Remarkable 8
Nvidia (NVDA) CEO Jensen Huang unveiled Alpamayo, a family of open models built for autonomous driving, during his keynote address at CES 2026 in Las Vegas on Monday. “Today, we are introducing Alpamayo, the world’s first thinking model for autonomous driving,” Huang said. “Not only does your car drive as you would expect it to drive, but it reasons out any situation it could come upon.” Mercedes-Benz’s (MBGAF)(MBGYY) new CLA model, the brand’s first vehicle featuring the MB.OS platform, will include Nvidia’s full-stack DRIVE AV software, AI infrastructure and accelerated compute. The two companies have worked on this for several years. Huang said the vehicles will be available in the U.S. during the first quarter of 2026, in Europe during the second quarter and will enter other markets and regions during the second half of 2026. “This is the first large-scale physical AI market,” Huang said. “We can all agree it is fully here. In the next 10 years I’m fairly certain a large percentage of the world’s cars will be autonomous.”
An increasing number of Chinese companies were now developing world-leading “open-weight” AI models, not just DeepSeek, the Center for AI Standards and Innovation (CAISI) at the Department of Commerce said on Friday. The CAISI report was the second US government evaluation of a leading Chinese model developer, after US President Donald Trump called for official assessments of frontier Chinese models’ capabilities and censorship in his AI Action Plan, released in July.
A new generation of AI tools is reshaping offensive cybersecurity, making vulnerability discovery faster, cheaper, and far more automated. NeuroSploitv2 exemplifies that shift. The AI-powered penetration testing framework works across multiple large language models, including Gemini, Claude, GPT, and Ollama, automating tasks that once required skilled human researchers. In practice, it’s beginning to replace entire phases of the penetration testing process. That transition is already real. Just six weeks ago, Cyera Research Labs used AI-augmented methods to uncover two critical vulnerabilities in Google’s Gemini CLI — flaws that enabled arbitrary command execution and forced rapid fixes from Google. The takeaway is simple: AI-driven security research is now operating at a pace traditional workflows struggle to match.
This AI Tool Can Plan and Execute Penetration Tests on Its Own | eWEEK
Nvidia Cosmos Reason 2 is an open-reasoning vision language model (VLM) that allows intelligent machines “to see, understand, and act in the physical world like humans,” according to Nvidia. Moreover, using Nvidia Cosmos Reason 2, physical AI can make decisions as humans do, using reason, prior knowledge, understanding of physics, and more. Lastly, Nvidia Isaac GR00T N1.6 is an open-reasoning vision language action (VLA) model specifically designed for humanoid robots, enabling full-body control and leveraging Nvidia Cosmos Reason for the additional benefits discussed above. The new models are all available on Hugging Face.
Nvidia’s physical AI models clear the way for next-gen robots - here’s what’s new | ZDNET
Just before the start of the new year, the AI world was introduced to a potential game-changing new method for training advanced models. A team of researchers from Chinese AI firm DeepSeek released a paper on Wednesday outlining what it called Manifold-Constrained Hyper-Connections, or mHC for short, which may provide a pathway for engineers to build and scale large language models without the huge computational costs that are typically required.
How DeepSeek’s new way to train advanced AI models could disrupt everything
“The next wave of AI won’t be defined by sheer scale, but by ubiquity -- by models small enough to run on a drone, in a car, in robots, on a phone or a computer laptop,” Mistral said in the release, pointing out that small models are often preferable for real-life use cases. By keeping costs and latency down, they can be more accessible than heavier, slower, more expensive models that require more infrastructure to run on. Mistral added that small models like Ministral 3 are also easier to customize, making them ideal for fine-tuning to enterprise workflows. The company emphasized that customization as the release’s main appeal for developers across all kinds of projects.
Mistral’s latest open-source release bets on smaller models over large ones - here’s why | ZDNET
“We anticipate France’s debt-to-GDP ratio, already at 113 per cent in 2024, will deteriorate by an additional 2 to 3 percentage points per year in the medium term,” said UBS. It has urged clients to cut their holdings in French short-dated bonds. They won’t be alone in pulling out and a sell-off will only intensify the crisis by increasing French borrowing costs. Expect downgrades from the ever late to the party credit ratings agencies to follow.
France’s pension crisis could soon spell trouble for Britain
“I’m writing this to tell you something simple. Growing older is not the closing act. It can be the most exquisite chapter – if you let yourself bloom again. Let these years ahead be your treasure years. You don’t need to be famous. You don’t need to be flawless. You only need to show up – fully – for the life that is still yours.”
Patricia Routledge death: Keeping Up Appearances star dies, aged 96
While some believe that machines soon will do everything humans can do, ushering in a new age of boundless prosperity, other predictions are at least more grounded. For example, Goldman Sachs predicts that generative AI will boost global GDP by 7% over the next decade, and the McKinsey Global Institute anticipates that the annual GDP growth rate could increase by 3-4 percentage points between now and 2040. For its part, The Economist expects that AI will create a blue-collar bonanza.
We told you so (here and here):
Boston Dynamics’ humanoid robot, Atlas, has taken a significant step beyond flashy lab demos. For the first time, the AI-powered robot is being tested in a real factory, learning how to do actual industrial work. CBS’s 60 Minutes was invited to Hyundai Motor Group’s new auto plant near Savannah, Georgia, where Atlas was put to work inside a parts warehouse. The visit offered a rare look at how close humanoid robots are to joining human workers on factory floors. Standing 5 feet, 9 inches tall and weighing about 200 pounds, Atlas is designed to match human scale while exceeding human endurance.
Unlike traditional factory robots bolted to one spot, Atlas can walk, lift, turn, and adapt to different tasks — qualities that make humanoid robots especially appealing for flexible factory work. How Atlas learns to work Earlier versions of Atlas relied on hand-written code. The current model is different. It is fully electric and powered by AI running on Nvidia processors. Rather than programming every move, Boston Dynamics now trains Atlas by teaching it. Engineers guide the robot through tasks using virtual reality controls or motion-capture suits. Those demonstrations generate data that feeds Atlas’ machine-learning models. “If that teleoperator can perform the task that we want the robot to do, and do it multiple times, that generates data that we can use to train the robot’s AI models to then later do that task autonomously,” said Scott Kuindersma, head of robotics research at Boston Dynamics, in an interview with 60 Minutes from CBS. Much of this learning happens in simulation. Thousands of digital versions of Atlas are trained at once under different conditions. Once a skill works well, it is uploaded to the real robot — and to every other Atlas unit.
Boston Dynamics’ Atlas Humanoid Robot Enters the Factory Floor
Not everybody on the right is applauding the US move in Venezuela, here is Evans-Pritchard, chief economic commentator of The Telegraph:
Beware of viewing energy analytics through the prism of the American media and the culture war. The critical fact is that sales of fossil-based “ICE” cars peaked eight years ago in China and are now in freefall – and China today is a bigger car market than the US and the EU combined.
The share of EVs and plug-in hybrids reached 53pc in November. This pattern is fast being replicated even in long-haul trucks. China is set on a course to eliminate all dependency on seaborne oil and gas supplies at a breakneck pace and for strategic reasons.
In the process, it has driven down the cost of EVs to levels that wipe out ICE rivals in any open market – almost anywhere in the world. Sales reached 51pc in Singapore in November and well over 40pc in Vietnam. Thailand is fast becoming an EV production hub for South East Asia, with output up 20-fold in a year.
Europe is tapping the brakes behind a wall of tariff protection, but sales of pure ICE cars have continued to slide nevertheless, falling from 46pc to 36pc in the 11 months to November compared to a year earlier. This is what Trump and his Maga pirates are up against in Venezuela, and more broadly in their anachronistic, pre-modern view of how the world works.
To the extent that the Orinoco ever produces serious oil it will add barrels to a structurally saturated market and will in the end cannibalise US shale on home ground.
Trump has pulled the US into a joint venture with a Leftist kleptocracy
Robot vacuums have learned how to mop, dodge cables, and even lift small obstacles. At CES 2026 in Las Vegas, Roborock added a new milestone to that list: stairs. The company, headquartered in Beijing, China, used its CES stage to unveil the Saros Rover, a robot vacuum that uses a wheel-leg design to climb staircases, handle slopes, and clean areas that have long been off-limits to robotic cleaners. Roborock describes the Saros Rover as the world’s first robot vacuum with a two-wheel-leg architecture. Each wheel-leg can move independently, allowing the robot to raise itself, lower itself, and adjust its balance as the ground changes.
China’s Roborock Unveils a Robot Vacuum That Can Climb Stairs
Self-driving car firm Mobileye has made a massive acquisition that signals where AI is heading next. The company announced it will acquire humanoid robotics startup Mentee Robotics for $900 million, marking one of the largest humanoid robot deals in history. Mobileye will pay $612 million in cash plus 26.2 million shares, representing a 455% premium over Mentee’s $162 million valuation from 10 months ago. The deal is expected to close in the first quarter of 2026.
Mobileye to Buy Israeli Startup Mentee Robotics for $900M | eWEEK
Well, well, healthcare has experienced a shift that might make doctors worried about job security go very pale. OpenAI has introduced ChatGPT Health, marking the first time AI has been specifically designed to serve as a dedicated personal medical assistant for everyday users. With over 230 million people globally asking health-related questions on ChatGPT every week, make no bones about it, the demand was already there. The features OpenAI’s announcement says the AI can explain lab results, create exercise plans, and recommend questions to ask during medical appointments. Unlike generic medical websites that offer one-size-fits-all information, this system learns from user interactions to deliver increasingly tailored health guidance through high-level instructions provided ahead of time. Advanced natural language processing that understands complex medical descriptions in plain English. Picture this: you can describe symptoms conversationally, and the AI translates these descriptions into medically relevant insights while maintaining appropriate caution about serious conditions requiring professional intervention. The system integrates with existing health records through B.well and connects with wellness apps like Apple Health, Function, and MyFitnessPal. This creates a health monitoring ecosystem that operates 24/7, with access to healthcare metrics collected by Apple Watch including heart rate, sleep patterns, and running form.
OpenAI Offers Medical Assistance With ChatGPT Health Launch | eWEEK