Among those advocating a different direction is Yann LeCun, who said existing AI systems such as ChatGPT, Claude and Gemini are effective for tasks including coding, mathematics and text generation but are not designed to achieve human-like or even animal-level understanding of real-world environments.

Speaking during the VivaTech conference, LeCun said these systems primarily learn statistical relationships from vast datasets rather than developing an understanding of how the physical world behaves. He argued that this limitation makes them unsuitable for many robotics applications that require flexible reasoning and adaptation.

LeCun, who previously served as chief AI scientist at Meta before leaving in 2025, now leads Paris-based Advanced Machine Intelligence Labs (AMI Labs), where researchers are developing an alternative AI architecture known as Joint Embedding Predictive Architecture (JEPA).

Earlier this year, AMI Labs announced it had raised more than $1 billion in seed funding. The investment included backing from Nvidia and a fund managing the private wealth of Jeff Bezos, making it one of Europe’s largest seed funding rounds.

Why Researchers Want AI Beyond Language Models

LeCun argues that LLMs are highly capable within structured and predictable domains because they generate responses based on patterns learned during training. However, he said those systems do not possess an underlying model of physical reality.

To illustrate the difference, he described balancing a pen on its tip. While a person instinctively knows the pen will fall without needing to predict the exact direction, an LLM may attempt to generate a statistically likely outcome rather than reasoning about the uncertainty of the situation itself.

According to LeCun, JEPA seeks to address this challenge by creating abstract representations of the physical world. Rather than predicting every possible outcome, the system is intended to identify which information is meaningful while ignoring unnecessary detail, allowing it to reason more efficiently about real-world situations.

He said this capability could eventually make AI systems better suited for robotics, where machines must continually interpret changing environments instead of responding to fixed prompts.

Robotics Continues to Drive AI Research

Improving AI reasoning has become a significant objective for robotics developers, who have invested billions of dollars in humanoid machines capable of operating in human environments.

Although robotic hardware has advanced rapidly, teaching robots to perform everyday household activities such as loading dishwashers or ironing clothing safely remains technically difficult and expensive.

LeCun said current LLM-based approaches are unlikely to solve these challenges effectively because they are not built to interpret complex physical interactions.

World Models Gain Momentum Across the Industry

Other researchers share the view that future AI systems will require more sophisticated reasoning than today’s language models provide.

Ingmar Posner, who directs the Applied AI Lab at the University of Oxford and also serves as an Amazon Scholar, said future AI systems should be capable of explaining cause and effect, identifying what matters in a situation and evaluating alternative actions.

His research group has spent several years developing what he describes as a mechanistic world model, designed to organize knowledge so that information can be efficiently retrieved, combined and modified when solving problems.

World models have existed as a research concept for decades but received renewed attention following influential work published in 2018 by David Ha and Jürgen Schmidhuber. Their research proposed that AI systems could learn by building internal simulations of the world rather than relying solely on memorized patterns.

Since then, several organizations have expanded work in the field. Google has developed Dreamer world models, including a version that learned to collect diamonds in the video game Minecraft by imagining future scenarios during decision-making.

Additional research includes Genie from Google DeepMind, Gaia from Wayve, and work at World Labs, founded in 2023 by Fei-Fei Li.

Commercial Deployment Remains Ahead

Posner said it remains difficult to predict how quickly these newer AI architectures will mature, noting that the rapid arrival of generative AI systems surprised many researchers who had expected such capabilities to take decades longer.

LeCun said AMI Labs plans to continue refining its JEPA-based system through the remainder of this year, with the goal of introducing initial industrial deployments next year if development progresses as planned.

Looking further ahead, he said broader-purpose AI systems capable of performing many tasks with limited additional training remain the long-term objective. Even if those systems eventually exceed human capabilities in certain areas, LeCun argued that people will continue to play the central role in defining goals, asking questions and directing how AI is applied.

Tags: Artificial Intelligence, World Models, Yann LeCun, AMI Labs, Large Language Models, Robotics, Machine Learning, ChatGPT, Google DeepMind, Nvidia, JEPA, AI Research

The post AI Researchers Shift Focus Beyond Large Language Models Toward World Models for Real-World Intelligence appeared first on Journos News - Breaking News, World News, Top Stories, Todays Headlines and Flash Reports.

The event brought together robotics companies and researchers from Japan, China and the United States, with exhibitors presenting humanoid systems designed for industrial, logistics and service-related applications. While Japanese firms were once regarded as leaders in humanoid robotics, several Chinese companies drew significant attention at the summit for systems aimed at lower-cost production and broader deployment.

Companies including Toyota Motor Corporation and Honda Motor Co. appeared alongside Chinese robotics groups such as Booster Robotics, LimX Dynamics and Unitree.

The summit highlighted a broader shift in the humanoid robotics sector, where Chinese companies are increasingly focusing on manufacturing scale and lower hardware costs. Several systems on display emphasized mobility, dexterity and repetitive task execution rather than fully autonomous general-purpose operation.

One of the more visible demonstrations came from China-based High Torque, which presented its Mini Pi Plus humanoid robot. The company said the machine was designed as a compact humanoid platform and priced from about $5,500. Demonstrations included dancing and coordinated movement routines, although the system was not presented as capable of performing broad household or industrial tasks.

The growing Chinese presence in humanoid robotics mirrors developments previously seen in industries such as consumer electronics, smartphones and electric vehicles, where Japanese manufacturers initially led technologically before losing market share to companies focused on lower-cost mass production.

Technology author Tim Hornyak, who attended the summit, described the situation as similar to Japan’s so-called “Galapagos syndrome,” a term used to describe highly advanced domestic technologies that struggle to gain global commercial traction.

Japanese robotics developers nevertheless argued that the country retains strengths in precision manufacturing and reliability. Honda demonstrated a motorized robotic hand capable of manipulating small components and threading a needle, a task requiring high levels of mechanical control and stability.

Keisuke Tsuta, an assistant chief engineer involved in Honda’s robotics development, said the company’s systems emphasized durability and production quality. Similar robotic hand technologies were also displayed by several Chinese companies at the exhibition.

The summit also reflected growing interest in humanoid systems as a response to labor shortages in Japan, particularly in logistics and transportation sectors. Tokyo-based technology company GMO presented work involving a humanoid platform intended to assist with cargo handling and operational tasks linked to Japan Airlines.

According to presentations at the event, the humanoid platform incorporated robotics technology from Unitree. Organizers and participating companies said such systems are intended to operate in environments designed for human workers, potentially reducing the need for specialized automation infrastructure.

Longtime robotics researcher Hiroshi Ishiguro of Osaka University also appeared at the summit alongside a humanoid robot modeled after himself. Ishiguro said Japan’s social acceptance of robotics could remain an advantage as humanoid technologies move toward wider deployment.

Data referenced during the event included findings from a recent Pew survey indicating that Japanese respondents expressed lower levels of anxiety toward artificial intelligence than respondents in the United States.

Despite rapid advances in humanoid demonstrations, several showcased systems remained focused on narrowly defined tasks or choreographed movement routines rather than generalized autonomous behavior. Companies at the summit largely framed current humanoid development as an ongoing engineering and manufacturing challenge rather than a fully mature commercial market.

The exhibition underscored how competition in humanoid robotics is increasingly shifting from experimental research toward scalability, hardware supply chains and production economics — areas where Chinese manufacturers are expanding rapidly.

The post Japan’s Humanoid Robotics Sector Faces Rising Pressure From Chinese Rivals appeared first on Journos News - Breaking News, World News, Top Stories, Todays Headlines and Flash Reports.

Sam Altman, the chairman of Oklo, is not just known for his role at OpenAI, the artificial intelligence company he co-founded. He’s also backing a revolutionary nuclear energy startup. Oklo, which aims to develop innovative nuclear power technology, is one of many nuclear ventures receiving support from Silicon Valley’s tech elite. But is this just a strategic move for AI progress, or does it hold potential for addressing wider global energy challenges?

The Rising Need for Energy

As AI technologies evolve, their demand for power grows. Data centers, essential for supporting digital infrastructure and AI development, already consume a large portion of global electricity, and this demand is expected to rise. In the United States, energy consumption from data centers has surged by 50% since 2020, with projections suggesting that it could account for 9% of total electricity consumption by 2030. This puts significant pressure on energy supplies, and without new sources of reliable, clean energy, tech companies risk jeopardizing their own operations and sustainability promises.

In response, many tech giants, including Microsoft, Amazon, and Meta, are turning to nuclear energy as a solution. Unlike solar or wind power, which are intermittent, nuclear offers a stable, 24/7 energy supply essential for data centers. Microsoft, for example, has even secured a deal to restart a reactor at the infamous Three Mile Island, aiming to power its AI ambitions.

Tech Billionaires and Nuclear Ventures

Altman is not the only tech mogul interested in nuclear energy. He’s also invested in Helion Energy, a fusion energy startup, while Facebook co-founder Dustin Moskovitz, LinkedIn co-founder Reid Hoffman, and venture capital firm Mithril (backed by Peter Thiel) have also joined forces with Helion. Other notable figures, such as Bill Gates, have backed nuclear initiatives as well. Gates chairs TerraPower, a company developing innovative nuclear reactors, and has poured over $1 billion into the effort.

Meanwhile, tech giants like Google and Amazon are backing nuclear startups like TAE Technologies and X-energy, and Amazon founder Jeff Bezos has invested in the Canadian startup General Fusion. These investments point to a collective recognition that nuclear energy, particularly next-generation reactors, could be key to powering AI’s future while addressing climate change.

The Push for Clean, Reliable Power

The investments from tech leaders in nuclear power are often framed as a solution to the growing need for clean, baseload energy. “We need clean, baseload power that is free of carbon dioxide and methane emissions, and nuclear is a reliable, affordable source of that energy,” says Megan Wilson, Chief Strategy Officer at General Fusion.

General Fusion, which is still in the early stages of proving its technology, aims to create a safer form of nuclear energy: fusion. Unlike traditional fission reactors, fusion combines atoms instead of splitting them, making it more stable and less prone to catastrophic failures. While fusion is still years away from being commercially viable, it’s a promising long-term solution for both the energy and AI sectors.

The Safety and Regulatory Debate

Despite the optimism around next-generation nuclear energy, the industry faces significant challenges, particularly around safety and regulation. Many critics are concerned that the heavy influence of tech companies could lead to relaxed safety protocols and insufficient regulatory oversight. Edwin Lyman, Director of Nuclear Power Safety at the Union of Concerned Scientists, voices concerns that Silicon Valley’s push for quick development could lead to regulatory changes that undermine public safety. He warns that the industry’s clout could result in safety and security rules being weakened.

However, proponents argue that these tech-backed nuclear projects, particularly those led by Bill Gates and Sam Altman, are focused on improving safety features. For example, TerraPower’s reactors include “passive” safety mechanisms that automatically cool the reactor if something goes wrong, aiming to avoid accidents like Chernobyl. Gates himself has expressed confidence in the regulatory scrutiny that his company faces, emphasizing that the Nuclear Regulatory Commission (NRC) will ensure that safety standards are met.

Oklo’s Vision: Small, Scalable Reactors

Oklo, founded in 2013 by MIT graduates, is focused on developing small, modular nuclear reactors that can produce cleaner, more efficient energy. These “fast reactors” are designed to use less fuel and recycle used nuclear fuel, making them cheaper and more sustainable. While Oklo has made significant progress, receiving approval to begin site investigations for one of its small reactors in Idaho, experts caution that the journey from concept to actual energy production will be a long and costly one.

Altman’s venture has seen strong stock performance, thanks to the increasing demand for clean energy to power AI, but the company is still far from generating revenue. Oklo’s reactors remain in the theoretical phase, and it will likely take years before they are commercially viable.

TerraPower: The Bill Gates Approach

Gates’ TerraPower is also racing toward developing next-gen nuclear reactors. The company broke ground in June on a demonstration plant in Wyoming, which is expected to be operational by 2030. TerraPower’s design is simpler than traditional reactors, using sodium for cooling instead of water. The plant is being built near a retiring coal power plant, where it will eventually provide power to a utility company, PacifiCorp.

Like Oklo, TerraPower’s reactors are designed with passive safety features, minimizing the risks associated with traditional nuclear plants. Gates has invested heavily in TerraPower because he believes nuclear energy is essential for combating climate change. While the project has garnered attention, it still faces public skepticism, particularly regarding its safety and regulatory hurdles.

Challenges and Future Prospects

While the future of nuclear energy remains uncertain, the push for cleaner, more reliable energy is intensifying. Tech companies, with their vast financial resources, are playing a pivotal role in driving these innovations. However, the public remains wary, and significant challenges remain, including safety concerns, regulatory approval, and high costs.

As AI technology continues to evolve and its energy demands grow, the need for a stable, sustainable energy supply will only increase. Whether nuclear energy can meet that need—while balancing safety and regulatory concerns—remains to be seen.

For now, tech leaders like Sam Altman, Bill Gates, and Jeff Bezos are all-in on nuclear, hoping their investments will fuel both the AI revolution and a cleaner energy future. But the road ahead will be a long one, and the risks are high.

Source

The post Tech Giants Invest in Nuclear Power for AI’s Future appeared first on Journos News - Breaking News, World News, Top Stories, Todays Headlines and Flash Reports.

Google has unveiled a groundbreaking quantum computing chip, named “Willow”, claiming it can solve complex problems in just five minutes—tasks that would take the world’s fastest supercomputers an astonishing 10 septillion years to complete. This development marks a significant step in the quest for ultra-powerful computing based on quantum mechanics.

What Makes Willow Revolutionary?

Google describes Willow as a chip that incorporates key breakthroughs in quantum computing and believes it is a milestone on the path to a functional, large-scale quantum computer. Hartmut Neven, head of Google’s Quantum AI lab and self-proclaimed “chief optimist,” expressed confidence that Willow would contribute to practical applications in the near future.

While still experimental, Willow is a glimpse into a future where quantum computers could revolutionize industries by simulating intricate systems and solving previously insurmountable problems.

How Quantum Computing Works

Quantum computers differ fundamentally from traditional ones. Instead of using bits (binary values of 0 or 1), they process data in qubits, which can exist in multiple states simultaneously. This ability, stemming from the bizarre nature of quantum mechanics, allows quantum computers to tackle calculations exponentially faster than conventional systems.

Key features of quantum particles include:

• Superposition: The ability to exist in two places or states at once.
• Entanglement: A mysterious connection between particles that persists even across vast distances.

Credits CNET: Google Quantum AI Reveals Willow Quantum Computing Chip

Potential Applications and Implications

Quantum computing holds immense promise for solving problems across various fields, such as:

• Healthcare: Accelerating drug discovery and understanding complex biological processes.
• Energy: Designing nuclear fusion reactors and improving battery technology.
• Security: Developing or breaking encryption systems.

However, this power also raises concerns. Quantum computers could potentially bypass modern encryption techniques, prompting companies like Apple to introduce “quantum-proof” encryption to protect sensitive data.

When Will Quantum Computing Become Practical?

Neven predicts that commercial-grade quantum computers capable of real-world applications won’t emerge until the end of the decade. Initial use cases are likely to involve simulations where quantum effects are crucial, such as pharmaceutical development, energy systems, and advanced materials research.

A Quantum Future Awaits

While Willow is not yet a universal solution, it represents a significant step toward the quantum era. The ability to simulate and solve real-world challenges using quantum mechanics could revolutionize science, industry, and technology.

Quantum computing is still years—and billions of dollars—away from full maturity, but Google’s new chip gives us a tantalizing glimpse of a future that could redefine what computers are capable of achieving.

Source

The post Google’s Willow Chip: A Quantum Leap Towards Supercomputing Supremacy appeared first on Journos News - Breaking News, World News, Top Stories, Todays Headlines and Flash Reports.

Nvidia has posted strong third-quarter earnings that exceeded both sales and earnings expectations, continuing to benefit from the growing demand for its AI chips. However, the company’s shares fell by 2% in after-hours trading despite these positive results.

Key Financial Highlights

• Revenue: $35.08 billion, surpassing the $33.16 billion expected by analysts.
• Earnings per Share (EPS): 81 cents adjusted, better than the 75 cents per share expected.

Looking ahead, Nvidia forecasts approximately $37.5 billion in revenue for the current quarter, plus or minus 2%, exceeding the $37.08 billion analysts had predicted. This implies a year-over-year growth of about 70%, a significant slowdown from the 265% growth reported for the same period last year.

Growth Overview

Nvidia’s revenue surged 94% year-over-year for the quarter ending October 27, continuing the company’s growth trajectory. However, this marked a decline from previous quarters, where Nvidia posted 122%, 262%, and 265% growth, respectively.

Despite a slowdown in growth, Nvidia remains the primary beneficiary of the ongoing artificial intelligence boom. In 2024, Nvidia’s shares have nearly tripled, making it the most valuable publicly traded company.

Data Center Division Driving Growth

Nvidia’s data center business, which handles sales of AI processors and related components, now constitutes the majority of the company’s revenue. The division generated $30.8 billion, reflecting a 112% increase from the previous year. This exceeded the $28.82 billion analysts had expected. However, not all of this revenue came from chips—about $3.1 billion came from sales of networking parts.

Net Income and Margins

Net income for the quarter grew to $19.3 billion, or 78 cents per share, compared to $9.24 billion, or 67 cents per share, in the same period last year. Nvidia’s gross margin also increased to 73.5%, slightly above analyst expectations, driven by higher sales of data center chips.

AI Chips: Blackwell and H200

Nvidia’s next-generation AI chip, Blackwell, is now in full production, with 13,000 samples already shipped to major clients such as Microsoft, Oracle, and OpenAI. Nvidia’s CEO, Jensen Huang, highlighted that Blackwell is in “full production,” and it is now in the hands of most of the company’s major partners who are working to integrate the chips into their data centers.

Nvidia’s CFO, Colette Kress, confirmed that shipments of Blackwell chips will increase next year, with the company expecting “several billion dollars” in Blackwell-related revenue in the fourth quarter. Additionally, sales of Nvidia’s current-generation AI chip, the H200, saw significant growth during the quarter.

Both the Hopper and Blackwell systems face supply constraints, with demand for Blackwell expected to outstrip supply for several quarters into fiscal 2026.

Gaming Business Performance

Nvidia’s gaming division generated $3.28 billion in revenue, exceeding StreetAccount’s expectation of $3.03 billion. The increase in gaming sales is attributed to higher demand for GPUs for PCs and laptops, as well as increased revenue from gaming console chips. Nvidia’s chip powers the Nintendo Switch, contributing to the rise in console chip sales.

Smaller Segments: Automotive and Professional Visualization

While Nvidia’s gaming and data center divisions are the company’s largest revenue drivers, its automotive and professional visualization segments also showed growth.

• Automotive: Sales in this segment reached $449 million, a 72% increase year-over-year, largely driven by demand for self-driving car chips and chips used in robotics.
• Professional Visualization: This segment accounted for $486 million in revenue, marking a 17% increase from the previous year.

Conclusion

Nvidia continues to lead the AI chip market with robust growth, driven primarily by its data center business and strong demand for its next-generation chips like Blackwell. While the company is experiencing a slight slowdown in growth rates, it remains poised for continued success, with significant revenue expected from AI chip sales and sustained demand across its gaming and automotive sectors.

Source

The post Nvidia Reports Strong Q3 Earnings, Driven by Soaring AI Chip Demand appeared first on Journos News - Breaking News, World News, Top Stories, Todays Headlines and Flash Reports.