AntiFox-linked intrusions used IIS web shells and valid AWS credentials to enumerate cloud services and access production secrets.
Category: robotics/AI – Page 12
The Rise and Rise of Alice Huang
Alice Huang leveraged her connections from investing in internet video games to become a data center executive. Here are the interlocking arrangements that would put her in Jensen Huang’s orbit.
The revenues of alleged Chinese spy Alice Huang’s online video games company, BVI-registered Osmanthus Vale Holdings Ltd, rose from $36,000 to $36 million in just three years. Then she sold it for $145 million to a Shenzhen firm that would facilitate her entry into the AI data center industry. The Wire China follows the filings.
PRISM helps robots move better by learning how forces and motion interact
A new neural network component, developed at University of Michigan Engineering, challenges a common assumption in robot learning: that the same general-purpose neural networks used for image recognition are equally suited to controlling physical movement.
In image classification, a neural network maps an image to a label but does not change the image. During robotic motor control, actions—like a humanoid robot moving a limb—change the robot’s physical state, which in turn changes what sensors observe next.
Robot movement and goals involve multiplicative relationships among physical variables. For example, the forces acting on one joint depend on the positions and velocities of other joints, while quantities such as mechanical power and contact multiply forces and motion.
Google Search
Google DeepMind has released Gemini 4 Carbon, built specifically for long-horizon agentic workflows, complex debugging, and scientific reasoning. While Gemini 4 Argon remains the standard for everyday software engineering tasks, Carbon handles tasks that push frontier models to their limits.
To showcase its cross-disciplinary reasoning and multimodal code synthesis, an intriguing open-source companion project was published alongside the release: Birdsong github/dopeytree/birdsong(https://www.google.com/search?q=https%3A%2F%2Fgithub.com%2Fdopeytree%2Fbirdsong)).
The model generates interactive 3D acoustic simulations of a songbird’s syrinx (vocal organ), modeling live sound synthesis across six bird species—serving as a compelling practical benchmark for physics-informed agentic generation.
BENCHMARK COMPARISON: GEMINI 4 CARBON vs. ARGON
• General Intelligence (Artificial Analysis Index v5.0) — Carbon: 59 | Argon: 50 (+9 pts jump)
• Advanced Math (RiemannBench) — Carbon: 80.0% | Argon: 76.0% (+4.0 pp)
• Scientific Tool Use (Terminal-Bench-Science 0.1) — Carbon: 64.8% | Argon: 52.9% (+11.9 pp boost in tool execution)
• Long-Context Reasoning (GraphWalks 256k–1M, BFS) — Carbon: 96.2% | Argon: 84.2% (+12.0 pp in long-context mapping)
Ada Palmer on Viking Ethics, Laws of History, and Terra Ignota
In June 2020, I asked a historian whether the United States was heading toward civil war.
Six years later, her answer is worth hearing again.
Ada Palmer studies censorship, heresy, and how power shifts during information revolutions. Her research asks what the printing press can teach us about the choices we face in the digital one. She is also the author of Terra Ignota, the award-winning #ScienceFiction series set in a 25th-century world of borderless nations and voluntary citizenship.
We talked for 3 hours and 7 minutes. We talked about Peter Turchin’s laws of #History and whether the past can predict anything. About what popular science like Cosmos gets wrong about the Inquisition. About what she learned reenacting the papal election of 1,492 with her students. About Viking ethics, of all things. And about whether one person still matters when history seems to move on its own.
One line from her has stayed with me:
“Your choices aren’t: keep the world I have or make a better one. Your choices are: let the world change towards the entrenchment of power… or work to shape that change.”
Microsoft presses Trump to scuttle ban on Chinese data center tech
America’s AI boom depends on a critical component made largely in China. And now Microsoft is urging Trump NOT to ban it.
It’s called an optical transceiver. These devices convert electrical signals into optical signals, allowing data to travel through fiber-optic cables between servers, switches, and routers inside data centers.
Here’s the problem: China accounts for roughly two-thirds of the global market.
Washington is concerned that Chinese-made transceivers could potentially be exploited to steal data, install malware, or disrupt U.S. data center operations.
But banning them could create another problem: a massive supply-chain bottleneck.
Microsoft and other tech companies warn that U.S. manufacturers simply don’t have the capacity to replace Chinese suppliers at scale. Billions of dollars’ worth of AI chips have already been purchased, but without the optical transceivers needed to connect the infrastructure, those chips could end up sitting idle.
And there’s another complication: one of China’s leading suppliers, Zhongji Innolight, has been added to the Pentagon’s list of Chinese companies alleged to have military ties.
Quantum could solve AI energy drain, says tech founder
Quantum computers could curb the vast energy demands of artificial intelligence data centers with their ability to run complex calculations, one of the tech’s pioneers told AFP.
Australian startup Diraq is conducting the first field test worldwide putting a small quantum computer based on a silicon chip in a commercial data center in Sydney.
Gigawatt-scale energy demand from large data centers—warehouse-like buildings full of servers used to train or run AI tools—requires new power plants or increases pressure on electrical grids, sparking community backlash.
AI-generated molecular fingerprints help identify existing compounds with therapeutic potential
Health care researchers have long sought to harness the power of computing to design molecules that bind to specific receptors and can be used as targeted treatments for diseases. Even with the increasing sophistication of artificial intelligence, however, this technological promise has yet to be realized.
Junmei Wang, professor of pharmaceutical sciences in the School of Pharmacy, is tackling the problem from a different perspective.
In a paper published Oct. 28 in the journal Science Advances, Wang outlined Full-Spectrum Generative Lead Discovery (FSGLD), a novel, practical and reusable framework that takes advantage of generative AI to screen for existing compounds with potential for therapeutic development.
Railway Says Stolen Customer AI Keys Led to $740,000 in Estimated Unauthorized Usage Charges
Railway says stolen AI API keys caused $740,000 in unauthorized charges after attackers decrypted secrets in up to 100,000 workspaces.
Microsoft Researchers Show AI Agents Can Be Backdoored by Editing Fewer Than 10 Neurons
Microsoft demonstrates how editing as few as six AI model neurons can plant backdoors that trigger command execution by agents.







