AllLLMRoboticsStartupsResearchEthicsBusinessHardwareAutonomousMedicineCreative

All news

10667
Microsoft Decision-1: the model that picks answers from options 35x faster than GPT-6 Sol
Business·TNW

Microsoft Decision-1: the model that picks answers from options 35x faster than GPT-6 Sol

Oct 12, 2026·2 min
TikTok CEO Shou Chew Expects an AI Scientific Breakthrough Within Two to Three Years
Business

TikTok CEO Shou Chew Expects an AI Scientific Breakthrough Within Two to Three Years

Oct 12, 2026·2 min
Nicolas Cage Refused to Sign Amazon's AI Waiver for Spider-Noir
Ethics

Nicolas Cage Refused to Sign Amazon's AI Waiver for Spider-Noir

Oct 12, 2026·2 min
IMF: AI Infrastructure Spending Is Holding Global Growth at 3%
Business

IMF: AI Infrastructure Spending Is Holding Global Growth at 3%

Oct 12, 2026·2 min
Six parents explain how they limit their kids' social media and AI use: trust in Big Tech is fading
Business·Guardian

Six parents explain how they limit their kids' social media and AI use: trust in Big Tech is fading

Six parents told the Guardian they're tightening limits on kids' social media and AI chatbots because they don't trust company assurances, as regulators accuse Meta…

Oct 12, 2026·2 min
Microsoft President: AI Is Regulated Less Than Milk and Washing Machines
Business·Guardian

Microsoft President: AI Is Regulated Less Than Milk and Washing Machines

Brad Smith called for AI "kill switches" and a legal duty to victims of breaches carried out by rogue AI agents.

Oct 12, 2026·2 min
DeepSeek to Invest $400M in Robotics Startup Founded by Ex-Tencent Scientist
Robotics·Bloomberg Tech

DeepSeek to Invest $400M in Robotics Startup Founded by Ex-Tencent Scientist

DeepSeek wants to lead the first round of a new AI-for-robots startup. The money will come from the company's own funding of more than $12 billion.

Oct 12, 2026·2 min
Booz Allen Hamilton's AI Agents Succeeded in 8 of 8 Test Attacks on an Industrial System
Robotics·3DNews AI

Booz Allen Hamilton's AI Agents Succeeded in 8 of 8 Test Attacks on an Industrial System

In Booz Allen Hamilton's tests, AI agents found vulnerabilities on their own, gained admin access and controlled a robotic arm. Result: 8 of 8 attacks succeeded.

Oct 12, 2026·2 min
Top themes
01Karpathy Estimates Only About 5,000 People Have Access to the Most Advanced AIOct 12, 2026·2 min02SIA: China Supplies the US Just 5% of Chip Equipment, but Over 90% of Some Rare EarthsOct 12, 2026·2 min03Satya Nadella Urges Companies to Treat AI Models as an Insider Risk and Install an "Emergency Brake"Oct 12, 2026·2 min04Yandex Blocks Some 360 and Disk Features After Drone Attack on Third Data CenterOct 12, 2026·2 min05Asbis bets on data centers and AI to sustain a 476% stock rallyOct 12, 2026·2 min
TikTok CEO Promises AI Science Breakthroughs Within 2–3 Years
Business·3DNews AI

TikTok CEO Promises AI Science Breakthroughs Within 2–3 Years

Oct 12, 2026·2 min
There Are More Programmers Now, Even Though AI Promised to Replace Them Within a Year
Business·Habr AI

There Are More Programmers Now, Even Though AI Promised to Replace Them Within a Year

Oct 12, 2026·2 min
AI Solved 370 Math Problems, Mathematicians Panic, Musk Seeks $40 Billion: This Week's Highlights
Business·Habr AI

AI Solved 370 Math Problems, Mathematicians Panic, Musk Seeks $40 Billion: This Week's Highlights

Oct 12, 2026·4 min
Vibe-coding in Rust: three macOS utilities and five browser projects
LLM·Habr AI

Vibe-coding in Rust: three macOS utilities and five browser projects

Oct 12, 2026·2 min
Qwen3.6-35B-A3B hits 109 tok/s on an RTX 5080 with an RTX 3060 on PCIe Gen2 x4
Hardware·Habr AI

Qwen3.6-35B-A3B hits 109 tok/s on an RTX 5080 with an RTX 3060 on PCIe Gen2 x4

An enthusiast ran a 35B model on an RTX 5080 and an RTX 3060 in a PCIe Gen2 x4 slot (about 2 GB/s) and got 109 tokens/s. Here's why the narrow link wasn't a…

Oct 12, 2026·2 min
How to Save AI Agent Tokens When Working with Large Codebases: 5 Patterns
LLM·Habr AI

How to Save AI Agent Tokens When Working with Large Codebases: 5 Patterns

When analyzing large projects, AI agents spend most of their tokens exploring code rather than writing it. Habr breaks down 5 universal context management patterns…

Oct 12, 2026·3 min
Tsinghua Team Pushes Generation Down to a Single Step, From Images to Streaming Video
Research·Jiqizhixin (机器之心)

Tsinghua Team Pushes Generation Down to a Single Step, From Images to Streaming Video

Tsinghua researchers are developing "sample distribution training" to enable one- or few-step generation, from still images to real-time video.

Oct 12, 2026·2 min
模思智能 Raises Over 1 Billion Yuan in Pre-A for a Real-Time Omni-Modal Model
Business·Jiqizhixin (机器之心)

模思智能 Raises Over 1 Billion Yuan in Pre-A for a Real-Time Omni-Modal Model

Oct 12, 2026·2 min
New Benchmark Tests Whether AI-Generated Images Obey the Laws of Physics
Research·Jiqizhixin (机器之心)

New Benchmark Tests Whether AI-Generated Images Obey the Laws of Physics

Oct 12, 2026·2 min
UNITAS: A 3D World Model for Robots Claims the Lead on Several Benchmarks
Robotics·Jiqizhixin (机器之心)

UNITAS: A 3D World Model for Robots Claims the Lead on Several Benchmarks

Oct 12, 2026·2 min
HIL-UMI: Dong Hao's Team Finds a Robot's Weak Spots Without the Robot
Robotics·Jiqizhixin (机器之心)

HIL-UMI: Dong Hao's Team Finds a Robot's Weak Spots Without the Robot

Oct 12, 2026·2 min
Mamba-based S²FI grades fruit quality from video, backed by a dataset of 6,460 clips
Robotics·arXiv cs.CV

Mamba-based S²FI grades fruit quality from video, backed by a dataset of 6,460 clips

Oct 12, 2026·2 min
SAMF Teaches Drones to Locate Themselves on a Satellite Map Using Thermal Imaging, No GNSS Required
Robotics

SAMF Teaches Drones to Locate Themselves on a Satellite Map Using Thermal Imaging, No GNSS Required

Oct 12, 2026·2 min
APEX Compresses Diffusion Latents 128x and Beats a 256-Token Model on ImageNet-512
Research

APEX Compresses Diffusion Latents 128x and Beats a 256-Token Model on ImageNet-512

Oct 12, 2026·2 min
SWIM saves compute: audio-video generation gains up to 30.7% over Best-of-N
Research

SWIM saves compute: audio-video generation gains up to 30.7% over Best-of-N

Oct 12, 2026·2 min
SAM 2.1 fine-tuned on facades and interiors: mIoU rises from 49.9% to 69.6%
LLM·arXiv cs.CV

SAM 2.1 fine-tuned on facades and interiors: mIoU rises from 49.9% to 69.6%

Researchers fine-tuned SAM 2.1 on 2679 images of 42 buildings, lifting mIoU across 11 building-element classes from 49.9% to 69.6%. The model was integrated into…

Oct 12, 2026·2 min
SAIL Keeps Its Target in Frame Using Audio and Text, and Beats Rivals on Two Benchmarks
Research·arXiv cs.CV

SAIL Keeps Its Target in Frame Using Audio and Text, and Beats Rivals on Two Benchmarks

A new preprint introduces SAIL, a framework for segmenting objects from video, audio and text. Two tokens counter target loss; on two benchmarks it beats AVS…

Oct 12, 2026·2 min