The Rise of the Thinking Agent and the New Era of Global AI Competition

The global artificial intelligence landscape just witnessed a significant tectonic shift. Beijing-based Moonshot AI has made waves with the reported capabilities of its advanced models, including a “thinking agent” designed to perform complex, multi-step reasoning. This development, often associated with their powerful Kimi chatbot, is not just another iterative update; it represents a fundamental change in how AI operates, moving from simple question-answering to autonomous problem-solving. The emergence of such powerful models from China signals a new phase in the AI race, one where reasoning and agency are the new frontiers.

The “Thinking Agent” Revolution

At the heart of this advancement is the concept of a “thinking agent.” Unlike traditional large language models (LLMs) that predict the next word in a sequence, a thinking agent is designed to formulate a plan, execute it using various tools, observe the results and refine its plan accordingly. Reports suggest that Moonshot’s advanced systems are capable of executing an astounding 200–300 sequential tool calls to achieve a single goal. This is a massive leap from current standards, where models typically struggle with more than a handful of sequential actions.

This capability transforms the user experience. Instead of asking, “What is the stock price of Company X?”, a user could task the agent with: “Analyze the last five years of financial reports for Company X, compare its performance with its top three competitors, taking into account recent market trends, and generate a comprehensive investment report.” The AI would then autonomously break this down into hundreds of smaller steps: searching for competitors, retrieving financial documents, performing calculations, reading market news and synthesizing the information into a final document. This shift from a passive information retrieval system to an active, problem-solving partner is the promise of the “thinking agent.”

A visualization of a thinking agent's complex decision-making process, involving hundreds of sequential steps and tool calls.
Above: A visualization of a thinking agent’s complex decision-making process, involving hundreds of sequential steps and tool calls.

Closing the Gap: China’s AI Ambitions

The significance of Moonshot’s achievement cannot be overstated. For years, the prevailing narrative was that the U.S. held an insurmountable lead in generative AI. However, the rapid rise of Chinese labs like Moonshot, which was founded in 2023, challenges this view. By developing models with such deep reasoning capabilities, Moonshot is demonstrating that Chinese AI is not just about scale but also about sophisticated architectural innovation.

The release of models with “thinking” capabilities, similar to OpenAI’s recent o1 model (codenamed “Strawberry”), shows that Chinese labs are keeping pace with, and in some aspects potentially surpassing, their Western counterparts in the race for artificial general intelligence (AGI). This is a clear indicator of a vibrant and fiercely competitive AI ecosystem in China that is quickly closing the technological gap. For more on the global AI race, see this analysis of the U.S./China AI competition.

Stacking Up Against U.S. Titans

The competition is no longer just about who has the largest model or the most GPUs. It’s about who can build the smartest, most autonomous agent. Moonshot’s focus on long-horizon reasoning and massive sequential tool use places it in direct competition with the most advanced projects from OpenAI, Google DeepMind and Anthropic.

While U.S. companies have excelled at building foundational models, Moonshot’s approach highlights a different strategic focus: building systems that can reason and act in the real world over extended periods. This capability is crucial for automating complex workflows in industries like finance, law and software development. The fact that a relatively new startup is pushing the boundaries in this specific area is a testament to the agility and ambition of China’s AI sector.

Real-World Applications: What Can it Do?

The ability to execute hundreds of sequential tool calls opens up a world of possibilities. In software engineering, an agent could be given a high-level feature request and autonomously write code spanning multiple files, debug its own errors and write test cases. In scientific research, it could formulate hypotheses, search through vast databases of papers, design experiments and analyze results.

For everyday users, this could mean a personal assistant that can plan a complex multi-city vacation, booking flights and hotels based on a set of preferences and budget constraints, all without human intervention. The move is from a conversational partner to a capable executive assistant.

The Broader Chinese AI Landscape

Moonshot AI is not an outlier but part of a larger, dynamic ecosystem. Other Chinese tech giants and startups are also making significant strides. Alibaba’s Qwen series of models, for instance, has consistently ranked at the top of open-source leaderboards, demonstrating world-class performance in coding and mathematics. DeepSeek and Zhipu AI are other key players pushing the frontiers of open-source and commercial AI models.

This collective effort, driven by intense domestic competition and strong government support, has created a formidable AI industry that is now a primary rival to the U.S. The landscape is no longer dominated by a single country but is a bipolar competition that is accelerating the pace of innovation globally.

A conceptual illustration of the balanced AI competition between major tech companies in the U.S. and China.
Above: A conceptual illustration of the balanced AI competition between major tech companies in the U.S. and China.

Wrapping Up

The emergence of Moonshot AI’s advanced “thinking” models is a watershed moment. It signals that the era of simple chatbots is ending, replaced by a new generation of AI agents capable of complex, multi-step reasoning and autonomous action. This development not only showcases the rapid advancements of China’s AI ecosystem but also intensifies the global competition for AI supremacy. As Chinese and U.S. labs push each other to new heights, we can expect to see an even faster rate of innovation, leading to AI systems with capabilities we are only just beginning to imagine. The future of AI is not being written in one country, but in a dynamic, high-stakes rivalry that is reshaping the technological world.

Sony settles with Tencent over ‘slavish’ Horizon clone

Earlier this year Sony sued Tencent for copyright infringement over its Light of Motiram game, calling it a “slavish clone” of Horizon Zero Dawn. Then, earlier this month, Tencent agreed to stop promoting and publicly testing the game. Now, the two companies have reached a “confidential settlement” and the case has been dismissed, according to court documents seen by The Verge. Light of Motiram has also disappeared from Steam and Epic’s game stores. 

“SIE and Tencent are pleased to have reached a confidential resolution and will have no further public comment on this matter,” Tencent’s spokesperson told The Verge. 

When Sony first filed its lawsuit in July 2025, it said that Tencent’s game appeared to copy aspects of not just Horizon Zero Dawn, but other franchise games including Horizon Forbidden West and Lego Horizon Adventures. That included the post-apocalyptic setting with humans and machines coexisting, the visual appearance of characters and even the marketing materials — something Engadget certainly noticed when Tencent first announced the game.

This article originally appeared on Engadget at https://www.engadget.com/gaming/sony-settles-with-tencent-over-slavish-horizon-clone-120042886.html?src=rss

The inside story of SpaceX’s historic rocket landing that changed launch forever

On Dec. 21, 2015, SpaceX launched the Orbcomm-2 mission on an upgraded version of its Falcon 9 rocket. That night, just days before Christmas, the company successfully landed the first stage for the first time. The story behind this remarkable achievement is nowhere more fully told than in the book Reentry, authored by Ars Technica Senior Space Editor Eric Berger and published in 2024. To mark the tenth anniversary, Ars is reprinting a slightly condensed chapter from the book that tells the inside story of this landing. The chapter begins in June 2015 with a tragedy, the disintegration of a Falcon 9 rocket carrying the CRS-7 cargo supply mission for NASA. It was the first time a Falcon 9 had been lost in flight.

Seconds after the Dragon-bearing Falcon 9 rocket broke apart over the Atlantic Ocean, David Giger shouted into his headset, “Dragon is alive!”

In the decade since he joined the company straight out of graduate school, Giger had taken on management of the entire Dragon program, reporting directly to Elon Musk. He watched the CRS-7 launch from mission control in Hawthorne not with a particular role, but rather providing a leadership presence. Giger could sense the Dragon mission team, mostly younger engineers, freeze up as video showed debris from the rocket showering back to Earth. A lot of the people involved in the hairy early flights of Dragon, including the C2 mission in 2012, had moved on to other positions at SpaceX or departed.

Read full article

Comments

 Introducing our new ‘Programming with AI’ unit

In the age of AI, we believe kids still need to learn to code. We’re also testing how AI tools can be used to help with teaching and learning programming. This blog dives into our new classroom unit ‘Programming with AI’, which we’ve developed for teachers to introduce students to learning to code using tools built on large language models (LLMs). We share an overview of the unit, and the early excitement of educators and learners with whom we’re exploring this new approach to teaching and learning programming.

Students completing activities from lesson 1
Students completing activities from lesson 1

Bridging the gap in programming education

As India is accelerating its efforts to modernise digital and computing education, it’s the ideal place for us to test how best to introduce learners to AI-assisted approaches to programming and teach them responsible AI use.

We decided to design a ‘Programming with AI’ unit of eight lessons that support students aged 14–16 to progress from block-based programming to text-based Python programming, Python being one of the most widely employed programming languages in industry and academia. In the new unit, students are guided to use LLM-based tools safely, ethically, and responsibly in order to undertake data analysis projects in Python. Students learn how to use the tools, such as Gemini and ChatGPT, to deepen their understanding, experiment with ideas, and connect classroom concepts with real-world technological applications.

A group of male students at the Coding Academy in Telangana.

The ‘Programming with AI’ unit is currently being piloted across five residential schools under our partner, the Telangana Social Welfare Residential Educational Institutions Society. In the coming months, we will evaluate the impact of this AI-supported approach to learning programming on student engagement, motivation, and overall learning outcomes.

Key design decisions during unit development

During the design phase of the unit, we identified three main challenges we wanted to address:

  1. The first was the transition from block-based to text-based coding. The students we were designing for are familiar with block-based programming like Scratch, but Python needs precisely typed code (syntax). This is a huge shift, and we wanted to offer a way to bridge the gap.
  2. We also had to figure out how to teach prompting LLM tools to students who are learning English as their second language.
  3. Finally, LLMs are non-deterministic. This means that LLM tools don’t always produce the same output, even when given the same prompt. They also do not generate correct code consistently.

To solve the first problem, we decided that the initial lessons in the unit should start by showing familiar Scratch blocks right next to the equivalent Python syntax. This direct comparison makes the transition from block-based to text-based programmer smoother.

Comparison of Python syntax with equivalent Scratch block
Comparison of Python syntax with equivalent Scratch block

To teach prompting, we built a gradual path. The unit starts with a debugging code activity, where students get code with errors and a prompt to make an LLM tool fix the errors. Later, they learn to break down a prompt into their goal (what output they want) and the context (additional information for the LLM tool). Eventually, they use a structured framework to prompt the tool to generate complete, bug-free code.

Handling the unpredictable outputs of LLM tools we turned into a learning activity. In one lesson, students are given different chatbot-generated code segments and asked to analyse and reflect on the code quality. This crucial exercise helps them build the critical thinking and code evaluation skills they need for programming with AI tools.

Illustration showing the progression of Python and generative AI concepts covered in the unit
The progression of Python and generative AI concepts covered in the unit

Over the eight lessons, there is a progression of Python concepts and generative AI concepts, as shown in the image above. They lead students to complete two challenging, real-world data analysis projects.

Preparing the educators

To ensure the successful pilot of the ‘Programming with AI’ unit, a comprehensive teacher training program was essential. We understood that teachers needed confidence and competence not only in the subject matter — programming — but also in guiding students to ethically and effectively use LLM tools.

STEM and ICT teachers participated in the training
STEM and ICT teachers participated in the training

The training was an intensive, three-day programme: two days were held in-person for deep collaboration and one day was virtual for reinforcing key information. We covered five critical areas:

  • Hands-on activities to ensure comfort with the Python programming language
  • Exploring the capabilities and limitations of LLM tools
  • Effective prompt writing, which teachers would in turn model for their students to use for code debugging and idea generation
  • Reviewing lesson flow and learning objectives to understand the structure of a lesson
  • Adapting specific pedagogies to integrate an LLM tool in teaching the Python programming language

This training ensured our educators were not just teaching a new unit, but felt like confident facilitators ready to lead the shift towards AI-supported learning.

Initial student reactions: Excitement, familiarity with LLM tools, and seamless transition

The pilot of the ‘Programming with AI’ unit began during the second week of December 2025 across five selected schools. It was immediately met with enthusiastic student engagement — they were highly excited to learn Python programming using LLM tools.

When teachers initiated discussions about AI, students demonstrated their familiarity with LLMs, chatbots, and related concepts. They were not only aware of generative AI but could readily name and describe the functions of LLM tools, telling us they were actively using these tools to create projects in various other subjects.

Students could easily connect with the activities in the lessons as they had foundational understanding of programming, which they developed while learning block-based coding in Scratch. This prior knowledge enabled their seamless transition to the new learning. Perhaps most impressive was the students’ rapid adoption of LLM tools to debug their code. This skill can ease what is traditionally a steep learning curve when students first learn text-based programming.

We are now eagerly anticipating their curiosity and progress in the upcoming lessons.

Looking ahead and collecting impact data

With the initial enthusiasm confirmed, the pilot program moves into its next critical phase. Teachers will continue to deliver the unit over the next few months in the five selected schools in Telangana.

STEM and ICT teachers participated in the training
The teachers and team at our ‘Programming with AI’ training.

During this extended observation period, our primary focus will be on rigorous data collection and impact assessment. We will closely monitor how students interact with the provided learning materials and how they use LLM tools throughout the unit. We will systematically assess whether the new teaching approach successfully helps students achieve the defined learning objectives for Python programming. Crucially, we will capture comprehensive data to understand the impact of this AI-integrated approach on engagement, skill acquisition, and problem-solving compared to traditional methods.

Based on the synthesis and analysis of all the gathered data, we will share our gathered insights early next year. The findings will inform our plans for more widely introducing the AI-integrated unit and for developing similar content for other settings and age groups.

The post  Introducing our new ‘Programming with AI’ unit appeared first on Raspberry Pi Foundation.

Video Game Hardware Sales Had a Historically Bad November In the US

U.S. video game hardware spending fell 27% year over year in November to $695 million, according to market analyst company Circana. “This is the lowest video game hardware spending total for a November month since the $455 million reached during the November 2005 tracking period,” Circana says. Furthermore, only 1.6 million units of hardware were sold in the U.S. in November, which is “the lowest total for a November month since 1995 (1.4 million).” The Verge reports: The rising costs of consoles probably didn’t help. The PlayStation 5 and Xbox Series of consoles both turned five in November, but customers looking to pick up one of the consoles brand new are having to grapple with higher prices following price hikes this year. Those hikes have led to an “all-time November high” for the average price paid for a new unit of video game hardware of $439, Circana says — a number that’s up 11 percent from 2024. (In November 2019, the average price was $235, according to Circana analyst Mat Piscatella.)


Read more of this story at Slashdot.

Another Starship Clone Pops Up In China

Longtime Slashdot reader schwit1 shares a report from Ars Technica: Every other week, it seems, a new Chinese launch company pops up with a rocket design and a plan to reach orbit within a few years. For a long time, the majority of these companies revealed designs that looked a lot like SpaceX’s Falcon 9 rocket. The first of these copy cats, the medium-lift Zhuque-3 rocket built by LandSpace, launched earlier this month. Its primary mission was nominal, but the Zhuque-3 rocket failed its landing attempt, which is understandable for a first flight. Doubtless there will be more Chinese Falcon 9-like rockets making their debut in the near future. However, over the last year, there has been a distinct change in announcements from China when it comes to new launch technology. Just as SpaceX is seeking to transition from its workhorse Falcon 9 rocket — which has now been flying for a decade and a half — to the fully reusable Starship design, so too are Chinese companies modifying their visions.

The trend began with the Chinese government. In November 2024 the government announced a significant shift in the design of its super-heavy lift rocket, the Long March 9. Instead of the previous design, a fully expendable rocket with three stages and solid rocket boosters strapped to the sides, the country’s state-owned rocket maker revealed a vehicle that mimicked SpaceX’s fully reusable Starship. Around the same time, a Chinese launch firm named Cosmoleap announced plans to develop a fully reusable “Leap” rocket within the next few years. An animated video that accompanied the funding announcement indicated that the company seeks to emulate the tower catch-with-chopsticks methodology that SpaceX has successfully employed.

But wait, there’s more. In June a company called Astronstone said it too was developing a stainless steel, methane-fueled rocket that would also use a chopstick-style system for first stage recovery. Astronstone didn’t even pretend to not copy SpaceX, saying it was “fully aligning its technical approach with Elon Musk’s SpaceX.” And then, on Friday, the state-aligned China.com reported that a company called “Beijing Leading Rocket Technology” took things a step further. It has named its vehicle “Starship-1,” adding that the new rocket will have enhancements from AI and is billed as a “fully reusable AI rocket.”


Read more of this story at Slashdot.

MIT Grieves Shooting Death of Renowned Director of Plasma Science Center

An anonymous reader quotes a report from the Guardian: The Massachusetts Institute of Technology (MIT) community is grieving after the “shocking” shooting death of the director of its plasma science and fusion center, according to officials. Nuno FG Loureiro, 47, had been shot multiple times at his home in the affluent Boston suburb of Brookline on Monday night when police said they received a call to investigate. Emergency responders brought Loureiro to a hospital, and the award-winning scientist was pronounced dead there Tuesday morning, the Norfolk county district attorney’s office said in a statement.

The Boston Globe reported speaking with a neighbor of Loureiro who heard gunshots, found the academic lying on his back in the foyer of their building and then called for help alongside the victim’s wife. The statement from the Norfolk district attorney’s office said an investigation into Loureiro’s slaying remained ongoing later Tuesday. But the agency did not immediately release any details about a possible suspect or motive in the killing, which gained widespread attention across academic circles, the US and in Loureiro’s native Portugal.

Portugal’s minster of foreign affairs announced Loureiro’s death in a public hearing Tuesday, as CNN reported. Separately, MIT president Sally Kornbluth issued a university-wide letter expressing “great sadness” over the death of Loureiro, whose survivors include his wife. “This shocking loss for our community comes in a period of disturbing violence in many other places,” said Kornbluth’s letter, released after a weekend marred by deadly mass shootings at Brown University in Rhode Island — about 50 miles away from MIT — as well as on Australia’s Bondi Beach. The letter concluded by providing a list of mental health resources, saying: “It’s entirely natural to feel the need for comfort and support.”


Read more of this story at Slashdot.

Dolby and LG introduce a modular home audio system for CES 2026

LG is teaming up with Dolby for a new collection of speakers that the companies are unveiling ahead of CES 2026. The LG Sound Suite features a modular lineup the H7 soundbar, M7 and M5 wireless surround speakers and the W7 subwoofer. The speakers can be combined in more than two dozen different configurations, from a pair to a full 13.1.7 channel surround sound home theater setup.

The main pitch of the LG Sound Suite is its application of Dolby Atmos FlexConnect. This is the audio brand’s tech for optimizing sound from wherever wireless speakers are placed. It’s meant to deliver optimal sound even from unusual locations, even when you’re limited by outlet locations, furniture placement or other quirks of a room’s layout. The collaboration with LG brings FlexConnect to a soundbar for the first time. When the H7 soundbar is used as the lead device, the suite will bring Dolby Atmos FlexConnect audio to any television. In a future software update, LG will also bring support for Dolby Atmos FlexConnect directly to some of its premium TVs, including select 2025 models.

This article originally appeared on Engadget at https://www.engadget.com/audio/speakers/dolby-and-lg-introduce-a-modular-home-audio-system-for-ces-2026-010000126.html?src=rss

Senate Confirms Billionaire Entrepreneur Jared Isaacman As New NASA Chief

Longtime Slashdot reader schwit1 shares a report from Politico: The Senate on Wednesday approved Jared Isaacman for the top job at NASA — an unprecedented comeback after President Donald Trump yanked his nomination this spring. Senators confirmed the billionaire private astronaut in a 67-30 vote. Trump renominated Isaacman for NASA administrator in November, after pulling his original nomination in May. He cited Isaacman’s relationship with SpaceX CEO Elon Musk, with whom Trump had just had a falling out, as the rationale for his decision. Isaacman’s surprise rebound followed months of political jockeying and help from high-profile figures in Trump’s orbit. […] Isaacman garnered backing from lawmakers during his hearing by confirming his support for NASA’s Artemis moon-landing mission, a key prerogative for Capitol Hill. He also committed to instilling urgency at the space agency, citing China’s space ambitions.


Read more of this story at Slashdot.

Mountain Biker Jumps Through Two Open Semi Trucks Traveling In Opposite Directions

This is a video of Red Bull mountain biker Matt Jones jumping his ride through the open rears of two autonomous Scania semi trucks traveling in opposite directions. It’s a tight window (less than a second), but after lots of planning and practice, he successfully makes the jump. That’s cool, but you give me four Red Bulls and I’ll try it with absolutely zero preparation and a flat front tire. “Because Red Bull gives you wings?” What? No, Red Bull makes my girlfriend pee the bed if she drinks one after 4PM.

NASA finally—and we really do mean it this time—has a full-time leader

Jared Isaacman, a pilot and financial tech billionaire, has commanded two groundbreaking spaceflights, including leading the first private spacewalk.

But his most remarkable flying has occurred over the last year. And on Wednesday, he stuck the landing by earning formal Senate approval to become NASA’s 15th administrator.

With a final tally of 67 to 30, Wednesday’s Senate confirmation came 377 days after President Trump first nominated Isaacman to serve as NASA administrator. Since that time, Isaacman had to navigate the following issues:

Read full article

Comments

[$] LWN.net Weekly Edition for December 18, 2025

Inside this week’s LWN.net Weekly Edition:

  • Front: Civil Infrastructure Platform; COSMIC desktop; Calibre adds AI; Maintainer’s Summit; ML tools for kernel development; linux-next; Rust in the kernel; kernel development tools; Linux process improvements; 6.19 merge window part 2.
  • Briefs: capsudo; Asahi Linux 6.18; Pop!_OS 24.04; Vojtux; KDE Gear 25.12; Rust 1.92.0; Quotes; …
  • Announcements: Newsletters, conferences, security updates, patches, and more.

The Oscars Will Abandon Broadcast TV For YouTube In 2029

The Academy has struck a multi-year deal to move the Oscars to YouTube starting in 2029, ending decades on ABC and making the ceremony free to stream worldwide with YouTube holding exclusive global rights. Variety reports: The Oscars, including red carpet coverage, behind-the-scenes content and Governors Ball, will be available live and for free on YouTube to viewers around the world, as well as to YouTube TV subscribers in the United States. Architects of the agreement said they hope the move to YouTube will help make the Oscars more accessible to “the Academy’s growing global audience through features such as closed captioning and audio tracks available in multiple languages.” […]

The Academy had been seeking a new broadcast licensing agreement for the better part of 2025. Over the summer, several expected and unconventional buyers, including NBCUniversal and Netflix, had come into the mix as potential suitors. Insiders believe that YouTube shelled out over nine figures for the Oscars, besting the high eight-figure offers from Disney/ABC and NBCUniversal. Under the most recent contract, Disney was paying around $100 million annually for the Oscars — but given the ratings declines for the kudocast, Disney/ABC were reportedly looking to spend less on license fees.

[…] It’s not a secret that the Academy and Disney/ABC would occasionally have disagreements over the best path for the Oscars, including the show’s length, which awards to present and who should host. Now, on a streamer with no time limits, the Oscars can be any length, and the Academy likely has carte blanche to do whatever it wants with the telecast. “They can do whatever they want,” says one insider. “You can have a six-hour Oscars hosted by MrBeast.”


Read more of this story at Slashdot.

China reportedly has a prototype EUV machine built by ex-ASML employees

A report from Reuters claims that scientists in China have created a prototype of a machine that could eventually be used to produce semiconductor chips capable of powering artificial intelligence. Sources told the publication that a team in Shenzhen completed the prototype of an extreme ultraviolet lithography machine earlier this year and it is allegedly now undergoing testing. The EUV machine was reportedly made by former engineers from Dutch semiconductor supplier ASML. Reuters states that China is targeting production of its own EUV chips beginning in 2028, although other experts have projected 2030 as a more likely date. 

EUV is a supremely complicated bit of technology; we have an explainer below that gets into some of the details. It is at the heart of the chips made by companies such as Intel and TSMC, so any company trying to compete would also need access to EUV. Although the Chinese prototype is not yet making chips, it is reportedly able to generate the extreme ultraviolet light needed for chip manufacturing.

If confirmed, this development would put China in control of tech much sooner than analysts had previously expected. To date, EUV has largely been kept out of reach by Western companies and used as a bargaining chip by the US government. Chinese President Xi Jinping has placed a high priority on the country being able to produce its own semiconductors. “The aim is for China to eventually be able to make advanced chips on machines that are entirely China-made,” a source told Reuters. “China wants the United States 100 percent kicked out of its supply chains.”

This article originally appeared on Engadget at https://www.engadget.com/big-tech/china-reportedly-has-a-prototype-euv-machine-built-by-ex-asml-employees-235833756.html?src=rss