All Updates
Page 700 of 752
February 18, 2026
Blackwell 50x Hopper Efficiency in DeepSeek Test
NVIDIA Blackwell Ultra achieves 50x tokens per MW throughput vs Hopper in DeepSeek-R1 tests, reducing million-token costs to 1/35th. It previews Rubin platform with 10x further gains. Key enablers include 130 TB/s NVLink and NVFP4 precision.
Chunwan Spotlights Yibin's AI City Rise
The Spring Festival Gala's 7-minute Yibin segment showcases the city as an emerging AI new city. It depicts Yibin's evolution from liquor capital to power battery hub and now AI innovation center. This reflects tech-driven green transformation under new quality productive forces.
Caltech Brings Fiber Performance to Silicon Chips
Caltech researchers achieved a breakthrough by developing technology for ultra-low loss optical signal transmission on silicon wafers. Performance in the visible light band nears traditional fiber optic levels. This advances on-chip photonics significantly.
Perplexity Abandons Ads for User Trust
AI startup Perplexity is ditching its advertising business amid competitors' ad integrations. The decision prioritizes user trust over quick monetization. This bucks the trend in AI commercialization.
Google-Ormat Geothermal Deal Powers Nevada DCs
Google plans to procure geothermal energy from Ormat Technologies to power its Nevada data center cluster. The deal enables 24/7 clean electricity supply for reliable operations. This supports sustainable energy for high-demand compute workloads.
Edge 145 Enhances Passwords & PDF Read-Aloud
Microsoft released Edge 145 stable (145.0.3800.58) to all users with password management, autofill, and PDF reading improvements. Enterprise/IT features included alongside security fixes. Everyday users gain better productivity tools.
Apple Leaps to M5 in AI Cloud Upgrade
Apple completed a major hardware upgrade for its Apple Intelligence cloud infrastructure. Proprietary servers jumped from M2 Ultra directly to self-developed M5 chips, skipping M3 and M4. This delivers substantial boost to cloud AI compute power.
6000 Gamers Pick DLSS 4.5 as Top Quality
ComputerBase organized a blind test with 6,000 gamers comparing NVIDIA DLSS 4.5, AMD FSR 4, and native rendering without labels. Players preferred DLSS 4.5 for superior image quality over native resolution and FSR 4.
Qualcomm Invests Up to $150M in India AI Fund
Qualcomm plans to invest up to $150 million to support India's growing tech and AI startup ecosystem. The funds will be deployed through Qualcomm Ventures to startups at various stages. Focus areas include AI applications in automotive, IoT, robotics, and mobile sectors.
WD Raises $3.09B Selling SanDisk Shares
Western Digital plans to raise $3.09 billion by selling shares in its flash memory subsidiary SanDisk, spun off nearly a year ago. Shares priced at $535-$555, up to 9.4% discount from $590.59 close.
X-MAP Profiles Misclassifications in Spam Detection
X-MAP is an explainable framework combining SHAP attributions and NMF to create topic profiles for correctly classified spam/phishing vs. legitimate messages. It detects misclassifications via Jensen-Shannon divergence from these profiles. Experiments achieve 0.98 AUROC and recover 97% of false rejections when used as a repair layer.
WGAN Boosts Synthetic Population Diversity
New method uses joint WGAN with gradient penalty to synthesize populations from multi-source data, tackling diversity and feasibility issues. Introduces regularization term for generator loss, outperforming baselines in recall (+7%), precision (+15%), and overall similarity (88.1 vs 84.6). Enhances agent-based models in transportation and urban planning.
WAC Boosts Web Agents with World Models
WAC integrates multi-agent collaboration where an action model consults a world model for strategic guidance on web tasks, grounding suggestions into executable actions. It employs a two-stage deduction chain with consequence simulation and judge model scrutiny for risk-aware action correction. Experiments show 1.8% gains on VisualWebArena and 1.3% on Online-Mind2Web.
Trace Rewriting Blocks LLM Distillation Theft
Researchers introduce methods to rewrite teacher model reasoning traces, deterring unauthorized knowledge distillation while embedding verifiable watermarks. Techniques include LLM-powered rewriting and gradient-based approaches that preserve answer correctness. Experiments demonstrate strong anti-distillation effects with maintained or improved teacher performance and reliable watermark detection.
Tech Billionaires Flock to Delhi AI Summit
Silicon Valley leaders from Google, Anthropic, and OpenAI will attend India's week-long AI Impact Summit hosted by PM Narendra Modi. The event gathers thousands of tech executives, officials, and AI safety experts alongside global south leaders like those from Kenya and Indonesia. Discussions focus on wresting control over rapidly advancing AI technology.
Secure Energy-Efficient Agentic AI Wireless Networks
This arXiv paper introduces a secure wireless agentic AI network with a supervisor AI agent assigning others for cooperative reasoning while unselected agents jam eavesdroppers. It formulates an energy minimization problem optimizing agent selection, BS beamforming, and transmission power under latency and accuracy constraints. Proposed ASC and LAW schemes reduce energy by up to 59.1%, validated on a Qwen-based system with strong benchmark performance.
Ruva: Transparent On-Device Personal AI Graph Reasoning
Ruva introduces a 'Glass Box' architecture for Personal AI using Personal Knowledge Graphs, allowing users to inspect and precisely redact AI knowledge. It replaces vector databases with graph reasoning to eliminate hallucinations, ensure accountability, and enable the 'Right to be Forgotten.' A demo and project are available online.
ResearchGym: AI Agents Research Benchmark
ResearchGym introduces a benchmark with five containerized environments from ICML, ICLR, and ACL papers, totaling 39 sub-tasks, where agents propose hypotheses and run experiments to beat baselines. GPT-5-powered agents show a capability-reliability gap, succeeding in only 6.7% of evaluations and completing 26.5% of sub-tasks. It identifies key failure modes like impatience and poor resource management, while occasionally achieving SOTA results unreliably.
Reasoning LLMs Beat Conversational in Risky Choices
A study of 20 LLMs reveals clustering into rational reasoning models (RMs), insensitive to framing and order, and less rational conversational models (CMs) with human-like biases and a description-history gap. RMs match rational agents across explicit and experience-based prospects. Mathematical reasoning training differentiates RMs from CMs.
R2U-Net Hits 0.900 DSC in Brain Tumor Segmentation
Introduces Attention-Gated Recurrent Residual U-Net (R2U-Net) Triplanar model for glioma segmentation, achieving 0.900 Dice Score on BraTS2021 Whole Tumor. Integrates residual, recurrent, and attention mechanisms for efficiency. Extracts features for survival prediction with 45.71% accuracy.