AllLLMRoboticsStartupsResearchEthicsBusinessHardwareAutonomousMedicineCreative

All news

10783
Chinese AI Startup Psibot Raises Nearly $100 Million at a $1.48 Billion Valuation
Startups·Bloomberg Tech

Chinese AI Startup Psibot Raises Nearly $100 Million at a $1.48 Billion Valuation

Jul 26, 2026·1 min
Khosla Ventures Is Raising $5.5 Billion, a Record Fund for an Early OpenAI Investor
Startups

Khosla Ventures Is Raising $5.5 Billion, a Record Fund for an Early OpenAI Investor

Jul 26, 2026·1 min
Shexu Technology Raises Over 100 Million Yuan in Series B for Industrial AI
Startups

Shexu Technology Raises Over 100 Million Yuan in Series B for Industrial AI

Jul 26, 2026·2 min
China to Tie New AI Data Centers to Green Power: 2026–2030 Renewable Energy Plan
Business

China to Tie New AI Data Centers to Green Power: 2026–2030 Renewable Energy Plan

Jul 26, 2026·2 min
Shexu (设序科技) closes Series B — 300 million yuan raised in total for industrial AI design
Autonomous·36Kr (36氪)

Shexu (设序科技) closes Series B — 300 million yuan raised in total for industrial AI design

Chinese startup Shexu, which builds industrial AI for engineering design, closed a Series B of more than 100 million yuan, bringing its total funding to 300 million…

Jul 26, 2026·3 min
Ptacek on the OpenAI incident: a 2025 open-weights model is enough for a sandbox escape
Autonomous·Simon Willison

Ptacek on the OpenAI incident: a 2025 open-weights model is enough for a sandbox escape

Security researcher Thomas Ptacek says even a 2025 open-weights model could repeat the sandbox escape and network scanning seen in the OpenAI incident. No frontier…

Jul 26, 2026·2 min
LumiBot Raises Tens of Millions of Yuan to Build Dexterous Robot Hands With Visuo-Tactile Sensors
Robotics·36Kr (36氪)

LumiBot Raises Tens of Millions of Yuan to Build Dexterous Robot Hands With Visuo-Tactile Sensors

Chinese startup LumiBot, founded by a former Wall Street investor, has closed an angel round of tens of millions of yuan to build dexterous robot hands with its own…

Jul 26, 2026·3 min
Ant Group Takes Stake in AI Robotics Startup Lexiang Intelligent
Robotics·36Kr (36氪)

Ant Group Takes Stake in AI Robotics Startup Lexiang Intelligent

An Ant Group investment arm has joined the shareholders of Lexiang Intelligent, a young Chinese startup building intelligent robots and AI applications.

Jul 26, 2026·2 min
Top themes
01Yuewen Unveils Buddy AI Agents for Writers, Comic Dramas and Copyright WorkJul 26, 2026·2 min02FutureTech at WAIC 2026: How the "One Person + a Swarm of AI Agents" Model Is Changing StartupsJul 26, 2026·3 min03Shanghai Extends STAR Market's Fifth Listing Standard: Loss-Making AI Companies Can Go PublicJul 26, 2026·3 min04Tabula: A Privacy-Preserving Foundation Model for Single-Cell Genomics and Cell Rejuvenation Factor DiscoveryJul 26, 2026·2 min05Ant Digital Unveils Agentar 2.0 at WAIC 2026, a "Super-Factory" for Business AgentsJul 26, 2026·2 min
Hybrid Training for VLA Models: Online-RL Quality at Half the Compute
Research·arXiv cs.LG

Hybrid Training for VLA Models: Online-RL Quality at Half the Compute

Jul 26, 2026·2 min
Memory Merge DQN: Merging Target Network Weights Instead of Hard Replacement in DQN
Creative·arXiv cs.LG

Memory Merge DQN: Merging Target Network Weights Instead of Hard Replacement in DQN

Jul 26, 2026·2 min
AI Decodes the Meaning of Speech From Brain Activity and Transfers to New People Without Retraining
Medicine·arXiv cs.LG

AI Decodes the Meaning of Speech From Brain Activity and Transfers to New People Without Retraining

Jul 25, 2026·2 min
Prefix-GRPO: Training Small Language Model Agents on Trajectory Prefixes
LLM·arXiv cs.LG

Prefix-GRPO: Training Small Language Model Agents on Trajectory Prefixes

Jul 25, 2026·2 min
OOD Benchmark Leak: How a Contaminated Class Dragged a Detector's AUROC Down to 0.326
Medicine·arXiv cs.LG

OOD Benchmark Leak: How a Contaminated Class Dragged a Detector's AUROC Down to 0.326

A hidden leak in an OOD benchmark gave an anomaly detector an AUROC of 0.326, below chance; once the class was removed, the metric rose to 0.911.

Jul 25, 2026·3 min
Graph Neural Networks vs. Gradient Boosting: Predicting Arsenic in US Water
Medicine·arXiv cs.LG

Graph Neural Networks vs. Gradient Boosting: Predicting Arsenic in US Water

Researchers compiled 74,000+ samples and compared ML models: graph neural networks account for spatial dependence and match gradient boosting at predicting arsenic…

Jul 25, 2026·2 min
mLSTM and Orthogonalized Reads: The Recall Gain Is Trainability, Not Memory
Research·arXiv cs.LG

mLSTM and Orthogonalized Reads: The Recall Gain Is Trainability, Not Memory

An arXiv preprint finds that orthogonalizing the mLSTM memory matrix (five Newton–Schulz iterations) improves trainability, not data storage, and the effect can be fully removed.

Jul 25, 2026·2 min
GLEAM: Spectral Losses for Turbulence Forecasting on Unstructured Meshes
Research·arXiv cs.LG

GLEAM: Spectral Losses for Turbulence Forecasting on Unstructured Meshes

Jul 25, 2026·2 min
LAARA vs. LoRA: Layer-Wise Adaptive Rank for Efficient Fine-Tuning
Creative·arXiv cs.LG

LAARA vs. LoRA: Layer-Wise Adaptive Rank for Efficient Fine-Tuning

Jul 25, 2026·2 min
SAE Interpretability Evaluation: The Measurement Pipeline Matters More Than the Model Architecture
LLM·arXiv cs.LG

SAE Interpretability Evaluation: The Measurement Pipeline Matters More Than the Model Architecture

Jul 25, 2026·2 min
FNO and UNO Neural Operators for Calculating Two-Dimensional Neutron Flux
Research·arXiv cs.LG

FNO and UNO Neural Operators for Calculating Two-Dimensional Neutron Flux

Jul 25, 2026·2 min
SUM: A Server-Side Fix for Forgetting in Federated Continual Learning, Up to 22% Gain
Research·arXiv cs.LG

SUM: A Server-Side Fix for Forgetting in Federated Continual Learning, Up to 22% Gain

Jul 25, 2026·2 min
STN-TGAT: A Graph Neural Network for Ranking S&P 500 Stocks and Building Portfolios
Research

STN-TGAT: A Graph Neural Network for Ranking S&P 500 Stocks and Building Portfolios

Jul 25, 2026·2 min
How to Explain Continual Learning in Neural Networks: Attention Rollout and Grad-CAM on Time Series
Research

How to Explain Continual Learning in Neural Networks: Attention Rollout and Grad-CAM on Time Series

Jul 25, 2026·2 min
Air Quality Arena: A Benchmark for Forecasting Air Pollution with Foundation Models
Medicine

Air Quality Arena: A Benchmark for Forecasting Air Pollution with Foundation Models

Jul 25, 2026·2 min
CruiseBench: A New Benchmark for Evaluating Neural Networks on Aircraft Engine Lifetime Prediction
Medicine·arXiv cs.LG

CruiseBench: A New Benchmark for Evaluating Neural Networks on Aircraft Engine Lifetime Prediction

Researchers have introduced CruiseBench, an N-CMAPSS-based benchmark that isolates the cruise phase of flight to fairly compare neural networks predicting aircraft…

Jul 25, 2026·2 min
HyenaND: A Subquadratic Operator That Replaces Attention on Multidimensional Data
Medicine·arXiv cs.LG

HyenaND: A Subquadratic Operator That Replaces Attention on Multidimensional Data

HyenaND processes images, volumes and PDEs in their native geometry and promises O(L log L) complexity instead of attention's quadratic cost, without losing…

Jul 25, 2026·2 min