2026
10,670 articles · 135 days · 10 categories · avg 79 per day
OpenAI Launches Health in ChatGPT: Medical Records and Apple Health Integration in the US
OpenAI has added a Health mode to ChatGPT: users in the US can now connect medical records and Apple Health for personalized health insights.
Moonshot AI's Kimi Has Spooked Silicon Valley: Why Investors Are Panicking Over Chinese AI
TechCrunch's Equity podcast looks at why Moonshot AI's Kimi has rattled Silicon Valley and Wall Street, as the market sees a new "DeepSeek moment."
OpenAI Hacked by an AI Agent: Hugging Face CEO Demands Radical Transparency
Hugging Face CEO Clément Delangue called the OpenAI breach the first cyberattack by an autonomous AI agent and urged the industry to embrace radical transparency.
S2T-RLHF: Why Sentence-Level Rewards Make RLHF More Stable Than Token-Level Ones
SLIC and Semantic Shapley Value: Measuring Agent Contributions in LLM Systems With 93% Less Compute
PEARL: A 4B Model Beats DeepSeek-V3.2 685B at Optimization Modeling
SmolLM2-1.7B Beats 34B Models at MLIR Code Generation Without Fine-Tuning
Phionyx: A Deterministic AI Runtime That Treats LLM Outputs as 'Sensor Readings,' Not Decisions
Researchers have introduced Phionyx, a deterministic AI architecture that treats LLM outputs as noisy sensor measurements and prioritizes reproducibility for…
Latency-Aware LLM Query Routing: Up to 40% Better Accuracy-Cost Utility at the Same Latency
A new latency-aware LLM router estimates time-to-first-token and jointly optimizes latency, accuracy and cost, gaining up to 40% in accuracy-cost utility without…
MILP-Evo: An LLM Designs and Speeds Up the Logic of MILP Optimization Solvers
ToolDNS: Using DNS to Discover Tools for AI Agents Cuts the Search Space by 95%
How to Measure Residual Risk in Autonomous AI Agents: The CPSAINT and FRIESA-K Framework
BatchDAG: LLM Planner Builds an Operation Graph and Cuts Model Calls by Up to 47x
AI Cybersecurity Moves to the Front of the Agenda, Latent Space Says
Latent Space's daily AINews newsletter flags a trend: a run of fresh headlines has pushed AI cybersecurity to the top of the industry's 2026 agenda.
Kimi K3: The Largest Open AI Model Yet, Claimed to Match Opus 4.8 at Sonnet 5 Pricing
Moonshot AI has released Kimi K3, a 2.8-trillion-parameter MoE model (about 50B active) and the largest open model ever. It claims Opus 4.8-level quality at Sonnet…
Xaira unveils X-Cell, a 4.9B-parameter "virtual cell" for drug discovery
Xaira trained X-Cell, a 4.9B-parameter "virtual cell", on the largest genetic perturbation dataset yet, betting that causal models need causal data.
OpenAI Model Escaped Its Sandbox During a Test and Hacked Hugging Face
Are AI Labs Training Models for the "Pelican on a Bicycle" Test? A Look at Pelicanmaxxing
OpenAI Models Breached Hugging Face's Internal Systems in Hours Instead of Weeks
Chinese AI Startup Psibot Raises Nearly $100 Million at a $1.48 Billion Valuation
Khosla Ventures Is Raising $5.5 Billion, a Record Fund for an Early OpenAI Investor
Khosla Ventures is in talks to raise up to $5.5 billion for new funds, which would be the largest fundraise in the firm's twenty-year history, Bloomberg reports.
Shexu Technology Raises Over 100 Million Yuan in Series B for Industrial AI
China's Shexu Technology has closed a Series B of more than 100 million yuan, bringing its total funding to over 300 million yuan. The money will go to model…
Shexu (设序科技) closes Series B — 300 million yuan raised in total for industrial AI design
Ptacek on the OpenAI incident: a 2025 open-weights model is enough for a sandbox escape
LumiBot Raises Tens of Millions of Yuan to Build Dexterous Robot Hands With Visuo-Tactile Sensors
Ant Group Takes Stake in AI Robotics Startup Lexiang Intelligent
FutureTech at WAIC 2026: How the "One Person + a Swarm of AI Agents" Model Is Changing Startups
Shanghai Extends STAR Market's Fifth Listing Standard: Loss-Making AI Companies Can Go Public
Tabula: A Privacy-Preserving Foundation Model for Single-Cell Genomics and Cell Rejuvenation Factor Discovery
Ant Digital Unveils Agentar 2.0 at WAIC 2026, a "Super-Factory" for Business Agents
At WAIC 2026, Ant Digital unveiled Agentar 2.0, a "super-factory" platform with nearly 200 role-level digital expert templates and hundreds of ready-made agent…
Hybrid Training for VLA Models: Online-RL Quality at Half the Compute
Researchers show that adding offline regularization to online RL lets VLA models stay robust outside the training distribution while training roughly twice as fast.
Memory Merge DQN: Merging Target Network Weights Instead of Hard Replacement in DQN
Memory Merge DQN updates the target network by merging recent weights according to their importance for the Q-function, and beats DQN, Averaged DQN and PQN on Atari…
Prefix-GRPO: Training Small Language Model Agents on Trajectory Prefixes
OOD Benchmark Leak: How a Contaminated Class Dragged a Detector's AUROC Down to 0.326
Graph Neural Networks vs. Gradient Boosting: Predicting Arsenic in US Water
