Meta

This research paper evaluates systematic biases in LLM-as-a-Judge evaluation pipelines, identifying style bias as a major issue. The study compares nine debiasing strategies across multiple model families and benchmarks to improve evaluation reliability.

The Chinese government has blocked Meta's $2 billion acquisition of the AI startup Manus, citing national security concerns. The decision highlights the increasing difficulty of cross-border tech deals amid the intensifying US-China AI rivalry.

China has blocked a deal involving Meta and Manus AI, signaling a significant shift in global tech investment and AI regulation. This move highlights the increasing tension between international AI development and nationalistic regulatory policies.

China's NDRC has blocked Meta's $2 billion acquisition of the agentic AI startup Manus. The decision requires the parties to unwind the transaction, despite the company having already relocated its headquarters and staff to Singapore.

China has blocked Meta's attempt to acquire the AI startup Manus for $2 billion. The decision prevents the massive expansion of Meta's artificial intelligence capabilities through this acquisition.

China has blocked Meta's attempt to acquire the AI startup Manus. This move represents a significant regulatory intervention in the cross-border acquisition of artificial intelligence companies.

Meta has partnered with Overview Energy to secure renewable energy for its data centers using space-based solar technology. The deal involves using satellites to beam infrared light to terrestrial solar farms, addressing the energy demands of growing AI compute power.

The article evaluates Amazon's valuation following significant deals involving AI chips and cloud services with Meta and Anthropic. It focuses on how these strategic partnerships and hardware investments impact Amazon's market position.

Researchers propose a new three-layer architecture that decouples personal data from shared model weights using composable LoRA adapters and deletable user proxies. This method allows for deterministic unlearning and prevents user data from being stored in the base model, enhancing privacy and preventing data extraction.

Meta and Microsoft are implementing significant workforce reductions, including layoffs and voluntary buyouts, as they shift resources toward massive AI investments. While companies cite the need to offset AI spending, some industry observers suggest these cuts may be part of broader corporate restructuring rather than a direct result of AI disruption.

Meta employees are using an internal leaderboard to track and compete based on their AI token usage. This trend, dubbed 'tokenmaxxing,' serves as a way for workers to demonstrate AI proficiency and productivity through high-volume model interaction.

Broadcom is strengthening its relationships with Meta and Google Cloud to support their AI infrastructure needs. The company is focusing on developing custom AI chips and networking solutions to meet growing demand from major tech players.

Meta is introducing new AI-driven safety features specifically designed to protect teenage users. These tools aim to mitigate potential risks and enhance the digital safety experience for younger demographics on their platforms.

Meta is implementing a tool called the Model Capability Initiative to monitor US-based employees' computer activity. This system records keystrokes, mouse movements, and screenshots to provide data for training the company's AI agents.

Startup CEOs are increasingly viewing high AI compute costs as a status symbol and a metric of success. This 'tokenmaxxing' trend involves replacing traditional human roles with high-cost AI token usage to scale businesses with minimal headcount.

Meta is implementing a new internal tool to record employee keystrokes and mouse movements to train its AI models. The company states this data is necessary to build more capable agents that understand how humans interact with computers.

Meta plans to track the mouse movements, clicks, and keystrokes of its US employees to create training data for AI agents. This initiative aims to help AI models better understand and execute everyday computer-based tasks like navigating menus.

Broadcom is extending its partnerships with Meta and Alphabet to develop custom AI chips. This update highlights the ongoing collaboration to provide specialized hardware for large-scale AI workloads through 2026.

Meta is reportedly collaborating with Broadcom to develop custom AI chips, aiming to reduce its reliance on Nvidia. This strategic move focuses on creating specialized hardware tailored to Meta's specific infrastructure needs.

Meta is reportedly developing a photorealistic, 3D AI avatar of Mark Zuckerberg to interact with employees. The bot is being trained on his public mannerisms and strategic perspectives to provide managerial guidance and feedback.

Civil society groups, including the ACLU, are calling on Meta to abandon plans to integrate facial recognition into its AI-powered smartglasses. The groups argue that such features could compromise privacy and facilitate surveillance by stalkers or government agencies.

Meta is increasing the price of its Quest VR headsets by $50–$100 due to rising costs of critical components like memory chips. This price hike is partially driven by Meta's massive capital expenditure shift toward AI infrastructure and superintelligence development.

The European Union is considering a mandate for Meta to alter its WhatsApp AI policies as part of an ongoing antitrust investigation. This move aims to address competition concerns regarding how Meta integrates its AI capabilities into its messaging platform.

Meta is developing a photorealistic, AI-powered 3D version of Mark Zuckerberg to interact with employees. The project aims to replicate his mannerisms and strategic thinking to maintain connection with the workforce.

Yann LeCun, a prominent figure in AI development formerly at Meta, has successfully raised over $1 billion for his new startup venture. The funding marks a significant investment in the next generation of AI leadership and innovation.

Meta and Hugging Face have introduced OpenEnv, an open-source framework designed to evaluate AI agents in real-world environments rather than simulations. The framework uses a standardized API to test how agents handle complex tasks like temporal reasoning and multi-agent coordination using real tools like calendars and browsers.