AMI Labs and World Labs Keep Product Development Under Wraps
AMI Labs and World Labs Keep Product Development Under Wraps AMI Labs, an AI research company focused on world models, has gained significant buzz and funding since its inception less…
Meta
47 primary articles · 13 secondary mentions
Primary coverage
Meta's Muse AI Assistant Now Available on Mac for User Actions
Meta's Muse AI Assistant Now Available on Mac for User Actions Meta has officially launched its Muse AI assistant for Mac, allowing users to execute actions directly on their computers.…
Instinct and Meta's Muse Launch Phone Calling Features for AI Assistants
Instinct and Meta's Muse Launch Phone Calling Features Instinct has announced its new Concierge feature, which includes the ability to make phone calls, as revealed by founder Noah Shinn on…
Meta to Unveil Luna Smart Glasses with Six Microphones and No Camera
Meta to Unveil Luna Smart Glasses with Six Microphones and No Camera Meta is set to introduce its new smart glasses model, Luna, which will not include a camera. This…
Threads Introduces Tools for Podcast Promotion and Audience Engagement
Meta's Threads is rolling out new tools specifically designed for podcast creators, enabling them to promote their shows and engage with audiences directly on the platform. This follows last year's…
Meta Introduces AI Agents for WhatsApp Business Setup and Management
Meta has launched a WhatsApp Business MCP server that allows developers to utilize AI coding agents, including Claude, Cursor, Codex, and ChatGPT, to automate various aspects of business setup. These…
US DOJ Supports Fair Use for AI Training Amid Copyright Controversy
The US Department of Justice asserts that training AI models on copyrighted text qualifies as fair use. This position directly contradicts a report issued by the US Copyright Office, which…
Instagram's AI Profiles Often Indistinguishable from Real Users, Says Meta
Meta's Instagram has acknowledged that users often cannot distinguish between AI-generated profiles and real people. In response, the platform plans to implement a new labeling system by late 2024, replacing…
Meta Releases Open-Weight Glimmer Model Amid Muse Spark API Restrictions
Meta has released the open-weight AI model Glimmer this week, aiming to broaden AI accessibility. CEO Mark Zuckerberg stated, "AI should be for everyone," highlighting the company's commitment to democratizing…
Meta's Muse Glimmer Model to Compete with OpenAI, Anthropic in AI Market
Meta's new Muse Glimmer model features 30 billion parameters and requires less than 20 GB of memory for its compressed weights. This model is part of a broader strategy by…
Meta AI uses a second AI agent as a memory coach to keep long tasks on track
Meta AI has developed a novel approach to mitigate the issue of AI agents forgetting previously diagnosed errors during complex task execution. This is achieved by introducing a secondary memory…
Meta Faces Lawsuit Over AI-Driven Layoffs Affecting 8,000 Employees
Meta laid off 8,000 employees, with a lawsuit claiming the company used internal AI systems to generate the layoff lists. This lawsuit is filed in a California federal court, alleging…
Tencent Pursues Majority Stake in Manus After Meta's $2B Deal Blocked
Tencent is in talks to acquire a majority stake in Manus, an AI agent startup valued at $2 billion. This move comes after Beijing blocked Meta's $2 billion acquisition of…
Meta Launches Muse Spark 1.1 API at $4.25 per Million Tokens
Meta's Muse Spark 1.1 API is priced at $4.25 per million output tokens, directly entering the competitive AI API market. This launch follows the release of Grok 4.5 yesterday, intensifying…
Meta Tests Prototype of AI-Powered Headsets with 'Super Sensing' Feature
Meta is currently testing a prototype of AI-powered headsets featuring 'Super Sensing', which utilizes a camera and microphone to record every moment of the wearer's life. This development marks a…
Meta's Muse Image Raises Legal Concerns Over Instagram Photo Usage
Meta's new image generation model, Muse Image, features a controversial @-mention capability that allows users to generate images of individuals using their Instagram photos. This functionality has raised significant legal…
Meta's AI Agents Progressing Slower Than Planned, Says Mark Zuckerberg
Mark Zuckerberg stated that the AI agents Meta reorganized around are progressing slower than planned. This internal assessment highlights a delay in the development timeline for Meta's AI initiatives, which…
Meta Limits Use of Claude and Codex to Protect Training Data Integrity
Meta has implemented a restriction on its engineers, prohibiting the use of Anthropic's Claude and OpenAI's Codex. This decision aims to prevent outputs from these AI tools from being integrated…
NEA’s Tiffany Luck says enterprises are still figuring out their AI ROI
In a recent discussion, NEA’s Tiffany Luck emphasized that many enterprises are still grappling with the complexities of measuring their return on investment (ROI) in artificial intelligence. This conversation comes…
Uber's AI Budget Surpasses Expectations Amid Claude License Cuts
META: Uber's AI spending has surged, while some firms are reducing Claude software licenses. Uber's annual AI budget reportedly exceeded expectations within just a few months. This surge in spending…
A Technical Taxonomy of LLM Agent Communication Protocols
Problem The paper addresses the lack of a unified framework for classifying and analyzing communication protocols among large language model (LLM) agents in multi-agent systems. As LLMs evolve, the need…
Knowledge Reutilization in Meta-Reinforcement Learning
Problem The paper addresses the limitations of existing end-to-end meta-reinforcement learning (MRL) methods, which often couple task inference with embodiment-specific control. This coupling can obscure non-parametric task semantics, reduce sample…
Why Meta Paid $14.3B For Scale AI And Alexandr Wang’s Data Empire - Forbes
Meta has made headlines with its $14.3 billion acquisition of Scale AI, a company renowned for its data infrastructure that supports AI development. This strategic move is particularly significant as…
Learning Fair Pareto-Optimal Policies in Multi-Objective Reinforcement Learning
Problem The paper addresses the gap in multi-objective reinforcement learning (MORL) concerning fairness in policy selection, particularly in scenarios with dynamic or unknown user preferences. Existing single-policy MORL methods, which…
Meta’s new ‘AI Mode’ on Facebook pulls from public info across its platforms
Meta has announced the rollout of a new AI Mode on Facebook, designed to enhance user experience by pulling from public information across its various platforms. This initiative, revealed on…
Benchmarking LLM Agents on Meta-Analysis Articles from Nature Portfolio
Problem The paper addresses the lack of comprehensive benchmarks for evaluating systematic scientific reasoning in meta-analysis, particularly in the context of literature retrieval, study selection, and statistical aggregation. Existing benchmarks…
A year after Meta tapped Alexandr Wang to build a new AI model, Zuckerberg has to sell it - CNBC
Meta is facing significant pressure as CEO Mark Zuckerberg must consider selling an AI model developed by Alexandr Wang, a year after engaging him for the project. This situation arises…
Meta reportedly moves to unwind $2B Manus deal after Beijing’s demand
Meta is in the process of dismantling its $2 billion acquisition of Manus, a move prompted by a directive from Chinese authorities. This development is significant as it highlights the…
Meta’s months-old AI unit is a soul-crushing gulag, say the engineers stuck inside it
Meta's newly established AI unit, which employs around 6,500 engineers, is reportedly experiencing significant internal strife. Employees have voiced their frustrations, claiming that the work environment is oppressive and stifling,…
AgentBeats: Agentifying Agent Assessment for Openness, Standardization, and Reproducibility
Problem The rapid advancement of agent systems across various domains has led to fragmented evaluation methodologies, primarily reliant on fixed, LLM-centric benchmarks that necessitate extensive integration and often result in…
Claw-SWE-Bench: A Benchmark for Evaluating OpenClaw-style Agent Harnesses on Coding Tasks
Problem The paper addresses the lack of standardized evaluation metrics for general-purpose agents like OpenClaw in coding tasks, particularly under the constraints of the SWE-bench framework. Existing benchmarks do not…
Adjoint Method versus Physics-Informed Neural Networks in PDE-Constrained Inverse Problems
Problem The paper addresses the lack of a standardized comparison between adjoint-based optimization and physics-informed neural networks (PINNs) in the context of PDE-constrained inverse problems. Previous studies often employed disparate…
Market Design for AI: Beyond the Copyright Binary
Problem The paper addresses the inadequacies of current market structures for human-generated content used in AI training, highlighting a gap in the literature regarding effective market design that balances technological…
Soft-Prompt Tuning for Fair and Efficient LLM Benchmark Evaluation
Problem The paper addresses the inadequacies in current benchmark evaluations of large language models (LLMs), which often misrepresent a model's knowledge due to reliance on specific formatting requirements. This issue…
Meta signs first AI data center deal in India with Reliance
Meta has officially entered into a partnership with Reliance to establish its first AI data center in India. This collaboration is pivotal as it underscores Meta's commitment to enhancing its…
It Takes One to Bias Them All: Breaking Bad with One-Shot GRPO
Problem — This work addresses a significant gap in understanding the robustness of alignment mechanisms in large language models (LLMs) against biased inputs. Despite extensive post-training alignment efforts, the authors…
From 0-to-1 to 1-to-N: Reproducible Engineering Evidence for MetaAI Recursive Self-Design
Problem — This work addresses the nascent concept of MetaAI, which lacks a comprehensive framework for evaluating AI systems that can modify their own design and operation. The authors argue…
Instagram AI chatbot breach may have affected over to 20,000 accounts, Meta discloses
Meta has revealed that a security breach related to its Instagram AI chatbot potentially affected more than 20,000 user accounts. This incident, which lasted nearly seven weeks, involved the chatbot…
Meta's Hatch AI agent could cost up to $200 a month and marks its first paid AI product
Meta has unveiled its Hatch AI agent, marking a significant step as the company's first paid AI product. This development comes at a time when Meta is seeking new revenue…
Meta AI chief sees opportunity in models’ giving health advice - ET BrandEquity
Meta AI's leadership is eyeing a promising avenue for artificial intelligence: offering health advice through advanced AI models. This move comes at a time when the intersection of technology and…
The Download: AI hacking beyond Mythos, and chatbots’ impact on our brains
Reports have emerged that hackers have successfully leveraged Meta's AI customer support system to steal Instagram accounts, highlighting a significant security vulnerability within the platform. This incident is particularly alarming…
Meta steals a tactic from Tesla and builds data centers in tents
In a surprising move, Meta is reportedly exploring the use of tent structures for its data centers, a tactic reminiscent of Tesla's approach to temporary facilities. This shift comes as…
Reinforcement Learning Elicits Contextual Learning of Unseen Language Translation
Problem This work addresses the limitations of existing methods for translating unseen or low-resource languages, which often rely on continued training or encoding specific language grammars. These approaches tend to…
Meta rolls out a new AI creator assistant on Facebook
Meta has unveiled a new AI creator assistant on Facebook, aimed at empowering content creators with insights and tools to enhance their engagement strategies. This rollout comes at a crucial…
Podcast: Hackers Asked Meta AI To Let Them In. It Worked
In a recent podcast episode, it was revealed that hackers managed to manipulate Meta's AI system to gain unauthorized access, highlighting significant security vulnerabilities within the technology. This incident is…
Be Fair! Can Machine Learning Engineering Agents Adhere to Fairness Constraints?
Problem — This work addresses the gap in the literature regarding the evaluation of machine learning engineering (MLE) agents in sensitive domains, particularly concerning fairness and compliance with regulatory standards.…
Meta’s AI agent for WhatsApp Business is now available globally
Meta has officially rolled out its AI agent for WhatsApp Business to users worldwide, marking a significant step in enhancing customer interactions through automated support. This launch is particularly timely…