All news

Robotics

UBTech's Thinker-WM Takes Three First Places, Scoring 0.740 on WorldOlympiad

Research

Google DeepMind Launches AI Agent to Hunt for Antimicrobial Peptides Among Millions of Proteins

LLM

Synchronous MoE decoding makes GPUs wait for the busiest group

Business·Jiqizhixin (机器之心)

A Company from Nanjing University's Orbit Bets on a Fourth Driver of AI Growth

A team from Nanjing University's orbit sees a fourth driver of AI growth after compute, algorithms and data: connectivity and cooperation among independent models.

MTEdit: 500 thousand editing trajectories teach models not to lose quality over 8 steps

Researchers built MTEdit, a dataset of 500 thousand multi-step image edits, and fine-tuned Qwen-Image-Edit to the best result among open models.

Xiaomi-Lens-0, a 4B model, beats Qwen3.5 by 5.6 points on RefCOCO

Xiaomi releases Lens-0: a 4B model built on Qwen3.5 that draws masks from its own visual latents, gains 5.6 points on RefCOCO and leads on cIoU.

OfficeTown Simulates 23 Organizations and Trains Agents for Office Work on 406 Tasks

Researchers built OfficeTown, a generative simulation of offices that produced a benchmark of 406 tasks and 4,836 criteria. Training on its data noticeably improved…

01Reflection releases Beam: 501B parameters and 30 points on the Intelligence IndexOct 16, 2026·3 min 02MARS Transfers Text-Based Refusal to Video and Images Without Data: A Test on 5 MLLMsOct 16, 2026·3 min 03Conditional Program shrinks DeepSeek's n-gram memory from 203 to 51 GB at 2 bitsOct 16, 2026·2 min 04ERB: AI Agents Build Their Own Model Tests and Match the Reference Benchmarks in 75% of PairsOct 16, 2026·2 min 05MambaAct speeds up π0.5 robots: action expert 10x smaller, latency down 64.8%Oct 16, 2026·3 min

Top themes

LM2LM: Qwen3.8-27B Beats GPT-6-sol at Ideas for Improving LLMs and Is 33% Cheaper

Robotics·Hugging Face Daily Papers

Codex Teaches Team Sites to Update on a Schedule and Adapt to Each Visitor

OpenAI showed how to build Sites in Codex: scheduled tasks refresh the data, and users' own plugins give each visitor a personalized view of the page.

Business·YouTube: OpenAI

OpenAI Releases GPT-Live-1.1: New Voices in 6 Languages and Fewer "checking..." Pauses

GPT-Live-1.1 is now available to all users, with new voices for French, German, Hindi, Mexican Spanish, Korean and Japanese, plus more accurate accents.

Research·Jiqizhixin (机器之心)

Tadpole Proposes a Foundation Model for Physics: From Prediction to Understanding

Tadpole brings foundation-model logic to science: one model instead of separate networks for weather, airflow, material deformation and electromagnetism.

Business·YouTube: Stanford Online

Stanford Online: AI Can Estimate Impact and Cost, but a Human Sets the Priority

Robotics·Jiqizhixin (机器之心)

EinsiaAI Releases EmbodiedEvo: Self-Evolving Agent Tops Robodojo

Hardware·Jiqizhixin (机器之心)

Business·Lovable Blog

Lovable launches Deploy on AWS: apps and data move into the customer's VPC

Oct 15, 2026·3 min

xFusion brings 800V HVDC and supernodes for "token factories" to OCP 2026

Business·YouTube: OpenAI

SentinelOne: OpenAI Models Do 4–6 Weeks of Reverse Engineering in 8 Hours

Business

ByteDance Seed Creates a Standalone Safety Team and Seeks 10 Top Specialists

Robotics

Artificial Analysis Launches AA-Robotics, Tests GPT, Claude and Gemini as a Robot's Brain

Research

SpecTr-GBV at EMNLP 2026 Combines Multi-Draft and Block Verification to Speed Up LLMs

Business·Jiqizhixin (机器之心)

AIPOCH Open-Sources Open-Science, a Lab Notebook for "Vibe Research" with AI

AIPOCH has released Open-Science as open source. The project promises to make AI-assisted research traceable, transferable and reproducible. Implementation details…

LLM·Jiqizhixin (机器之心)

Honeycomb Gives Video Models Constant-Size Memory for Loop Closure

Researchers unveil Honeycomb, a fixed-size memory that helps a video model recognize a room after the camera completes a full circle, without context growing with…