Həftəlik İcmal #2 — Elm, Texnologiya və Kosmos
Hər kateqoriya üzrə mühüm xəbərlər və analitik icmal.
Hər həftə təqdim etdiyimiz bu analitik icmalda texnologiya, elm, kosmos və akademik tədqiqat sahələrində qlobal miqyasda rezonans doğuran ən mühüm inkişafları bir araya topladıq. Məzmun mənbələrin etibarlılığı və sənaye əhəmiyyəti əsasında ciddi şəkildə seçilmişdir.
🚀 Kosmos və Roket Buraxılışları
1. APOD: 2026 August 22 – Mostly Perseids
Mənbə: NASA Breaking News
APOD Science APOD APOD: 2026 August 22… Today’s APOD Archive Submissions Index Search Calendar RSS Education About Discuss APOD Astronomy Picture of the Day Discover the cosmos! Each day a different image or photograph of our fascinating universe is featured, along with a brief explanation written by a professional astronomer. Mostly Perseids Explanation: Recorded […]
2. NASA Shares Views of August Solar Eclipse from Ground, Air, Space
Mənbə: NASA Breaking News
On Aug. 12, a total solar eclipse darkened skies over Greenland, Iceland, and Spain. As the Moon covered the Sun, it briefly revealed the Sun’s wispy outer atmosphere — the corona — to those in the path of totality who were lucky enough to have clear skies. NASA researchers and photographers were along the eclipse […]
3. TB 26-04 Updates and Modernization of NASA’s Chemical Equilibrium with Applications (CEA) Code
Mənbə: NASA Breaking News
For more information, contact Mark K. Leader, Glenn Research Center, mark.leader@nasa.gov Download the PDF version NASA’s Chemical Equilibrium with Applications (CEA) code is a foundational tool for propulsion system analysis. It provides equilibrium chemistry, rocket performance, shock, and detonation calculations used across NASA and the broader aerospace community. NASA Engineering and Safety Center (NESC) Activity […]
4. Webb Opens Treasure Chest
Mənbə: NASA Breaking News
NASA’s James Webb Space Telescope captured this Aug. 6, 2026, infrared image of part of the Carina Nebula, a star-forming region also home to the Cosmic Cliffs. This feature, called the “Treasure Chest,” is an object known as a cometary globule. A cometary globule is an isolated cloud of gas and dust with a dense, […]
5. NASA’s Artemis II Crew Set to Receive Congressional Space Medal of Honor
Mənbə: NASA Breaking News
President Donald J. Trump will award each of NASA’s Artemis II crew members the Congressional Space Medal of Honor at 11 a.m. EDT on Friday, Aug. 28, during a ceremony at the agency’s Johnson Space Center in Houston. NASA astronauts Reid Wiseman, Victor Glover, and Christina Koch, and CSA (Canadian Space Agency) astronaut Jeremy Hansen, […]
6. SpaceX launches 100th orbital mission of 2026
Mənbə: Spaceflight Now
The Starlink 15-20 mission represented the 99th Falcon 9 rocket launch of the year. Liftoff from pad 4 East happened at 12:25 a.m. PDT (3:25 a.m. EDT / 0725 UTC).
7. SpaceX launches 29 Starlink satellites from Cape Canaveral following last-minute scrub
Mənbə: Spaceflight Now
The Starlink 10-39 mission includes the 2,000th Starlink satellite to fly this year. Liftoff from pad 40 at Cape Canaveral Space Force Station is scheduled for 11:14 a.m. EDT (1514 UTC).
8. LandSpace becomes first commercial Chinese company to land an orbital-class booster
Mənbə: Spaceflight Now
The August 19 landing of its ZhuQue-3 rocket marked the first landing of a Chinese-made rocket using landing legs as well.
9. Astronauts remove failed antenna; run out of time to install a spare
Mənbə: Spaceflight Now
It was the 282nd EVA devoted to station assembly and maintenance, the fifth so far this year, the second for Menon and the first for Adenot, now the first French woman to walk in space. Total ISS EVA time now stands at 74 days 10 hours and 44 minutes since station assembly began in 1998.
10. SpaceX exceeds 11,000 satellites in low Earth orbit
Mənbə: Spaceflight Now
The Starlink 17-50 mission added another 24 broadband satellites to the company’s megaconstellation. Liftoff from pad 4E at Vandenberg Space Force Base happened at 9:01 p.m. PDT (12:01 a.m. EDT / 0401 UTC).
💻 Süni İntellekt və Proqramlaşdırma
1. Scrap (2006)
Mənbə: Hacker News
Bu mövzu üzrə rəsmi müzakirələr və yeniliklər davam edir. Ətraflı məlumat üçün keçidə baxa bilərsiniz.
2. NanoGPT Speedrun Frontier
Mənbə: Hacker News
Bu mövzu üzrə rəsmi müzakirələr və yeniliklər davam edir. Ətraflı məlumat üçün keçidə baxa bilərsiniz.
3. Why your local LLM feels dumber than it is
Mənbə: Hacker News
Bu mövzu üzrə rəsmi müzakirələr və yeniliklər davam edir. Ətraflı məlumat üçün keçidə baxa bilərsiniz.
4. ElevenLabs, TwelveLabs, ThirteenLabs
Mənbə: Hacker News
Bu mövzu üzrə rəsmi müzakirələr və yeniliklər davam edir. Ətraflı məlumat üçün keçidə baxa bilərsiniz.
5. Hister – A private, full content search index that you control
Mənbə: Hacker News
Bu mövzu üzrə rəsmi müzakirələr və yeniliklər davam edir. Ətraflı məlumat üçün keçidə baxa bilərsiniz.
6. From Atari to EVE Online: Building on 15 Years of AI Research in Games
Mənbə: Google DeepMind
Google DeepMind partners with game studios to prototype breakthrough AI gameplay.
7. Introducing Gemini 3.7 Flash
Mənbə: Google DeepMind
Bu mövzu üzrə rəsmi müzakirələr və yeniliklər davam edir. Ətraflı məlumat üçün keçidə baxa bilərsiniz.
8. Putting sign language AI into users’ hands
Mənbə: Google DeepMind
Introducing sign-language-to-text (SL2T), our breakthrough model powering new sign language features for Deaf and hard of hearing users.
9. WeatherNext: AI model achieves breakthrough in forecasting cyclones
Mənbə: Google DeepMind
Bu mövzu üzrə rəsmi müzakirələr və yeniliklər davam edir. Ətraflı məlumat üçün keçidə baxa bilərsiniz.
10. Gemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaboration
Mənbə: Google DeepMind
Gemini Robotics ER 2 helps robots reason, collaborate, and solve real-world tasks. It represents a step change in video understanding, tool orchestration, and multi-robot collaboration for robotic applications.
⚡ Hardware, İT və Mühəndislik
1. Stop Hunting, Start Solving: Accelerating Root Cause Analysis with Agentic AI
Mənbə: IEEE Spectrum
About this Webinar Turn Yield Excursions into Faster, More Confident Root Cause Analysis When a yield issue emerges, the answer rarely lives in a single system. Critical clues are spread across metrology data, tool traces, chemical analysis, and facilities systems, while growing data volumes make traditional dashboards slow, fragmented, and difficult to act on. What You’ll Learn: Discover how a purpose-built semiconductor analytics platform can help engineers connect insights across domains without moving data. See how Agentic AI, semiconductor-specific visualizations, and push-down compute enable faster investigation of yield excursions and process issues, even across billions of data points. In the session, a live demonstration shows how to conduct a multi-domain root cause investigation using Spotfire® Industry Pro. Key Takeaways: Understand why siloed manufacturing data delays yield recovery and inflates costs Learn how Agentic AI automates complex cross-domain analytics and visualization generation Explore methods for scaling high-performance analytics across massive fab datasets Who Should Attend: Yield, Process, and Integration Engineers; Fab and Manufacturing Operations Managers; Quality and Reliability Engineers; and Data & Analytics leaders supporting wafer fabs, foundries, OSATs, and IDMs who need to identify issues faster while maintaining confidence in decision-making. Save Your Spot! Join this webinar to learn how leading semiconductor teams are accelerating root cause investigations, scaling analytics across massive datasets, and transforming disconnected data into actionable manufacturing intelligence. Reserve your seat today. Register now for this free webinar!
2. Gaining Leadership Backing for Your Innovations
Mənbə: IEEE Spectrum
This article is part of our exclusive career advice series in partnership with the IEEE Technology and Engineering Management Society. Imagine this: You have a strong idea for a new product for your company. Your coworkers encourage you to move forward because they believe it could be the organization’s next big success. The idea clearly falls outside your department’s responsibilities, however, and you have no role in the product line. What should you do? Sit and wait for “the right group” to pick it up, or push the idea forward without knowing how or what it might mean for your current position? Such situations occur frequently. Many end up as missed opportunities, even though they could have significantly advanced the company’s technological or market position. Some organizations actively support such initiatives, allocating specific periods during the workday for employees to focus on developing their own ideas. Companies known for that include Google and 3M . They allow employees to pursue projects with a portion of their time, such as one day per week. Research that I conducted indicates it pays off for employee performance. Bootlegging and skunkworks At some companies, managers know such projects exist, but they deliberately turn a blind eye, allowing them to continue. Some employees persist through bootlegging or skunkworks projects. Bootlegging projects have not been approved by a manager or funded by the company. Skunkworks projects involve a small team within the company that has been given authority and funding to secretly research and develop potentially groundbreaking innovations during their off-hours. The term comes from Lockheed’s Skunk Works division, set up in 1943 in a rented circus tent to build the P-80 fighter jet in secret. It took just 143 days. The 3M Post-it Note came out of the company’s “15 percent culture,” described as a permitted bootlegging policy. It gives employees paid time off to pursue their own ideas. The company traces the philosophy to its longtime president and later chairman William L. McKnight . Company scientist Arthur Fry used the policy in 1974 to turn a colleague’s dormant adhesive into the first Post-it prototypes, after his own bookmarks kept falling out of his hymnal. There are several examples of high-visibility skunkworks projects. At Apple , Steve Jobs pulled roughly 20 people— pirates , as he called them—out of the company to build the original Macintosh computer in a building nicknamed Texaco Towers. In Walter Isaacson ’s biography Steve Jobs , he frames the idea as modeled on the skunkworks approach. Google’s Gmail system is frequently—and incorrectly—cited as a product of the company’s “20% time” policy. In a 2014 interview with Time magazine, the system’s creator, Paul Buchheit , said Gmail was in fact an official assignment. What the Gmail incubation did share with classic skunkworks projects was secrecy: For much of its three years in development, it was kept hidden from most people inside the company. If you want to drive change in your organization, build a promoter triad around your idea. At Alphabet , Google X—now known simply as X —operated as a secretive “ moonshot ” lab, kept hidden from most Google employees, according to a 2011 article in The New York Times . Google’s self-driving car project graduated from X to become Waymo , and Google Glass was likewise incubated there. The X team is now developing the second edition of Glass Enterprise , a successor aimed at industrial rather than consumer use. Amazon runs a comparable model through Lab126 , which, according to an article in Fast Company , evolved from a small skunkworks Amazon subsidiary into a hardware maker with nearly 3,000 employees. Lab126 delivered the Kindle in 2007 and the Echo in 2015. Then there are so-called submarine projects , which employees work on without permission and despite explicit disapproval. They can lead to disciplinary action and termination. Innovation management Innovation management theory offers a more structured and robust approach. It argues that successful organizational change requires support at several levels, according to “Teamwork for Innovation: The ‘Troika’ of Promoters,” published in R&D Management . The promoter theory, developed around 25 years ago, consistently shows that change projects are far more likely to succeed when they are supported on multiple organizational levels. A good idea alone is not enough; you need a network of technology, process, and power promoters to turn a concept into a fully implemented, scalable solution. First, you need a technology promoter: the person who has the idea, such as a new product, and possesses technical expertise and specific knowledge about the field or industry. Art Fry at 3M would be such an individual. How can you put that into practice as an individual? Start by clearly formulating your idea into a concise concept paper or one-page summary including benefits, technical feasibility, and potential business impact. Identify potential technology promoters (experts who can validate and refine your idea), and approach them early to strengthen the technical foundation. In parallel, map the relevant stakeholders and decision-makers, and identify process promoters who understand how decisions are made in your company. They could be colleagues in innovation, R&D, or business development who understand your idea and how it can benefit the company. The second is a process promoter: someone who might not know all the technical details but understands the organization’s formal and informal networks and knows how to navigate its processes, committees, and decision-making paths. This person can ensure the idea reaches the right stakeholders at the right time. In the 3M case, it would be a person from the organizational management department, often called an innovation manager . The key role here is to connect inventors such as Fry with people from other departments needed for further project development, such as manufacturing, quality control, and sales. Lastly, there’s the power promoter: a person in a leadership position who might not know the technical details but can allocate resources, eliminate obstacles, and maneuver through the company’s political dynamics. This individual has hierarchical power and acts as a sponsor of the idea or project. In the case of Fry, the person could be, say, the chief technology officer, but it also could be a middle manager who has the power for an individual field of action. The three-level promoter structure applies regardless of whether the change concerns a new product, new service, or internal process innovation. Engage potential power promoters by presenting a low-risk, small-scale pilot and a clear value proposition. Leaders are more likely to support ideas that are well prepared, vetted for potential risks, and backed by a small coalition. Building the promoter triad In short, don’t work in isolation. Systematically build alliances across expertise, networks, and hierarchical levels to create lasting change. If you want to drive change in your organization, build a promoter triad around your idea. The tech experts and leadership promoters are easier to identify. Process promoters are often found in corporate innovation management, R&D management, or strategy functions, but they also can emerge in line units with strong internal networks. Innovation management, as the promoter model describes it, looks nothing like the management structure most engineers are trained to expect. Traditional technical management runs on a single reporting line. With the promoter model, influence is spread across three people—technology, process, and power promoters—who may be in different departments, at different levels of seniority, and who might never share a reporting line. What holds the trio together isn’t a formal structure; it’s the idea itself, for as long as it takes to move the idea forward. That makes innovation management closer to networked, matrix-style leadership than to the pyramid most engineers picture when they hear the word management . It’s worth understanding both models before you decide which kind of impact you’re actually optimizing for. The Institute has covered the tension from the individual’s side in “ Tips for How to Think Like an Entrepreneur ,” “ Management Versus Technical Track ,” both published in partnership with the IEEE Technology and Engineering Management Society , and “ What to Consider Before You Accept a Management Role ” from the IEEE Spectrum Career Alert newsletter. All are worth a look if you’re weighing a formal management track against staying close to the technology itself. Remember: You don’t have to build your promoter network alone or only inside your own company. IEEE societies , sections and chapters , and technical committees , as well as the networking platform IEEE Collabratec , function as a ready-made cross-company network. They are practical places to find technology promoters with deep expertise in a field you don’t fully own yet, or to meet process and power promoters at other organizations who have built a promoter coalition around a similar idea. For more tips on how to advance your career, check out our Career Advice for Engineers, From Engineers collection .
3. IEEE Presidents’ Scholarship Honors Teen Innovators
Mənbə: IEEE Spectrum
About 16 percent of the global population—more than 1 billion people—live with some form of disability, according to the World Health Organization . Many of the disabilities affect independence and mobility. Three high school students working on inventions to help people with disabilities restore movement, translate thoughts, and navigate rough terrain had their work showcased at Regeneron’s International Science and Engineering Fair (ISEF) , held in May in Phoenix. Their projects earned them this year’s IEEE Presidents’ Scholarship awards. IEEE President Mary Ellen Randall presented the awards at a ceremony held during the fair. They also received an IEEE President’s coin, which students said was a highlight of their experience. Hollie Tang won this year’s IEEE Presidents’ Scholarship of US $10,000 for her wheelchair navigation system . The award is payable over four years of undergraduate university study and includes a complimentary IEEE student membership. Partap Sidhu , the second-place winner, received a $600 scholarship for his mind-controlled lower-limb exoskeleton. Third-place winner Calvin Shang Hung received a $400 scholarship for his rough-terrain robot. Sidhu and Hung also got complimentary IEEE student memberships. Established by the IEEE Foundation and administered by IEEE Educational Activities , the Presidents’ Scholarship recognizes high school students who demonstrate an exceptional grasp of electrical engineering, computer science, or another IEEE field of interest. Controlling movements with a tongue Holly Tang won the 2026 IEEE Presidents’ Scholarship of US $10,000 for her Tonguage project, which is a noninvasive, computer-vision-based human-machine interface. Lynn Bowlby Tang, a sophomore at Wilson High School in Hacienda Heights, Calif., secured the top prize for her Tonguage project: a noninvasive, computer-vision-based human-machine interface. Using tongue movements and a standard camera, the interface lets users control a computer and other digital tools as well as assistive technologies including wheelchairs. The tongue pad, one of the system’s core features, allows the user’s tongue to function as a directional cursor, while eye blinks serve as mouse clicks. Tonguage translates the person’s tongue and eye motions into actionable commands in several ways, such as the tongue’s position inside the mouth and continuous movement patterns. The system’s multimodality combines input from the tongue with other facial cues. The system includes a face-tracking feature for error prevention that verifies commands are coming from the intended user, disregarding anyone else who moves into the camera’s frame. That is a critical safety measure for a wheelchair-navigation application, Tang says. Accessibility was central to Tang’s mission. She built the system to run on relatively affordable, readily available laptop cameras rather than more costly specialized hardware. “Mobility conditions don’t discriminate,” she says. “They can affect anyone of any income, gender, and socioeconomic status.” Tang initially imagined Tonguage as a simple substitute for a keyboard and mouse. The more research she did, though, the more she realized that it could offer autonomy through applications such as wheelchair navigation, robotic arm control, and gaming, she says. “We’re so focused on trying to give people autonomy over just basic human tasks that we often leave out things like gaming,” she says. “They deserve the freedom to play games and enjoy entertainment as well.” Tang, who plans to pursue biomedical engineering, says a visit to a rehabilitation center solidified her purpose. “Including empathy in your technological solution is so important,” she says. “Empathy is hard to teach in a classroom, but it can be learned through experience, and through actually meeting people whose lives your work might change.” Mind-controlled exoskeleton Sidhu, a junior at Bethpage High School , in New York, took second place for NeuroGait , a mind-controlled, lower-limb exoskeleton . He says he was inspired by his volunteer work at a community center that lacked elevators. He saw individuals with mobility issues struggle to navigate the three flights of stairs. NeuroGait operates by reading the Bereitschaftspotential (BP), a faint electrical pattern that emerges one to two seconds before a person consciously initiates movement. Using a custom electroencephalogram (EEG) headset and a convolutional neural network (CNN), the system classifies intended movements and sends commands to a 3D-printed exoskeleton. Rather than rigid motors, the suit relies on pneumatic artificial muscles that Sidhu designed to mimic human anatomy. “The pneumatic artificial muscle in itself is so compliant that it’s able to adjust to the limitations of the human body,” he says. The technical specifications are striking: The CNN achieves a 99.9 percent accuracy in detecting a person’s intended movement, while the full system—from the brain’s signal to physical movement—operates at 95.2 percent accuracy, according to the results from 500 trials Sidhu conducted. Perhaps most impressively, Sidhu built the entire system for about $276, less than 1 percent of the $40,000 to $100,000 price tag of commercial exoskeletons, according to a 2025 revenue report from Roots Analysis. He says he hopes to bring NeuroGait to the community center where the idea for the project began. He attributes his success to staying current with research from institutions and organizations such as Boston Dynamics and MIT . “To be successful in research,” he says, “you have to know what’s being done right now.” A spider-inspired robot Hung, a sophomore at El Cerrito High School , in California, took third place for Math Into Motion: Robotic Hexapod for Hazardous Environments . The six-legged robot is designed to traverse terrain too unstable for humans or conventional robotic systems. With only weeks before the science fair deadline for entries and no prior electrical engineering experience, Hung began with an idea inspired by his interest in spaceflight: an insectlike robot. He had spent years watching rovers such as Curiosity and Perseverance struggle on uneven surfaces, leading him to hypothesize that a hexapod design would be better for rugged ground. As the project progressed, the humanitarian applications for his robot became clearer, he says. Watching news reports of the earthquake that struck Türkiye in 2023, as well as conflicts around the globe, Hung adapted his robot for use in disasters. The hexapod’s stable tripod walking gait, in which three legs stay grounded while the other three move, makes it well suited for navigating in collapsed buildings to locate survivors or to carry sensitive supplies such as insulin in conflict zones. The current version moves using three mathematical techniques. Inverse kinematics converts a target leg position into the motor angles needed to reach it. Linear interpolation breaks each movement into a series of smaller steps for smoother motion. And Euclidean transformations translate the robot’s travel direction into instructions that each leg can follow, regardless of the way a leg happens to be facing. Hung taught himself how to design a printed circuit board. He also taught himself 3D modeling, coding, and soldering. Figuring out the complicated mathematical transformations to coordinate legs facing different directions proved to be the toughest hurdle, he says. After seven months of development and trial and error, a critical circuit board failure in his third version nearly ended the project, he says. “There was a really strong moment of ‘Should I just give up?’” he recalls. He simplified the design and rebuilt it from the ground up. “I just decided to double down,” he says. The fourth version of the robot was the first that successfully walked across his living room floor. He advises aspiring engineers that “if you find the right project and it truly becomes your passion, designing it almost starts to feel like fun, and that’s what carries you through.” As the three young innovators demonstrate, the future of engineering goes far beyond technical ingenuity. Much is rooted in empathy and a commitment to human welfare. Through initiatives such as the IEEE Presidents’ Scholarship, the IEEE Foundation showcases and nurtures bright minds poised to shape the next era of assistive technology and robotics. For Tang, Sidhu, and Hung, the ISEF stage is just the beginning. They can look forward to impactful careers dedicated to advancing technology for the benefit of humanity.
4. From AI Copilots to Agent Swarms
Mənbə: IEEE Spectrum
The impact of AI on software development has been both profound and ever-evolving. Last year, I wrote about AMD’s plans to use AI not just for generating new lines of code, but also for other steps in the software development lifecycle (SDLC), such as triaging problems, debugging code, and testing the software. At the time, we were hoping for a 25 percent productivity boost from AI use over the course of two or three years. But with each new release, the capabilities of large language models (LLMs) improve dramatically—accelerating software development, increasing the quality of AI-generated code, and fundamentally reshaping how software is engineered. Now, just one year later, we have surpassed our productivity target, achieving a 30 percent overall productivity boost through AI. On top of that, we are rethinking not only how we use AI within the SDLC, but the structure of the SDLC itself. We believe that the biggest AI revolution in software engineering is still ahead. So far, we have largely been teaching AI how we perform tasks and asking it to mimic existing workflows. In many ways, this constrains AI to human patterns of thinking. The next transformation will come from collaborative swarms of AI agents capable of discovering solutions independently. Agents of today AMD began developing AI systems for code generation, testing automation, bug analysis, and code review in 2024. At the time, our objective was to achieve 25 percent AI-generated production code by 2027 while gradually automating larger portions of the SDLC. Measuring productivity is inherently challenging , but from the outset we have consistently tracked one objective metric: the percentage of source code generated by AI. Importantly, we count only code that passes all reviews and testing and is ultimately included in the final product. While AI-generated code is certainly not the only contributor to productivity gains, it is one of the few metrics that can be measured objectively and consistently. By this metric, we have crossed the 20 percent mark at the beginning of this year and are now progressing towards 50 percent across entire codebase. In some software components, more than 80 percent of the code is now generated using AI. Agentic AI has enabled us to include AI in every step of the life cycle: For code analysis and triage , agents are trained to analyze problem reports, identify and group similar requests, and highlight which code snippets are likely to need modification. For debugging and code generation , agents are directed to analyze a bug request and implement required code changes. For testing , the agents generate unit tests, and if those are passed, identify necessary integration and product-level tests. And finally, for the approval and release stage, agents prepare architecture summary, code change review, and full test results for engineers’ review and approval—and, if approved, integrate the changes into the next release. Agents of tomorrow Today, engineers create AI agents in their own image: They teach AI what they know about the system, how they would fix an issue, and how they would implement a change. This is already a major technological advancement. Engineers can create multiple “AI versions” of themselves, allowing these agents to work in parallel, scaling their expertise far beyond the limits of individual productivity. The limitation, however, is that these AI agents are still constrained by human thinking and human-defined approaches. AMD We believe the next major transformation in software engineering will occur when collaborative AI agent swarms can independently identify and develop solutions, guided by humans on what to solve rather than constrained by human assumptions about how the job should be done. Instead of providing detailed instructions on how to solve a problem, engineers will define the issue, the desired outcome, and the quality, performance, and system constraints, allowing AI agents to determine the optimal path to a solution. A swarm of AI agents will then work in parallel to generate, evaluate, and refine multiple solution approaches. These agents will automatically validate correctness, measure performance, test trade-offs, and compare alternative implementations against defined success criteria. Finally, AI agents will prepare ranked solution options, along with validation results and performance metrics, for engineer review and approval. The agents won’t be enhancing each step of the SDLC—they will be rewriting the SDLC themselves. To get to this point, we need to change how agents are trained. Today, improvement occurs one engineer and one agent at a time: An engineer reviews the output, refines the prompt, and repeats the process. To scale beyond this model, agents must continuously learn from one another, reuse successful strategies, and improve collaboratively across projects and teams. We are already moving in this direction by using multi-agent workflows extensively through agentic harnesses, such as Codex and Claude Code, while simultaneously developing our own internal multi-agent systems to support the next generation of AI-driven software engineering. A good example is our AI-driven effort to resolve issues in our Radeon Software eXperience (RSX). RSX is a user interface component that allows users to configure and monitor graphics driver behavior. In October 2025, we began using AI agents to automatically debug and fix reported RSX issues. Out-of-the-box AI tools delivered limited results, resolving only 6 percent of issues. The percentage of software issues fixed automatically by AI agents in AMD’s Radeon Software eXperience (RSX) has been growing steadily, reaching 75 percent in June 2026. As we analyzed failures and identified ways to improve, we built a learning loop—initially a largely manual process—to understand where the agents were falling short and how to improve them. Rather than retraining the underlying models, we refined the objectives given to the agents, allowing them to iteratively explore multiple approaches, evaluate the results against defined success criteria, and converge on better solutions. At the same time, advances in models and agent run-times further increased effectiveness. Together, these improvements significantly increased our resolution rate from 6 percent to more than 75 percent of RSX issues resolved by agentic loop. To make agents and agent swarms truly productive, we need a continuous learning loop that feeds errors and human interventions back into future agent workflows. The opportunity is to engineer this loop around clear, measurable goals. Each cycle captures new insights, making the entire AI engineering workflow smarter and more effective. Over time, this self-reinforcing loop—not just the underlying model—will become a key driver of AI progress. The evolving role of human engineers At AMD, we view AI as a means of increasing productivity, improving quality, and enabling employees to focus on higher-value work. Our goal is to empower our workforce with AI, not to reduce headcount. To support this transformation, we are investing heavily in AI education and training across the company. The way we work is evolving rapidly, and we want every AMD employee to be prepared to leverage AI confidently, responsibly, and effectively. As AI agents continue to improve, engineers will spend less time manually implementing solutions, focusing more on defining specifications, validating outcomes, and making the strategic decisions that drive innovation.
5. Digital Signal Processing Pioneer Bede Liu Dies At 91
Mənbə: IEEE Spectrum
Bede Liu, a digital signal processing pioneer, died on 7 May. He was 91. Liu was widely regarded as one of the founders of modern digital signal processing, a field that applies mathematical algorithms to analyze, modify, and transmit signals including sound, images, and video. The IEEE Life Fellow taught electrical engineering at Princeton for more than 50 years. From 1994 to 1997, he chaired the university’s electrical and computer engineering department . Liu’s research aided the transition from analog to digital processing of sound, images, and video. His work helped establish many of the mathematical and engineering techniques that underpin modern communications, multimedia systems, and consumer electronics. Although little known outside engineering circles, his work is embedded in technologies used by billions of people. The low-power digital signal processors that make cellphone calls, streaming video, and Internet communications possible can be traced to research he conducted in the 1970s and ‘80s. Liu received the 2018 IEEE Jack S. Kilby Signal Processing Medal for “sustained contributions to the analysis and the development of low-complexity realizations of digital signal processing algorithms.” “We stream music and video. We take photos with our phones, and we send them around. We don’t even think about it,” IEEE Life Fellow H. Vincent Poor said in an obituary for Liu . “But it’s all because of the signal processing, image processing, and video processing that’s been developed over the years, as well as other technologies that have grown up beside it and enabled it, like semiconductors. The development of these processing advances was exactly what Bede was a major part of.” Poor is a professor of electrical and computer engineering at Princeton. An impactful scholar and teacher Liu was born in Shanghai in 1934. During his childhood, his family relocated to Taiwan amid the upheaval of the Chinese Civil War . His father, Henry Liu Sr., was an electrical engineer. Liu earned his bachelor’s degree in electrical engineering in 1954 from the National Taiwan University , in Taipei. After graduating, he and his family moved to the United States. Liu and his father attended the Polytechnic Institute of Brooklyn (now the New York University Tandon School of Engineering ) together. They earned their master’s degrees in electrical engineering in 1956. Liu continued his studies at the school, earning a doctoral degree in electrical engineering four years later. In 1959 he was awarded a Bell Labs fellowship and worked at the company’s Murray Hill, N.J., location until he joined Princeton in 1962. “Liu was a highly impactful scholar and teacher—always thinking ahead of future needs and changing technologies,” said IEEE Life Fellow Peter J. Ramadge , a Princeton professor emeritus of engineering. Cellphones make use of a considerable amount of digital signal processing, Liu once noted. Many of the field’s advances, he added, involved making sophisticated processing practical on devices with limited computing power—which is the challenge that confronted generations of engineers designing portable electronics. Liu’s research contributions helped shape both the theory and practice of digital signal processing. With Abe Peled , a former graduate student, he authored the 1976 textbook Digital Signal Processing: Theory, Design, and Implementation , which is a standard reference for engineers. Published before digital signal processing had fully emerged as a distinct discipline, it helped define the subject for practitioners and students around the world. Liu also published 250 technical papers and was granted 12 U.S. patents. His papers are available to read on the IEEE Xplore Digital Library . The first patent granted to him and Peled was in 1976 for a hardware design that processed bits in parallel, rather than in sequence. The innovation greatly increased computing efficiency for data including sound and communication signals. Peled says Liu “demonstrated an openness to new ideas and a willingness to challenge the orthodoxy of the EE department at that time—which leaned heavily toward more theoretical information theory.” A mentor to well-known engineers Liu’s influence extended beyond his own research. He advised 53 doctoral students, many of whom went on to distinguished careers in academia and industry, including leadership positions at Google and IBM . One former student, computer scientist Robert Kahn , helped create the architecture of the modern Internet. Kahn, an IEEE Life Fellow, received the 2024 IEEE Medal of Honor . “His former students were very successful,” Poor said of Liu, “and I think that’s a testament to his skill as a mentor.” “Liu was a highly impactful scholar and teacher—always thinking ahead of future needs and changing technologies.” —Peter J. Ramadge Together with several Ph.D. students, Liu developed methods of filtering and compressing digital signals to mitigate errors and dramatically reduce the computation needed for signal processing. As digital signal processing moved from laboratories into commercial products, the impact of Liu’s ideas spread across industries. His research helped spawn the development of lower-cost and lower-power electronics and contributed to advances in mobile communications, multimedia technology, industrial automation, and biomedical imaging. A focus on media integrity and copyrights In the 2000s, Liu turned his attention to media integrity and copyright issues. “With the increasing accessibility of digital media source material, the protection of ownership and the prevention of unauthorized alteration has become an important concern,” he wrote in his 2002 book, Multimedia Data Hiding . The book, which he co-wrote with his former doctoral student IEEE Fellow Min Wu, discussed the theory, techniques, applications, and security of digital watermarking—hidden signals that could identify a genuine copy of a song, image or video to prevent unauthorized distribution or tampering. A Princeton team that included Liu, Wu, and another of his doctoral students uncovered serious vulnerabilities in watermarking technologies being considered by an industry consortium. They found that the standardization efforts were immature and would not protect against digital piracy. “Now nearly every copy of a Hollywood film given to a critic or theater carries a unique digital forensic watermark to prevent unauthorized redistribution,” said Wu. A force in the community Liu, an active IEEE volunteer, served on the IEEE Board of Directors in 1984 and 1985. He was the 1982 president of the IEEE Circuits and Systems Society . He was a member of the U.S. National Academy of Engineering , an academician of China’s Academia Sinica , and a foreign member of the Chinese Academy of Sciences . Outside the classroom, he was recognized for his humility, humor, enthusiasm, and generosity. When thinking of Liu, IEEE Life Fellow Kenneth Steiglitz says, cheer is the first word that comes to mind. Liu was “always ready with a positive remark, a quick smile or, maybe, some tips on the right way to cook a duck,” says Steiglitz, professor emeritus of computer science at Princeton. Liu encouraged his students to take on ambitious, unconventional projects, and he inspired students and colleagues with his adventurous spirit.
6. Real-time data for a better response to disease outbreaks
Mənbə: MIT News
The startup Kinsa uses its smart thermometers to detect and track the spread of contagious illness before patients go to the hospital.
7. Economist Antoine Levy is all over the map
Mənbə: MIT News
The PhD student is fascinated by local variations in economic activity, and how they drive national policies.
8. Two projects receive funding for technologies that avoid carbon emissions
Mənbə: MIT News
Asegun Henry, Paul Barton, and Matěj Peč will lead research supported by the MIT Energy Initiative's Carbon Capture, Utilization, and Storage Center.
9. The factory of the future, batteries not included
Mənbə: MIT News
Everactive provides an industrial “internet of things” platform built on its battery-free sensors.
10. Are we still listening to space?
Mənbə: MIT News
Despite the planet’s seeming standstill, graduate students continue to use LIGO to identify astrophysical events.
🔬 Elm, Biotexnologiya və Kəşflər
1. JWST finds early galaxies may be 4 times more massive than thought
Mənbə: ScienceDaily
Astronomers using JWST have discovered that massive early galaxies contain far more small, faint stars than expected. That hidden population could make some of these galaxies three to four times more massive than previous estimates. The finding makes it even harder to explain how enormous, mature galaxies formed so soon after the Big Bang. It could also suggest that planets around low-mass stars were more common in the early universe than scientists realized.
2. Exercise may work better for keeping weight off than losing it
Mənbə: ScienceDaily
Exercise may not be a powerful weight-loss tool on its own, but it appears to play a much bigger role in preventing lost weight from returning. By preserving muscle, supporting metabolism, improving fat burning, and regulating appetite and blood sugar, regular activity can make long-term weight maintenance easier.
3. Peppermint oil lowers blood pressure in just 20 days
Mənbə: ScienceDaily
Peppermint oil may have an unexpected benefit for people with mildly high blood pressure. Adults who took a small dose twice daily for 20 days saw their systolic blood pressure fall by an average of 8.5 mmHg, while a placebo group showed little change. Researchers say the inexpensive oil could eventually provide a simple additional option for managing elevated blood pressure.
4. This yeast supplement may boost cancer-fighting immunity
Mənbə: ScienceDaily
A simple yeast-based food supplement may help restore the immune system’s ability to fight cancer, according to new research in mice. Scientists found that yeast beta-glucan reprogrammed early immune cells in the bone marrow, leading to stronger, longer-lasting cancer-fighting responses against colorectal, skin, and breast cancer cells.
5. This frozen fiber makes light and sound interact 1,000x more strongly
Mənbə: ScienceDaily
Freezing the liquid core of an optical fiber produced an extreme environment where light and sound interact more than 1,000 times more strongly than in ordinary fibers. Researchers used the effect to create optoacoustic memory, potentially paving the way for lower-energy photonic computers and advanced quantum technologies.
6. Mysteries of the commonplace
Mənbə: Science News
Editor in Chief Nancy Shute reviews new research in static electricity and other science revelations in our September magazine issue.
7. ‘Thunderquakes’ do more than shake the ground — they map it
Mənbə: Science News
Scientists used fiber-optic cables to turn “thunderquakes” into underground maps that could help reveal hidden hazards beneath cities.
8. Plants under attack leave behind a surprising call for backup
Mənbə: Science News
Bacteria may build a memory of herbivore attack in the soil and coax new plants to attract wasps to rein in destructive pests.
9. Crossword: Lab Animals
Mənbə: Science News
Solve the crossword from our September 2026 issue, in which we celebrate some creature features.
10. Shock wave therapy for erectile dysfunction is on the rise
Mənbə: Science News
A sound wave–based therapy commonly used to dislodge kidney stones is gaining traction as a treatment for erectile dysfunction.
📚 Akademik Preprintlər və Tədqiqatlar
1. Physics-Informed Neural Networks as Fast Surrogate Models for Electrochemical Flow Reactors
Mənbə: arXiv Physics
arXiv:2608.19209v1 Announce Type: new Abstract: This work presents a physics-informed neural network (PINN) for modeling a transient two-dimensional electrochemical flow reactor with diffusion, migration, convection, and nonlinear anodic Butler--Volmer kinetics. The model is trained without labeled concentration data by embedding the governing transport equation and all initial and boundary conditions into a composite loss function. Spatial and temporal coordinates together with anodic overpotential, temperature, inlet concentration, maximum flow velocity, and diffusivity are used as inputs, allowing the network to predict concentration fields over a broad operating domain. Validation against finite-difference-based solutions shows strong agreement for representative transient and near-steady cases, with a mean relative space--time error of $(9.99 \pm 0.65)\times10^{-3}$ (sub-percent level) across the conditioning domain. PINN inference is faster than a traditional finite difference solver by a factor of $5.34$, thus reducing runtime by $81.3\%$. Generalization tests further show that the surrogate remains robust under in-domain boundary-focused sampling and controlled extrapolation, although anodic overpotential is the most challenging parameter due to its exponential effect on interfacial kinetics. The results indicate that physics-informed neural networks can serve as accurate and efficient parametric surrogates for electrochemical transport problems and provide a foundation for low-computational-cost digital-twin modeling of electrochemical flow reactors.
2. The near-wall cycle for skin-friction generation revealed through explainable deep learning
Mənbə: arXiv Physics
arXiv:2608.19213v1 Announce Type: new Abstract: Skin friction in wall-bounded turbulence is produced by intermittent near-wall motions, yet conventional coherent-structure definitions do not identify which individual events generate wall-shear stress, and thus friction. We train neural networks to predict the future velocity field and wall-shear-stress distribution in turbulent channel flow, and use SHAP-derived importance maps to identify the the input regions most influential for each prediction. The dominant events appear as paired objects: an upstream velocity-relevant region is associated with high-momentum motion toward the wall, while a downstream friction-relevant region marks the enhanced wall-shear-stress left in its wake. Tracking these pairs reveals a recurrent cycle of growth, streamwise elongation, decay, and occasional splitting into new wall-shear-producing events. These results redefine near-wall coherent structures not by what the flow looks like, but by what they do to wall-shear stress.
3. Radiotracer photoluminescence for element-specific identification of color centers
Mənbə: arXiv Physics
arXiv:2608.19219v1 Announce Type: new Abstract: We report on the implementation of a radiotracer photoluminescence spectroscopy setup at the ISOLDE radioactive ion beam facility at CERN, enabling element-specific identification of optically active defects in solids. The method combines radioactive ion implantation with optical spectroscopy, allowing the temporal evolution of photoluminescence signals to be correlated directly with nuclear decay. The setup is currently optimized for color centers in diamond and related wide-bandgap materials and enables room-temperature measurements. The system consists of an optical microscope coupled to a fiber-fed Czerny-Turner spectrometer with a liquid-nitrogen-cooled CCD detector, providing the stability required for long-duration measurements. As a proof-of-principle, radioactive $^{75}$Ga was implanted into diamond as a precursor to produce $^{75}$Ge impurities. The photoluminescence band extending from 600 nm, corresponding to the well-known GeV$^{-}$ center, exhibits an exponential decay with a half-life of $82.3^{+2.5}_{-2.3} \mathrm{min}$, in agreement with the known $\beta^{-}$ decay half-life of $^{75}$Ge of $82.78(4) \mathrm{min}$. This establishes a direct and unambiguous correlation between the observed spectral feature and its germanium origin. These results demonstrate the capability of the setup to perform element-specific optical spectroscopy and extend radiotracer methods to color centers in wide-bandgap materials, taking advantage of the uniquely broad range of radioactive isotopes available at ISOLDE.
4. UPLOAD-ANYON: Progress Towards an Optimised Twisted Anyon Cavity with High-Sensitivity AM-Noise Detection for an Ultralight Axion Dark Matter Search
Mənbə: arXiv Physics
arXiv:2608.19225v1 Announce Type: new Abstract: We propose a superconducting single-mode microwave haloscope based on helical cavity resonators for the detection of ultralight dark matter axions over the mass range $10^{-18}$ to $10^{-13}\,\mathrm{eV}$. Building on the single-mode helical-cavity concept introduced by Bourhill et al. [Phys. Rev. D 108, 052014 (2023); arXiv:2208.01640], we use an inverse-design framework to develop practical resonator geometries compatible with superconducting niobium fabrication. The optimisation employs a figure of merit derived to minimise the measurement time required to achieve a fixed experimental sensitivity. Relative to the heuristic M\"obius-geometry benchmark, the best subtractively manufacturable bulk-niobium design achieves a figure of merit more than three orders of magnitude larger. An experimentally informed microwave interferometric readout model, incorporating measured electronics noise and active suppression of pump amplitude noise, is used to project the sensitivity of the proposed experiment. For an acquisition time of three months, the haloscope is projected to reach $g_{a\gamma\gamma}<10^{-11}\,\mathrm{GeV}^{-1}$ across more than four orders of magnitude in axion mass. The projected sensitivity extends approximately one order of magnitude below the current exclusion limits set by CAST, providing a practical pathway towards a high-sensitivity direct search for ultralight dark matter axions.
5. Serenity-S1, a CMS Tracker Readout and Data Processing Card for HL-LHC
Mənbə: arXiv Physics
arXiv:2608.19226v1 Announce Type: new Abstract: The CERN Large Hadron Collider will be upgraded to the High-Luminosity LHC, increasing the particle-collision rate and the demands on detector readout. A key element of this upgrade is the new CMS Tracker, whose silicon sensors will generate data at an aggregate rate of about 10 TB/s. This data must be routed efficiently to processing systems and filtered before storage. Serenity-S1 is an ATCA readout and data-processing card based on an AMD Virtex UltraScale+ VU13P FPGA with 128 high-speed serial transceivers. Twenty 12-channel Samtec FireFly optical modules provide up to 120 transmit and 120 receive channels at line rates of up to 25 Gb/s, while an additional optical module supports CMS Tracker links at up to 10 Gb/s per channel. The modular optical layout, combined with the large FPGA, makes Serenity-S1 suitable for several applications within and beyond CMS. ATCA provides the power and cooling required for operation in the CMS service cavern. Low-level management and the IPMI interface are provided by an OpenIPMC mezzanine, while an AMD Kria K26 system-on-module running AlmaLinux 9 performs higher-level board management and application control. Firmware and software frameworks provide common infrastructure so that users can focus on application-specific logic. The paper describes the hardware, firmware and software architecture of Serenity-S1 and its deployment in CMS Tracker integration and R&D activities.