All Updates

Page 671 of 1671

May 11, 2026

πŸ”
Google AI Blogβ€’83d ago

AI-Powered Google Finance Launches in Europe

This week, the new AI-powered Google Finance is launching across Europe. It offers full local language support and a reimagined experience with a suite of powerful capabilities.

#fintech#europe-expansion#language-ai
πŸ‡¨πŸ‡³
cnBeta (Full RSS)β€’83d ago

AI Code Surge is Linux Kernel New Norm

Linus Torvalds released Linux 7.1-rc3, noting exceptionally large kernel patch sizes this cycle. This surge is largely due to developers heavily using AI coding tools. He states this is now the established new normal in Linux kernel development.

#ai-coding#kernel-patches#dev-trends
πŸ‡¨πŸ‡³
TechNodeβ€’83d ago

ByteDance Hikes AI Spend 25% to $28B

ByteDance has boosted its AI infrastructure spending by 25% to RMB 200 billion ($28 billion) this year. The increase stems from rising memory chip costs pushing up expenses. The initial budget was set at RMB 160 billion.

#capex-increase#memory-chips#china-ai
πŸ’°
ι’›εͺ’体‒83d ago

Accelink Surges on Optical Module Self-Reliance Shift

Accelink Technologies hits the daily limit up amid changing investor focus in photonics. Funds shift from speculating on orders to emphasizing autonomous controllability. The domestic substitution trend in optical modules takes flight.

#photonics#china-supply-chain
πŸ“Š
Bloomberg Technologyβ€’83d ago

Emerging Stocks Record High on AI Bets

Emerging-market equities are set to close at a record high driven by AI trades. Investors dismissed concerns over stalled US-Iran peace talks. This reflects strong appetite for AI amid geopolitical tensions.

#stock-rally#ai-trades#emerging-markets
πŸ‡¨πŸ‡³
cnBeta (Full RSS)β€’83d ago

Win11 Xbox Mode Boosts Nvidia AMD GPUs

Tests reveal Windows 11 Xbox mode delivers major performance gains for Nvidia and AMD GPUs. Microsoft counters SteamOS threat via Project K2 OS improvements. The mode turns PCs into full Xbox experiences.

#gpu-optimization#windows-gaming#project-k2
πŸ‡¨πŸ‡³
cnBeta (Full RSS)β€’83d ago

Jensen Huang: Seize AI Revolution Opportunity

Nvidia CEO Jensen Huang spoke at Carnegie Mellon University's 128th commencement on May 10. He compared graduates' careers starting with the AI revolution to his own at the PC era's dawn. Huang urged them not to fear AI but embrace the once-in-a-generation opportunity.

#ai-careers#nvidia-ceo#graduation-speech
βš›οΈ
量子位‒83d ago

Fields Winner: ChatGPT 5.5 Pro Outputs Paper in 17 Mins

Fields Medalist Terence Tao tested ChatGPT 5.5 Pro, producing paper-level math results in 17 minutes, alarming for math majors. Tao notes that 'digesting' AI outputs remains a human task.

#math-reasoning#fields-medalist#ai-research
πŸ“Š
Bloomberg Technologyβ€’83d ago

Pictet Fund Allocates 30% Cash to AI Stocks

Pictet Asset Management's $3.5 billion multi-asset fund has boosted equity exposure significantly. It shifted up to 30% of cash-equivalent holdings into AI heavyweights. Investments span Asia and the US amid risk revival.

#fund-strategy#ai-investment#risk-revival
🐼
Pandailyβ€’83d ago

SENASIC Tops China Auto Wireless SoC

CATL-backed SENASIC has passed its Hong Kong IPO hearing, becoming China's largest automotive wireless SoC company. CATL's rare three consecutive investment rounds underscore this milestone. It signals major EV battery giants expanding into automotive sensing chips.

#auto-chip#ev-sensing#china-ipo
πŸ”₯
36ζ°ͺβ€’83d ago

Advanced Chip Packaging Lead Times Hit 1+ Year

Demand for advanced packaging and testing equipment surges beyond expectations amid seal test factories' expansions. Upstream supply chains face order backlogs and customer capacity competition. Some equipment delivery now exceeds 1 year, delaying new fabs.

#semiconductor#supply-chain#packaging-delay
πŸ”₯
36ζ°ͺβ€’83d ago

QClaw Launches File Space Feature

Tencent Cloud's QClaw officially launched the File Space feature. A single authorization connects local files, Tencent Docs, and IMA knowledge base seamlessly. It closes the loop from material search, AI processing, output, to collaboration in one workbench.

#cloud-feature#ai-workflow#file-integration
πŸ“Š
Bloomberg Technologyβ€’83d ago

AI Chip Boom Pressures Korea-Taiwan Rates

South Korea and Taiwan's AI-fueled chip booms are swelling trade surpluses. This pressures central banks to raise interest rates later this year. Goldman Sachs highlights a K-shaped economic recovery.

#economy#trade-surplus#central-banks
πŸ’°
ι’›εͺ’体‒83d ago

Bosch Advances in Boundless Auto Fusion

Bosch is forging ahead without boundaries in smart automotive innovation. Competition in intelligent vehicles shifts from single-point feature stacking to cross-domain fusion challenges.

#smart-automotive#cross-domain-fusion#auto-strategy
πŸ“Š
Bloomberg Technologyβ€’83d ago

Korea Eyes Hyundai Robots for Army Shortage

South Korea's military is exploring a partnership with Hyundai Motor Co. for front-line robotics. This addresses deepening troop shortages via AI-powered unmanned systems. Seoul is accelerating investments in such tech.

#robotics#military#unmanned-systems
πŸ“Š
Bloomberg Technologyβ€’83d ago

AI Targets Minds Over Jobs

Tom Slater's paper argues AI's primary threat is to human judgment, learning, and expertise, not job displacement. In a Bloomberg podcast, Merryn Somerset Webb discusses this with Slater, manager of Scottish Mortgage Investment Trust. They explore implications for the future workforce and investment strategies.

#cognition-risk#workforce-impact#ai-investing
πŸ“„
ArXiv AIβ€’83d ago

Weblica Scales RL for Web Agents

Weblica is a framework for scalable, reproducible web environments using HTTP caching to replay stable visuals and LLM synthesis grounded in real websites. It enables RL training across thousands of diverse tasks, capturing web complexity. Weblica-8B outperforms open-weight baselines on navigation benchmarks with fewer inference steps.

#web-agents#llm-synthesis
πŸ“„
ArXiv AIβ€’83d ago

Survey: LLM Agent Memory Evolution

This arXiv survey proposes a three-stage evolutionary framework for LLM agent memory: Storage (trajectory preservation), Reflection (refinement), and Experience (abstraction). It analyzes drivers like long-range consistency, dynamic environments, and continual learning. It highlights proactive exploration and cross-trajectory abstraction in the advanced Experience stage.

#agent-memory#continual-learning
πŸ“„
ArXiv AIβ€’83d ago

State Rep & Termination for Recursive Reasoning

This paper proposes representing reasoning states in recursive systems as epistemic state graphs that encode claims, evidential relations, open questions, and confidence weights. It introduces the order-gap metricβ€”the distance between states from expand-then-consolidate vs. consolidate-then-expandβ€”to signal when to terminate iteration. A main result provides a necessary and sufficient condition for the linearized order-gap to be informative near fixed points, with applications to agents, tree-of-thoughts, and more.

#epistemic-state#order-gap
πŸ“„
ArXiv AIβ€’83d ago

Spectral Tool Detects Hidden AI Coalitions

New method detects coalitions in multi-agent AI via mutual-information graphs from hidden states and spectral partitioning. Validated in MARL environments and LLMs, it uncovers structures before behavioral changes and outperforms scalar MI metrics. Offers scalable monitoring for emergent AI organizations.

#multi-agent#ai-safety#representations
Page 671 of 1671