A jury ruled in favor of Sam Altman, OpenAI, and Greg Brockman, finding them not liable for Elon Musk's claims of breaking founding contracts and unjust enrichment.

At Dell Technologies World, Dell and NVIDIA unveiled new AI infrastructure including the Dell PowerEdge XE9812 based on NVIDIA Vera Rubin NVL72, delivering up to 10x lower cost-per-token for agentic AI inference. Dell CEO Michael Dell projected worldwide AI infrastructure spending could reach $3-4 trillion by 2030, with token consumption growing 3,400%. NVIDIA CEO Jensen Huang emphasized that demand is 'utterly parabolic.' Enterprise AI has moved from pilots to agentic AI and inference at scale. The Dell AI Factory with NVIDIA provides end-to-end solutions from deskside to data center, including confidential computing and support for open models.

Ian Buck hand-delivered the first NVIDIA Vera CPU systems to Anthropic, OpenAI, SpaceXAI, and Oracle Cloud Infrastructure. Vera is purpose-built for agentic AI workloads, featuring 88 custom cores, 1.2 TB/s memory bandwidth, and 50% faster per-core performance.

MemPrivacy is a framework from MemTensor, HONOR Device, and Tongji University that protects user privacy in LLM-powered agents by replacing sensitive data with typed placeholders on-device before cloud transmission, preserving memory utility. It introduces a four-level privacy taxonomy (PL1-PL4) with configurable masking thresholds. The best model achieves 85.97% F1 on MemPrivacy-Bench, and memory utility loss is limited to within 1.6% compared to up to 41.87% with irreversible masking.

Skilled is a terminal dashboard that reads local history files from AI coding tools like Claude Code, OpenCode, Codex, Grok, and Droid, providing real-time usage statistics including frequency, weekly trends, hourly distribution, per-project breakdowns, and audit heuristics. Fully local, zero network, no accounts, no configuration.

Berkeley Lab's MatterChat framework bridges large language models with physics-based interatomic potential models, enabling AI to understand 3D atomic structures for materials science. It outperforms GPT-4 in predicting material properties and could accelerate discovery of new materials.

Modern language models face an optimization challenge due to uneven token distributions: parameters associated with common tokens receive frequent gradient updates while rare token parameters may go hundreds of steps without a meaningful signal. Standard SGD uses the same learning rate for all parameters, causing rare token weights to remain near initialization. Adam's variance normalization automatically gives larger effective learning rates to underrepresented features, enabling them to converge quickly. A controlled NumPy experiment with a six-token vocabulary spanning four orders of magnitude in frequency shows that SGD yields weights near 0.15 for the rarest token (true weight 1.0), while Adam keeps all rare token weights close to 1.0.

As attackers get more sophisticated and persistent, IT workers have to step up their game as well. Here's how to do that in 2026.

SandboxAQ has integrated its Large Quantitative Models (LQMs) with Anthropic’s Claude, enabling researchers to use natural language to run complex simulations for drug discovery, materials science, and more. The first model available is AQCat Adsorption Spin for catalyst discovery, with drug discovery models coming soon.

The Agentic AI Foundation (AAIF) announced the addition of 43 new members in the past quarter, including 4 Gold, 27 Silver, and 12 Associate members, bringing total membership to 180 organizations. New members span financial services, government, academia, and enterprise technology, contributing expertise across AI infrastructure. AAIF Executive Director Mazin Gilbert emphasized the industry shift toward open standards for scalable agentic AI. Gold members include F5, GoDaddy, Stripe, and TRON.

The pope’s encyclical will address ‘the protection of the human person in the age of AI’, the Vatican says. In the first major text of his papacy, Pope Leo will address the rapid rise of artificial intelligence.

Bug bounties have served as one of cybersecurity's core pressure valves for decades, giving independent researchers a structured way to disclose vulnerabilities before attackers can exploit them. But a deluge of AI-assisted reports is upending parts of that system.

After three weeks of testimony, the Musk v. Altman lawsuit was dismissed due to the statute of limitations, but the trial revealed a widespread lack of credibility among AI leaders. The case involved control struggles at OpenAI, safety review issues, and mutual accusations. Public concern about AI is at an all-time high, and the behavior of industry leaders has deepened distrust.

Learn how to prompt Amazon Nova 2 Lite for content moderation using structured and free-form approaches based on the AILuminate standard. The techniques allow updating moderation policies via prompt edits without retraining, and benchmark results against other models are provided.

Security readiness has overtaken cost and skill shortages as the top barrier to AI adoption, according to a new Linux Foundation report. 67% of respondents feel pressure to accelerate AI deployment despite security concerns. 48% of organizations cite security as the top obstacle, up from 17% in 2024. The report also finds that AI is driving tech job growth, with net hiring projected to rise 31% in 2026, and that upskilling is the preferred strategy to close capability gaps.

Ofcom to update codes of practice amid rise in ‘revenge porn’ and AI-generated deepfakes targeting women and girls. Social media, messaging platforms and online forums that publish intimate image abuse – often intended to humiliate women and girls – are being instructed to follow new guidelines to stop it spreading. Ofcom said it would change its codes of practice to force service providers to detect and quash intimate image abuse – sometimes called “revenge porn” – and crack down on AI-generated deepfakes. A wave of deepfakes emerged in January when Elon Musk’s Grok AI was widely used to create sexualised videos of women in bikinis.

Wave Terminal is a terminal app on steroids that integrates AI, file manager, web browser, system info, and process tools. This article covers installation, usage, and AI features.

Microsoft announced at Open Source Summit North America 2026 the public preview of Azure Linux 4.0 and general availability of Azure Container Linux, alongside contributions to the Agentic AI Foundation (AAIF) and open governance tools for agentic systems.

After around two hours of deliberation, the jury has reached a unanimous verdict in Musk v. Altman, the tech trial of the year. The group found that two claims were barred by the statute of limitations, and a third failed thanks to the dismissal of one of these.

Aderant used Amazon Quick to unify search across six vendor systems and automate documentation workflows, achieving 90% faster search times and 75% documentation acceleration. The implementation led to significant efficiency gains, including 95% reduction in client history research time and 200% increase in documentation output.

Odysseyml launches Agora-1, a multi-agent world model that enables real-time simulation for up to four human or AI participants simultaneously. By decoupling simulation from rendering, it ensures consistent multi-perspective generation, suitable for AI research, game development, and robotics.

Remote control for GitHub Copilot CLI sessions is now generally available on github.com and GitHub Mobile. Developers can start a session in VS Code or the CLI, then monitor and adjust it from another device. Features include real-time monitoring, mid-flight instruction changes, permission approvals, and a seamless cross-device workflow, with privacy by default.

Manus launches Scheduled Tasks 2.0, enabling context-aware recurring tasks that run in the same task context, reuse project setups, and add scheduled actions to Manus-built web apps.

Configuring Hyprland is challenging; AI can help but requires corrections. Here's how Codex performed.

Deep Agents was previously designed in a generic way to work well across model families. Today we’re adding model-specific profiles to adjust prompts, tools, and middleware. We ship profiles for OpenAI, Anthropic, and Google models out of the box, which we see leads to a 10–20 point jump on a subset of tau2-bench over the default harness.

This article presents a parameter-efficient fine-tuning approach using LoRA and DoRA to adapt NVIDIA Cosmos Predict 2.5 for robot video generation on a single GPU. It covers data preparation, adapter initialization, training with rectified flow loss, inference, and evaluation metrics.

Amazon's upgraded AI assistant, Alexa Plus, can now generate podcasts on virtually any topic, allowing users to steer conversations and adjust length before generation. The feature draws from 200 partner news publications.

Omnia launches Insights, a feature that generates prioritized action plans to improve brand visibility in AI engines, based on real citation data. Each plan includes impact level, rationale, implementation steps, and citation gap analysis.

This post demonstrates implementing four Lambda-based custom code evaluators for a financial market-intelligence agent, covering schema validation, numerical accuracy, workflow compliance, and PII detection. It shows how to combine custom evaluators with built-in ones and call other AWS services for grounded fact-checking and real-time alerting.

This article discusses the structural shift in the data job market: SQL and Python have been demoted from differentiators to prerequisites. Analysis of over 700 data scientist postings from early 2026 shows machine learning and AI skills are now top demands. Four new differentiators are identified: data modeling, performance optimization, infrastructure awareness, and practical AI skills (LLMs, RAG, prompt engineering). The article provides actionable advice for acquiring each skill.