is working on an AI Character feature that will enable users to create personalized chatbots tailored to their specific needs and preferences. The feature was discovered in WhatsApp beta for Android version 2.25.1.26 Users will be able to customize chatbots for various purposes, including productivity and entertainment WhatsApp plans to provide suggestions to help users navigate the chatbot creation process Technical details: The AI Character feature appears to be an extension of Meta's existing AI Studio capabilities,...
read Jan 15, 2025Inside Meta’s race to do anything necessary to beat OpenAI
OpenAI and Meta are engaged in a high-stakes competition to develop advanced AI systems, with internal Meta documents revealing controversial data acquisition strategies. The core revelation: Meta executives discussed using pirated content from Library Genesis (LibGen), a book piracy website, to train their AI models while attempting to conceal this usage. Internal communications show Meta's primary objective was to match GPT-4's capabilities by any means necessary Company leadership viewed LibGen's content as "essential" for achieving state-of-the-art performance Executives discussed removing copyright information and metadata from training data to avoid legal complications Strategic considerations: Meta's internal deliberations highlight the growing challenge...
read Jan 14, 2025Meta allegedly trained its AI models on pirated torrent content
Meta's use of pirated content from torrent sites to train its Llama large language model has been revealed through court documents, leading to copyright litigation. Key developments: Court documents in the "Kadrey et al. v. Meta Platforms" case have exposed internal communications suggesting Meta's use of unauthorized content for AI training. Novelists Richard Kadrey and Christopher Golden filed the lawsuit in 2023, alleging Meta used their copyrighted works without permission Judge Vince Chhabria ordered the release of unredacted documents that were previously hidden from public view Internal communications show Meta employees expressing concerns about downloading torrented content on corporate laptops...
read Jan 14, 2025Meta to eliminate 5% of staff in shift to AI and AR development
Meta has announced plans to cut 5% of its workforce, approximately 3,600 employees, as part of a strategic shift to prioritize artificial intelligence and augmented reality development. Strategic reshaping: Mark Zuckerberg is implementing performance-based cuts to create space for new talent focused on AI innovation and smart glasses development. Employees deemed as "low performers" will be notified by February 10 and receive generous severance packages The company will retain some underperforming employees if management sees potential for future improvement This represents Meta's largest workforce reduction since 2023's "year of efficiency" cuts Performance management overhaul: Zuckerberg aims to accelerate the process...
read Jan 13, 2025You can now build custom chatbots on WhatsApp with AI character creation
Meta plans to introduce AI character creation capabilities directly within WhatsApp, allowing users to build custom chatbots through the messaging platform. Key development: Meta is developing features that will enable WhatsApp users to create personalized AI characters directly within the Android app, based on findings in beta version 2.25.1.26. The feature will allow users to build custom AI chatbots by describing their desired personality traits, expertise, and unique qualities The implementation appears similar to the existing Meta AI Studio tool currently available on Instagram's website WhatsApp plans to provide trait suggestions to help users during the character creation process A refresh...
read Jan 13, 2025Meta to deploy AI systems in 2025 that can perform the work of midlevel software engineers
Meta CEO Mark Zuckerberg has announced plans to automate midlevel software engineering work through artificial intelligence, while also implementing significant changes to content moderation and company culture. Key developments: Meta aims to deploy AI systems in 2025 that can perform the work of midlevel software engineers, potentially transforming how the company develops its applications. Zuckerberg anticipates AI will eventually handle all coding across Meta's applications Current midlevel software engineers at Meta earn mid-six-figure compensation packages The initial implementation may be costly, but the company expects long-term efficiency gains Content moderation shift: Meta is moving away from third-party fact-checking in favor of...
read Jan 10, 2025Meta-CoT framework enhances AI reasoning with explicit thought processes
A research team from multiple institutions has introduced Meta Chain-of-Thought (Meta-CoT), a new framework designed to enhance the reasoning capabilities of Large Language Models (LLMs). Key innovation: Meta-CoT builds upon traditional Chain-of-Thought prompting by explicitly modeling the reasoning process that leads to specific thought chains, representing a significant advancement in how AI systems approach problem-solving. The framework focuses on teaching LLMs not just what to think, but how to think through complex problems Meta-CoT incorporates multiple components including process supervision, synthetic data generation, and search algorithms The approach aims to mimic more sophisticated human-like reasoning patterns in artificial intelligence systems...
read Jan 10, 2025Meta faces new lawsuit for using copyrighted books to train AI
Meta faces accusations of deliberately using pirated books to train its AI systems, with CEO Mark Zuckerberg allegedly approving the use of unauthorized content despite internal concerns. Key allegations: The lawsuit, filed by prominent authors including Ta-Nehisi Coates and Sarah Silverman, claims Meta knowingly used pirated books from the LibGen dataset to train its Llama language model. Internal Meta communications revealed that executives were aware LibGen contained pirated content The dataset was allegedly distributed through peer-to-peer torrents Documents produced during discovery suggest Zuckerberg approved the use of LibGen despite concerns from Meta's AI executive team Legal context: This case represents...
read Jan 10, 2025Meta removes AI Hitler personas from Facebook
Meta's Facebook platform was found hosting AI-powered chatbots impersonating Adolf Hitler and other prohibited figures, violating the company's own AI Studio policies, according to an NBC investigation. The policy landscape: Meta's AI Studio, launched last year, allows users to create chatbots on Facebook Messenger, Instagram, and WhatsApp under specific guidelines. The platform prohibits chatbots from impersonating living people, copyrighted characters, religious figures, terrorists, and mass murderers Meta claims to review all chatbots before they go live to prevent "inaccurate, offensive, or otherwise objectionable" content The company recently made controversial changes to its "hateful conduct" standards, allowing for broader speech on...
read Jan 9, 2025Virgo’s AI model analyzes endoscopy videos using MetaAI’s DINOv2
MetaAI's DINOv2 computer vision technology is enabling Virgo, a San Diego-based medical company, to analyze endoscopy videos using artificial intelligence, with the goal of improving patient outcomes and advancing precision medicine. The innovation: Virgo has developed VirgoCloud, a system that captures and processes endoscopy videos, while also creating AutoIBD, an AI model that identifies potential candidates for inflammatory bowel disease clinical trials. The company has amassed over 1.75 million procedure videos, reportedly the largest dataset of its kind VirgoCloud connects to existing endoscopic equipment to capture, compress, and encrypt procedure videos The system transmits data to a HIPAA-compliant web portal...
read Jan 9, 2025Meta secretly trained its AI models on a Russian ‘shadow library,’ court docs show
Meta's use of pirated books database LibGen to train its AI language models has been revealed through court-ordered document unredaction, marking a significant development in an ongoing copyright lawsuit filed by authors. The core revelation: Meta accessed and utilized Library Genesis (LibGen), a controversial pirated content database, for AI model training, despite internal concerns about the legality and optics of this approach. Internal company discussions about using LibGen data were escalated to CEO Mark Zuckerberg Meta employees expressed hesitation about accessing LibGen data from corporate laptops The company's AI team ultimately received approval to use the pirated materials Legal context...
read Jan 9, 2025How to use Meta AI to enhance your selfies
Meta's new AI photo editing tools allow users to creatively enhance and transform their selfies across Facebook, Instagram, Messenger, and WhatsApp platforms through simple chat-based interactions. Key Features and Functionality: Meta AI offers a range of photo manipulation capabilities that let users modify everything from basic appearance touchups to creating fantastical scenarios. Users can access the tools by logging into any Meta platform and navigating to the chat or story section The AI assistant can adjust cosmetic elements like glasses color and lipstick shade More advanced features include adding objects, changing backgrounds, and transforming the entire narrative of an image...
read Jan 8, 2025Meta develops AI memory layer architecture to boost LLM accuracy and recall
Meta has introduced "scalable memory layers," a new architectural approach that enhances large language models' factual knowledge while reducing computational demands. Key innovation: Meta AI researchers have developed scalable memory layers that allow language models to store more factual information using sparse activation patterns, making them more efficient than traditional dense layers. The new architecture adds parameters to increase learning capacity without requiring additional compute resources Memory layers use key-value lookup mechanisms to encode and retrieve knowledge Unlike dense layers where all parameters are active simultaneously, memory layers only activate a small portion of parameters at a time Technical implementation:...
read