Cloud

MIT Unveils Dynamic Reasoning Method to Boost LLM Efficiency Researchers at MIT have engineered a more intelligent method for large language models (LLMs) to allocate computational resources during reasoning, significantly increasing their efficiency. This new technique enables LLMs to dynamically adjust the amount of computation used based on the complexity of a given question. This approach marks a departure from common methods that assign a fixed computational budget to every problem, which often wastes resources on simple queries and fails to solve more complex ones. By enhancing the reliability and efficiency of LLMs for complex reasoning, this development could lower the energy consumption of generative AI systems and enable their use in more critical, time-sensitive applications. The research is being presented at the Conference on Neural Information Processing Systems. ...

OpenAI Invests in Thrive Holdings to Accelerate Enterprise AI Adoption OpenAI is taking an ownership stake in Thrive Holdings to accelerate the adoption of artificial intelligence in enterprise settings. This strategic partnership will initially focus on the accounting and IT services industries, which feature workflow-heavy processes that can see immediate benefits from OpenAI’s platform. As part of the collaboration, OpenAI will embed its research, product, and engineering teams directly within Thrive Holdings’ companies. The primary goal is to enhance speed, accuracy, and cost-efficiency while improving the overall quality of services. This initiative aims to create a repeatable model for AI integration that can be expanded to other industries, representing a significant move to more deeply integrate AI into core business operations. ...

AI Achieves High Accuracy in Early Dementia Detection Using EEG Signals Researchers at Örebro University have developed two novel AI systems capable of detecting dementia by analyzing EEG (brain-wave) data. These systems can effectively differentiate between healthy individuals and those with dementia, including Alzheimer’s disease and frontotemporal dementia. One model, which utilizes a combination of temporal convolutional and LSTM networks, has demonstrated an accuracy of over 80%. A second model, designed with a focus on privacy and efficiency through federated learning, has reportedly achieved an impressive accuracy of over 97%. ...

Perplexity’s Comet AI Browser Faces Serious Security Vulnerability Allegations Perplexity’s Comet AI browser is facing significant scrutiny over alleged security vulnerabilities. Security researchers from SquareX claim to have discovered a hidden API that could allow attackers to execute arbitrary code on a user’s device, potentially leading to data theft or a full system compromise. Perplexity has vehemently denied these claims, labeling them as “fake news” and stating that no such vulnerability exists. This controversy follows previous security warnings from Brave and Guardio regarding other issues with Comet, including prompt injection vulnerabilities that could manipulate the AI into performing unintended actions. ...

Agibot’s A2 Humanoid Robot Sets World Record with 106-km Walk A Chinese-developed humanoid robot, the A2 from Shanghai-based Agibot, has set a new Guinness World Record for the longest continuous walk by a humanoid robot. The robot completed a 106.286-kilometer journey from Suzhou to Shanghai without shutting down. The trek, which ran from November 10th to November 13th, showcased the maturity of the robot’s hardware, advanced balance algorithms, and exceptional endurance. The A2 was powered by a rapid hot-swap battery system, enabling it to remain active for the entire duration. This record-breaking walk marks a significant milestone for the large-scale commercial deployment of humanoid robots in various industries. ...

AI & Machine Learning Breakthroughs: Tracing Ancient Life and Powering Research This collection of articles highlights recent advancements in Artificial Intelligence and Machine Learning. A groundbreaking new machine learning method has successfully identified the oldest chemical traces of life on Earth. In other news, a German research team has gained access to a powerful supercomputer to advance generative AI research, while new insights reveal how machine learning is fundamentally transforming global currency exchange. ...

Jeff Bezos Launches ‘Project Prometheus’ AI Startup for Manufacturing Four years after stepping down as Amazon’s CEO, Jeff Bezos is reportedly co-CEO of a new AI startup, Project Prometheus. The company will focus on developing artificial intelligence for engineering and manufacturing sectors and has allegedly secured a massive $6.2 billion in funding. Bezos will share the CEO role with tech executive Vik Bajaj. This new venture enters a highly competitive AI market, where substantial venture capital investments are being made in companies like OpenAI, raising questions about the long-term financial sustainability of the industry. ...

NTT Sets New AV Safety Standard with Launch of Mobility Subsidiary Japanese telecommunications giant NTT has established a new subsidiary, NTT Mobility, to consolidate its expertise in the self-driving vehicle sector. The new company, launching in mid-December, will offer a comprehensive service for deploying and operating autonomous driving services across Japan. This initiative aims to overcome challenges like ensuring safe autonomous vehicle provision and developing infrastructure suited for Japan’s roads, with a service launch goal of April 2027. Concurrently, an NTT-developed technology for assessing video quality from autonomous vehicle remote-monitoring systems has been adopted as a new international standard. The model, named Parametric object-recognition-ratio-estimation, verifies in real-time if video streams are clear enough for safe object detection, alerting operators when quality drops below critical safety thresholds. ...

OpenAI CEO Sam Altman Pushes for Government-Owned AI Infrastructure OpenAI CEO Sam Altman has called for governments to invest in and own their artificial intelligence infrastructure. In a statement, Altman clarified that OpenAI is not seeking government loan guarantees for its data centers, emphasizing that companies should bear their own business risks. He suggested that a government-owned AI infrastructure, including a “strategic national reserve of computing power,” would serve the public interest. This comes as OpenAI faces scrutiny over its significant investment plans and the immense costs associated with advancing AI technology. Altman believes the risk of having insufficient computing power is greater than the risk of overbuilding. ...

NVIDIA & Oracle to Build DOE’s Largest AI Supercomputer with 100,000 Blackwell GPUs NVIDIA and Oracle have announced a collaboration to build the U.S. Department of Energy’s largest AI supercomputer. The project will involve two new systems at Argonne National Laboratory, named Solstice and Equinox. The Solstice system will feature a record-breaking 100,000 NVIDIA Blackwell GPUs, while the Equinox system will include 10,000. Together, these systems are expected to deliver a combined 2,200 exaflops of AI performance. The supercomputers will be used to accelerate scientific discovery and research in areas such as healthcare and materials science. This initiative is part of a broader effort to advance the United States’ leadership in artificial intelligence. ...