IEEE Spectrum
PulseAugur coverage of IEEE Spectrum — every cluster mentioning IEEE Spectrum across labs, papers, and developer communities, ranked by signal.
15 day(s) with sentiment data
-
AI accessibility challenged by global resource disparities
Ensuring AI accessibility globally is challenging due to disparities in technological, commercial, and geopolitical resources among nations. This digital divide hinders some countries from achieving sovereignty in their…
-
Researchers propose ARA format to replace PDF for AI-native scientific papers
A new research artifact format called ARA (Agent-Native Research Artifacts) is proposed as a successor to the traditional PDF for scientific papers. Developed by researchers from multiple institutions, ARA aims to make …
-
AI training method uses light beams to alter memory
Researchers have devised a novel technique for training AI systems by using light beams to directly modify their memory. This method involves directing light onto a receiver, which then alters the AI's memory through ge…
-
Consumer hardware sufficient for LLM inference, not training
Small to medium-sized Large Language Models (LLMs) can be effectively run on consumer-grade hardware, negating the need for massive data centers for inference. Only the model training process requires extensive server f…
-
US Bans Chinese Robots Amid Supply Chain Concerns · 1 source tracked
The U.S. is implementing new policies to ban Chinese robots, citing national security concerns related to potential data backdoors. This move, spearheaded by figures like Congressman John Moolenaar, targets complete rob…
-
NASA satellite uses Google's Gemma 3 for in-space image analysis
A NASA satellite has successfully utilized Google's Gemma 3 AI model to perform image analysis while in orbit. This application demonstrates the potential for lightweight AI models to process data in space, which could …
-
New Genie Coefficient Aims to Measure AI Agent Intent Understanding
A new metric called the Genie Coefficient has been proposed to measure how well AI agents understand and fulfill user intent. As AI systems gain more autonomy and control, accurately interpreting human meaning becomes c…
-
Researchers propose new AI-friendly paper format to aid scientific collaboration
A group of 37 researchers from top universities and tech companies has proposed a new format for scientific papers, called an "Agent-Native Research Artifact" (ARA). This new format is designed to be more efficient for …
-
R&D Waste Persists Despite AI Adoption, New Report Finds
A new report from IEEE Spectrum and Wiley, sponsored by Patsnap, indicates that despite increasing AI adoption, many organizations are still wasting significant R&D budgets on projects that fail to reach commercializati…
-
NVIDIA unveils first AI-focused Superpod for rapid computation
IEEE Spectrum's Computing Editor, Dina Genkina, was given a tour of NVIDIA's inaugural Superpod. This specialized computing system is built to handle AI workloads, featuring numerous racks of GPUs designed for rapid com…
-
Robert N. Charette concludes IEEE Spectrum tenure after 20+ years
Robert N. Charette, a long-time contributing editor for IEEE Spectrum, is concluding his tenure after over two decades of writing about software engineering and risk management. Charette, known for his insightful articl…
-
AI model folds DNA into custom shapes for medical applications
A new generative AI model has been developed that can fold DNA into custom shapes based on user designs. These intricate DNA structures hold potential for applications in designing immunotherapies and advanced drug deli…
-
AI hyper-scaling digital inequality, concentrating compute and skills globally
Artificial intelligence is exacerbating global digital inequalities, with a significant concentration of AI compute and cloud services in the United States. This concentration leads to dependency for most countries, lim…
-
AI enhances radar and electronic warfare against agile threats
A white paper from IEEE Spectrum and Wiley explores the application of AI and machine learning in cognitive radar and electronic warfare systems. It details how mode-agile threats are outpacing traditional static system…
-
AI's power demands are rewriting electrical grid rules
The increasing power demands of artificial intelligence are creating significant operational challenges for electrical grids. These demands are not just about the sheer volume of energy consumed, but also about the unpr…
-
Flash memory advancements could speed up AI recall
Researchers are exploring how advancements in flash memory technology can accelerate AI recall processes. This development aims to improve the speed at which artificial intelligence systems can access and utilize stored…
-
AI's Growing Energy Footprint Detailed by IEEE Spectrum
A recent article from IEEE Spectrum highlights the significant and growing energy consumption associated with artificial intelligence technologies. The piece delves into the substantial electricity demands of AI, partic…
-
AI models struggle with spreadsheet tasks, research shows · 2 sources tracked
Large language models have historically struggled with spreadsheet manipulation, a limitation highlighted by recent research. This difficulty stems from the complex, structured nature of spreadsheets, which differs sign…
-
AI adoption paradox: heavy users fear job loss, smaller models enable offline use · 4 sources tracked
New research indicates a paradox in AI adoption: workers who utilize AI the most are also the most concerned about job displacement. Concurrently, smaller AI models are being developed to enable AI use in environments w…
-
Sideways chip stacking boosts memory density for AI needs
Researchers are exploring a novel method of stacking silicon chips sideways to enhance memory density. This technique aims to address the increasing memory demands posed by artificial intelligence applications. The appr…