Jiqizhixin (机器之心)
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
SGLang and Diffusion Texts: How Chinese Engineers Accelerate Context to Infinity
The large language model industry has entered a phase where simply increasing the number of graphics cards in a cluster is no longer sufficient.
Recommendation Systems: Why Algorithms Soon Will Stop Selecting and Start Creating
Remember when you spent half an hour scrolling Netflix and ended up going to bed without picking anything? The problem isn't that there aren't enough movies.
Sora 2: Chinese Project Breaks Monopoly of Closed Neural Networks
We've been living for nearly a year now in a reality where OpenAI showed us the future of video generation but hasn't let us touch it.
H Company: Why Investors Gave 180 Million Dollars to a Startup Without a Product
Imagine this situation: you come to investors, you have no ready product, no massive cluster of thousands of GPUs, and not even a website with fancy charts.
SkyReels-V3: Chinese Neural Network Makes Elon Musk Work for It
While Western giants like OpenAI keep their best developments under lock and key, the Chinese team from Kunlun Wanwei decided to go all-in.
Alibaba's Golden Triangle: How Its Own Chips Save the Cloud Empire
While the Western world held its breath awaiting the next shipment from NVIDIA, a completely different game was unfolding in the East.
SkyMusic: Why Chinese Kunlun Tech Spooked Suno and Udio
Let's be honest: until recently, most AI-generated tracks sounded like a robot trying to mimic singing underwater.
ByteDance and Li Han: How to Make Neural Networks Work, Not Just Talk
We've spent too long treating large language models as advanced search engines or amusing conversation partners.
Tencent Cloud Lighthouse: Deploy Clawdbot in One Click
Remember those days when setting up a server felt like an exorcism session? If you wanted to run even the simplest bot, you'd spend half a day wrestling with SSH keys, Nginx configs, and endless…
CRISPR: AI Teaches Genetic Scissors to Stop on Time
Imagine you're trying to edit complex software code, but your text editor occasionally randomly deletes entire chunks of data in neighboring files.
Vidu Q2: Now AI Copies Not Just Pictures but Acting Performance
Vidu Q2: Now AI Copies Not Just the Image, But the Acting Performance While the global community waits in anticipation for at least a closed beta of OpenAI's Sora, Chinese developers from Shengshu…
TTT-Discover: Stanford and NVIDIA Force AI to Think on the Fly and Outpace Scientists
Scientific progress has always hit the same wall: hypothesis verification. You can be a brilliant physicist, but simulating a complex process in a fluid or calculating the structure of a new material…
SplatSSC: Gaussian Splats Learned to Complete Reality from a Single Photo
Robots and autonomous vehicles have long suffered from "flat" vision. To understand what lies around a corner or how deep a corridor extends, they either had to be loaded with expensive lidars or…
Ant Group Teaches Robots Life: New VLA Model Outpaces Pi0.5
While Western venture capitalists are showering money on startups like Physical Intelligence, a player has appeared on the horizon that they clearly didn't expect.
OpenAI Prism: Scientific Articles Now Written Without Pain (and Without Overleaf)
The academic world has long remained a conservative bastion, dominated by complex typesetting systems and unwieldy citation managers.
The Distillation of Mind: Why Neural Networks Are Harmed by Teachers That Are Too Smart
Distillation of Mind: Why Neural Networks Suffer from Overly Clever Teachers Imagine you're trying to explain quantum electrodynamics to a first-grader.
Anthropic vs Clawdbot: Why Claude No Longer Tolerates Namesakes
In the world of modern artificial intelligence, naming is not just creativity but a minefield. If you decide to launch a service and its name echoes one of the "big three" LLMs, be prepared for…
LingBot-VLA: Ant Group Teaches Robots to Understand the World Without Extra Words
Imagine you bought a new phone and didn't need to relearn how to use it — your fingers already know where to press.
Kimi K2.5: Moonshot AI Goes Open-Source and Breaks Competitors' Plans
Remember how a year ago everyone was crazy about Kimi because of its phenomenal ability to "digest" huge volumes of text?
ICLR 2026: Why Getting Into AI Elite Became Harder Than Ever
Imagine this situation: you've spent half a year of your life, burned a small country's budget on GPU rentals, and consumed an ocean of coffee, only to end up with a terse rejection notice.
NVIDIA Earth-2: Jensen Huang Decided to Replace All Meteorologists at Once
While we habitually grumble at the app on our phones that promised sunshine but delivered rain, NVIDIA is quietly building a digital twin of the planet.
AI Agents: Three Pillars That Separate a Chatbot from a Digital Employee
If you thought GPT-4 or Claude 3 were the pinnacle of achievement, I have news for you. A large language model by itself is simply a very well-read, but extremely absent-minded conversationalist.
Glass Ceiling Shattered: Ant Group Taught Robots to See the Invisible
Imagine you bought a robotic assistant for a couple of tens of thousands of dollars, and on the very first day it crashed into a panoramic window or tried to grab thin air instead of a glass of…
Qwen3-Max-Thinking: Alibaba Releases Its Own "Thinker" Against OpenAI
The world of AI has definitively stopped being a race for the number of parameters and turned into a competition of "who thinks better." While Western giants carefully meter access to their smartest…
Qwen 2.5-Max: Alibaba Caught Up to Google While You Slept
The AI world has grown accustomed to the OpenAI and Google duopoly, but Chinese engineers from Alibaba Cloud decided it was time to change the rules of the game.
Sora for Marketplaces: Chinese Developers Created the Perfect Sales Video Pipeline
While Western giants are cautious about realism and copyright, the Chinese tech sector decided to bring advanced technologies down to earth into commercial reality.
LLM Black Box: Why We Still Don't Understand How They Think
We are accustomed to thinking that engineers are people who know exactly how their mechanism works down to the last bolt. In the case of large language models (LLMs), this confidence crumbles to dust.
Maia 200: Microsoft Builds Its Silicon Castle (Without Nvidia)
While the entire technology world held its breath waiting for the next shipment of scarce Nvidia accelerators, Microsoft decided it was time to finally take its fate into its own hands.
Diploma Be Gone: How the 'Wild Path' Leads to OpenAI and Anthropic
Forget everything you've been told about needing a PhD to work in the top tier of artificial intelligence.
OpenAI Under Siege: Why Sam Altman's Leadership Is No Longer a Given
Remember those glorious days when every update from OpenAI forced competitors into panic mode, frantically rewriting their roadmaps? It seems that period of dominance has come to an end.