Infrastructure
World's First 16TB M.2 SSD Launches
Taiwan's Exascend has launched the first 16TB M.2 2280 SSD, priced at 110,000 yuan. Prior 16TB SSDs were SATA 2.5-inch only.
cnBeta (Full RSS) · 193d ago
Full archive
Every story we have kept, newest first.
Looking for the daily editions? → Past editions
Page 1108 of 1375
March 19, 2026
Infrastructure
AMD Ryzen 9750X/9650X Get Clock Boost, 120W TDP
AMD upgraded Ryzen 7 9750X and Ryzen 5 9650X with 0.1GHz higher clocks and TDP doubled to 120W. No new desktop CPU architectures this year from AMD or Intel, just refreshes.
cnBeta (Full RSS) · 193d ago
Research
17yo Cracks AI Core Issue, Musk Praises
A Shenzhen 17-year-old high school student solved a fundamental AI problem after just one year in the field. Elon Musk praised Moonshot AI (Kimi) team's achievement as 'impressive' on social media.
cnBeta (Full RSS) · 193d ago
Infrastructure
TSMC Pushed to Shift 40% Capacity to US
TSMC has pledged $165B in US investments amid pressure but faces demands for $100-200B more. US targets relocating 40% of TSMC's capacity stateside.
cnBeta (Full RSS) · 193d ago
Applications
Wayve Self-Driving Taxi Masters London Chaos
Wayve CEO demos autonomous Ford Mustang navigating King's Cross junctions, controlling speed, steering, brakes, and indicators independently. The AI handles unprotected turns and busy traffic seamlessly.
The Guardian Technology · 193d ago
Applications
Guchi Robots Automate Car Final Assembly
Guchi Robotics, founded in 2019 by Chen Liang, builds robots for installing wheels, dashboards, and windows on cars for BYD and Nio without human help. Currently automates 20% of final assembly, targeting the remaining 80%.
The Guardian Technology · 193d ago
Applications
Ringtime Raises €1.8M for AI Recruitment Agents
Ringtime has raised €1.8M to deploy AI agents that automate blue-collar recruitment by handling calls, voicemails, and interviews. The tool tackles high candidate dropout rates in fast-paced hiring.
The Next Web (TNW) · 193d ago
Audio
Reson8 Secures €5M for Custom Speech AI
Amsterdam-based Reson8 raised €5M in pre-seed funding to scale its customizable speech AI platform. It offers high-precision, industry-specific recognition tailored for European languages, challenging US-dominated solutions.
The Next Web (TNW) · 193d ago
Infrastructure
Guanghuan Compute Hits 4000P Scale
Guanghuan Xintong's AI compute business exceeds 4000P scale with annual contracts over 100M RMB. Deployed in Beijing Kexin Shengcai and Tianjin Zanpu centers for high-performance services.
36氪 · 193d ago
Open Source
Crayfish Symbol Evolves to AI Icon
Crayfish 'xiaolongxia' shifts from 2015 film festival party symbol to OpenClaw AI framework's Chinese nickname, marking tech takeover of entertainment. From hot-money socializing to AI short dramas and comics disrupting long video.
虎嗅 · 193d ago
Infrastructure
Cars, Robots Need 300GB RAM Soon
Micron Technology predicts autonomous cars and humanoid robots will require 300GB or more of DRAM. The memory maker has already grown revenue by $10 billion in a single quarter.
The Register - AI/ML · 193d ago
Video
Xiao Yunque Seedance 2.0 Agent Generates Short Dramas One-Click
The article tests Xiao Yunque's latest AI Agent powered by Seedance 2.0, which understands drama scripts and generates complete short dramas with one click. The author effortlessly created the 'Dragon King Son-in-Law' short drama.
Ifanr (爱范儿) · 193d ago
Business
Cook: iPhone Huge Potential as Digital Core
Tim Cook affirms iPhone's vast tech and experience potential, remaining central to digital life. Apple advances in spatial computing, secret AR glasses, and AI-driven pendant.
IT之家 · 193d ago
Research
Third Net for EV Energy Self-Sufficiency
Academician Sun Fengchun advocates building a third green energy network on highways for 7000B kWh solar potential. EVs become mobile storage via routers for self-sufficiency.
虎嗅 · 193d ago
Infrastructure
China Boosts NEV Chips & Autonomous Driving
MIIT, NDRC, and SAMR held a NEV industry meeting on March 17 to normalize competition, enhance innovation, and expand consumption. Priorities include price monitoring, auto finance regulation, and accelerating auto chips, software, and autonomous driving breakthroughs.
36氪 · 193d ago
Infrastructure
Micron Storage Booms on AI Demand
Micron's next-quarter guidance exceeds market expectations, driven by sustained price hikes in traditional storage products. AI demand is fueling this surge, prompting questions on breaking the industry's cyclical constraints.
钛媒体 · 193d ago
Research
ShuttleEnv: Badminton RL Simulation Environment
ShuttleEnv is an interactive, data-driven simulation environment for badminton, supporting RL and strategic analysis in adversarial sports. Grounded in elite-player match data, it uses probabilistic models for realistic rally dynamics without physics simulation.
ArXiv AI · 193d ago
Research
RideJudge: AI Framework for Ride Disputes
RideJudge introduces a progressive framework aligning visual and logical reasoning for ride-hailing dispute adjudication, addressing limitations in multimodal LLMs. Key innovations include SynTraj for trajectory synthesis, Adaptive Context Optimization, and Ordinal-Sensitive RL.
ArXiv AI · 193d ago
Research
PIER RL Cuts Shipping Fuel Waste 9-Fold
PIER is a physics-informed offline RL framework that learns fuel-efficient, safety-aware maritime routing from historical AIS data and ocean reanalysis, without needing online simulators or forecasts. It reduces mean CO2 emissions by 10% versus great-circle routing and slashes catastrophic fuel waste from 4.8% to 0.5% of voyages on 2023 Gulf of Mexico routes.
ArXiv AI · 193d ago
Models
LLMs Struggle in Clue Reasoning Test
Researchers built a text-based multi-agent Clue game to evaluate LLMs' multi-step deductive reasoning with GPT-4o-mini and Gemini-2.5-Flash agents. Across 18 simulated games, agents secured only four correct wins, struggling with consistent reasoning.
ArXiv AI · 193d ago
Research
Kumiho: Graph-Native Memory for AI Agents
Kumiho introduces a graph-native cognitive memory architecture for AI agents, grounded in formal AGM belief revision semantics. It unifies cognitive memory and versioned assets using primitives like immutable revisions and URI addressing.
ArXiv AI · 193d ago
Models
InfoDensity Rewards Dense Reasoning Traces
Large Language Models generate verbose reasoning traces that waste compute. InfoDensity proposes an RL reward framework using AUC-based and monotonicity signals, weighted by length, to favor concise, high-quality steps.
ArXiv AI · 193d ago
Research
Geometry-Switching Fixes Agent Cascade Failures
New method introduces cascade-aware routing for multi-agent AI systems using spatio-temporal sidecars and adaptive geometry switching between Euclidean and hyperbolic models. It boosts Genesis 3 benchmark win rates from 50.4% to 87.2%, especially in tree-like graphs prone to exponential failures.
ArXiv AI · 193d ago
Research
GenAI Workflow for Participatory Planning Models
Researchers propose a templated LLM workflow to accelerate problem conceptualization in socio-environmental planning under deep uncertainty. It uses models like ChatGPT to extract components from stakeholder descriptions, unify perspectives, and generate Python implementations iteratively.
ArXiv AI · 193d ago
Research
Draft-and-Prune Boosts Auto-Formalization Reliability
Draft-and-Prune (D&P) is an inference-time framework that improves auto-formalization by drafting diverse natural-language plans, pruning incorrect formalizations, and aggregating via majority voting. It achieves 78.43% accuracy on AR-LSAT with GPT-4 and 78% with GPT-4o, outperforming baselines like MAD-LOGIC and CLOVER.
ArXiv AI · 193d ago
Research
CRAFT: Hidden-State RL for Jailbreak Defense
CRAFT is a red-teaming framework that aligns reasoning models using hidden representations to boost robustness against jailbreaks. It combines contrastive learning with RL to separate safe/unsafe trajectories in latent space.
ArXiv AI · 193d ago
Research
AI Scientist via Synthetic Task Scaling
Researchers developed a synthetic environment pipeline for training machine learning agents using the SWE-agent framework. The pipeline generates verified ML challenges grounded in real Huggingface datasets with self-debugging.
ArXiv AI · 193d ago
Research
Anthropic's 80K Claude User Survey
Anthropic surveyed ~80,000 Claude users across languages, revealing strong demand for expert-level AI capabilities with 80%+ reporting practical value. Concerns include reliability issues and loss of autonomy, especially cognitive decline fears in East Asia like Japan.
ITmedia AI+ (日本) · 193d ago
Applications
Metcash Adds AI Search to B2B Platform
Metcash has introduced AI-powered search to its B2B marketplace. The SAP-based platform is now augmented with Coveo's AI technology.
iTNews Australia · 193d ago