Tensormesh raises $4.5M to squeeze more inference out of AI server loads
With the AI infrastructure push reaching staggering proportions, there’s more pressure than ever to squeeze…
Meet OAT: The New Action Tokenizer Bringing LLM-Style Scaling and Flexible, Anytime Inference to the Robotics World
Robots are entering their GPT-3 era. For years, researchers have tried to train robots using the same autoregressive (AR) models that power large language models (LLMs). If a model can predict the next word in a sentence, it should be able to predict the next move for a robotic arm….
ASUS Zenbook S16, Zenbook 14, and Vivobook AI laptops open for pre-bookings ahead of 12th February India launch
ASUS has announced pre-bookings in India for its latest lineup of Zenbook and Vivobook laptops powered by AMD Ryzen AI processors. Headlining the portfolio is the ASUS Zenbook S16 (UM5606GA), alongside new Zenbook and Vivobook models aimed at AI-assisted productivity, creative workflows, multitasking, and everyday computing. The new range brings Copilot+ PC capabilities, dedicated NPUs,…
Meta’s Oversight Board takes up permanent bans in landmark case
Meta’s Oversight Board is tackling a case focused on Meta’s ability to permanently disable user accounts. Permanent bans are a drastic action, locking people out of their profiles, memories, friend connections, and, in the case of creators and businesses, their ability to market and communicate with fans and customers. This is the first time in…
A Coding Guide to Demonstrate Targeted Data Poisoning Attacks in Deep Learning by Label Flipping on CIFAR-10 with PyTorch
In this tutorial, we demonstrate a realistic data poisoning attack by manipulating labels in the CIFAR-10 dataset and observing its impact on model behavior. We construct a clean and a poisoned training pipeline side by side, using a ResNet-style convolutional network to ensure stable, comparable learning dynamics. By selectively flipping a fraction of samples from…
OpenAI introduces ChatGPT for Healthcare with HIPAA-compliant AI products
OpenAI has announced OpenAI for Healthcare, a new suite of secure AI products designed to help healthcare organizations deliver more consistent, high-quality patient care while reducing administrative workloads and supporting HIPAA compliance. The initiative includes ChatGPT for Healthcare, which is available starting today and is already rolling out to major healthcare institutions such as AdventHealth,…
TikTok agrees to deal to cede control of US business to American investor group
TikTok has reached a deal to cede a substantial portion of its U.S. operation to a group of American investors, thus ending a years-long tussle in which the federal government has sought to force the platform to do just that. The new partnership is described as a “new TikTok U.S. joint venture” in an internal…
OpenAI has Released the ‘circuit-sparsity’: A Set of Open Tools for Connecting Weight Sparse Models and Dense Baselines through Activation Bridges
OpenAI team has released their openai/circuit-sparsity model on Hugging Face and the openai/circuit_sparsity toolkit on GitHub. The release packages the models and circuits from the paper ‘Weight-sparse transformers have interpretable circuits‘. https://arxiv.org/pdf/2511.13653 What is a weight sparse transformer? The models are GPT-2 style decoder only transformers trained on Python code. Sparsity is not added after…
WhatsApp update brings Missed Call Messages, Voice Chat Reactions, and more
WhatsApp has announced a new holiday update, introducing a wide range of features across calls, chats, Status, and Channels. The update focuses on improving communication flexibility, engagement, and content creation, and is rolling out globally across mobile and desktop platforms. New Call Features Enhance Communication WhatsApp is bringing several upgrades to calling, making missed interactions…
Sakana AI raises $135M Series B at a $2.65B valuation to continue building AI models for Japan
As U.S. giants like Google, OpenAI and Anthropic race to develop the large language models that underpin their AI products, startups such as Sakana AI, Mistral AI, DeepSeek and AI21 Labs are carving out their own niche with specialized models designed for specific regions, industries, or unique features. In pursuit of that goal, Tokyo-based Sakana…
AI Interview Series #2: Explain Some of the Common Model Context Protocol (MCP) Security Vulnerabilities
In this part of the Interview Series, we’ll look at some of the common security vulnerabilities in the Model Context Protocol (MCP) — a framework designed to let LLMs safely interact with external tools and data sources. While MCP brings structure and transparency to how models access context, it also introduces new security risks if…