FlashFeed

Full text

With AI data centers swallowing up every bit of memory storage around, it should come as no surprise that some creative engineers might be looking at alternatives. Most aren't going to be considering solutions from the 1950/1960s, of course, but that's precisely what one self-professed tinkerer did to create the "world's worst USB device".

Tech YouTuber polymatt recently posted a video of the reasoning, technology, and construction of the aforementioned device , and it's well worth watching, even if you have no interest in something that can only store 64 bits of information. The reason why is that instead of using modern storage techniques, the storage medium of choice is magnetic-core memory.

These were all the rage during the 1950s and 60s, before silicon-based storage became the norm, and some of its more famous applications include ENIAC , the IBM 704 , and the Apollo Guidance Computer . Essentially, magnetic-core memory works by using tiny rings of ferrimagnetic materials, through which multiple wires are woven.

By using the wires to generate a magnetic field, the ring's magnetic polarity can be altered, with the direction equating to 0 or 1 in terms of binary data. No power is required to maintain the polarity, so magnetic-core memory is classed as non-volatile storage, just as NAND flash is.

Unfortunately, despite using really small ceramic rings, polymatt's USB device tops out at a mere 64 bits of storage. In contrast, IBM's 1957 magnetic-core unit had a capacity of 147,456 bits, though it did weigh several hundred kilograms and apparently cost $6,000 per month to rent.

64 rings means 64 bits. Only 16 million times more to have 1 GB drive. (Image credit: Polymatt/YouTube) Anyway, if you do watch the video, you'll see just how fiddly and delicate the whole project was for polymatt, though it's a little funny to note that the Espressif ESP32 microcontroller (used to handle the USB interface and overall read/write management of the device) has embedded flash storage that's millions of times larger than the little magnetic-core device.

Although there's no real point in comparing it to any external SSD that you can buy right now , I can't help but think that something has been lost with silicon-based storage. Sure, it's massively faster and far greater in capacity than anything you could achieve with magnetic-core memory, but a handful of chips on a PCB just doesn't have the same vibe.

Polymatt's creation is a thing of the past, but it's beautiful to look at, and I so want to make my own now. Looks like my Lego Milky Way will just have to wait a little longer to get completed now.

Full text

The first dinosaur found in Antarctica belonged to a group that included the largest animals ever to walk the planet, a new study finds.

A backbone from the 82 million-year-old giant was discovered more than 40 years ago, but at the time, researchers assumed it came from an ancient marine reptile. Now a new study, published June 29 in the journal Acta Palaeontologica Polonica , has revealed that it was actually a titanosaur — the group of long-necked sauropods that included the largest land animals on record.

"At first glance this appears to be an unremarkable fossil, but it holds an important place in the history of Antarctic exploration as the first dinosaur fossil found on the continent," study first author Paul Barrett , a paleontologist at the Natural History Museum in London, said in a statement .

While this fossil now holds the record for the first dinosaur fossil found in Antarctica, other dinosaur fossils were identified on the continent after its discovery, so it's not the only known Antarctic dinosaur. In fact, researchers have identified a variety of dinosaurs on Antarctica, with another sauropod fossil identified as a titanosaur in 2011.

The newly identified dinosaur was around 20 to 23 feet (6 to 7 meters) long. That's very small compared with the largest-known titanosaurs, which could grow up to 123 feet (37.5 m) long. However, as the fossil is just a fragment of a vertebra, researchers are unable to narrow down which species the Antarctic titanosaur belonged to, and it's possible that the individual was only a juvenile when it died.

Mike Thomson, a British Antarctic Survey geologist, found the fossil during an expedition to James Ross Island in 1985. The island, which has been the site of multiple dinosaur discoveries, is located off the northeastern Antarctic Peninsula, south of the 600-mile-wide (965 kilometers) Drake Passage that separates South America from Antarctica.

Antarctica is well known for its icy landscapes, but when titanosaurs roamed Earth, the continent was still attached to South America and full of temperate forests. Antarctica's dinosaurs were so far south that they would have lived in constant twilight during the winter months, according to a news article published by the Natural History Museum.

The authors of the new study identified the dinosaur using high-resolution CT scans, which enabled them to look inside the fossil. Dating back to the Cretaceous period (143 million to 66 million years ago), the titanosaur lived in the last age of the non-avian dinosaurs, before an enormous asteroid hit what is now Mexico's Yucatan Peninsula 66 million years ago and wiped them out.

RELATED STORIES

'Last titan' of Thailand discovered, and it's the longest-necked dinosaur on record from Southeast Asia

Giant 'cow of the Cretaceous' discovered almost 100 years ago identified as new duck-billed dinosaur

'He began to cry, and almost fell to the floor': The fluffy fossil that finally showed the world that birds are dinosaurs

The Antarctic titanosaur helps researchers better understand how dinosaurs spread across Earth's southern continents, which at the time were combined into a supercontinent called Gondwana. The presence of titanosaurs on Antarctica suggests that they may have used Antarctica to travel from what is now South America to New Zealand.

Other dinosaurs identified on Antarctica include small herbivores, armored ankylosaurs and bipedal predators like Imperobator , which would have shared the forests with the newly identified titanosaur. While researchers are beginning to piece together Antarctica's ancient ecosystems, they still have a lot to learn about the dinosaurs that lived there.

"There are likely many more dinosaurs to be discovered on the continent," Barrett said. "As climate change causes ice to retreat we may indeed find further evidence of this past biodiversity."

Full text

This weekend, the United States will celebrate its 250th birthday, a momentous occasion that has been anticipated across the land going back to the beginning of the new year. The country is awash in patriotism in many ways, from city streets lined with red, white, and blue decorations to advertisements honoring the milestone, and even a new Larry David series revisiting famous moments from American history with cringe humor and petty grievances. None of this would have happened, though, without the actions of key Revolutionary War players who fought against tyranny and secured independence from Great Britain. Everyone knows General George Washington , the founding fathers, and Paul Revere , but a new graphic novel Kickstarter, courtesy of the creator of the original Prince of Persia video games, Jordan Mechner , from Oni Press and Magnetic Press, is shining a light on the unsung heroes of the war effort.

Full text

Opened in 1971 on the site of a former Royal Air Force base, few places in Northern Ireland tell the story of the Troubles as powerfully as the former HM Prison.

Referred to as 'Maze Prison', 'Long Kesh' or 'the H-Blocks', it became one of the most infamous prisons in Europe and worldwide, housing both republican and loyalist paramilitary prisoners during decades of conflict.

The prison is best known for pivotal events that shaped modern Irish and British history, including the Blanket Protest, the 1981 Hunger Strike led by Bobby Sands (whose face you may have seen on murals around Belfast), and the dramatic 1983 mass escape of 38 Irish Republican Army prisoners.

This was the largest prison escape in the history of the British Isles.

These events transformed the Maze from an ordinary prison into an internationally-recognised symbol of political conflict.

Today, little remains of the vast prison complex after most of it was demolished in the 2000s. Even though a stalemate exists on its destiny, several significant structures still survive, including the Grade A-listed H-Block 5 and the prison hospital, preserved due to their historic importance.

Although the site is not generally open as a tourist attraction, visitors interested in dark tourism, political history or the legacy of the Troubles often stop nearby to appreciate the immense historical significance of what once stood here.

Whether viewed as a place of tragedy, resilience or political change, Long Kesh remains one of the most historically significant and thought-provoking sites in Northern Ireland, offering a powerful reminder of a conflict whose legacy continues to shape the region today.

Full text

EXCLUSIVE Pentera Labs’ red teamers compromised a developer’s AI agent via his Claude Desktop app and ultimately turned that access into full remote code execution on the dev’s machine – demonstrating how an attacker could turn a trusted, chatty AI assistant into a double agent operating on their behalf. “Claude’s got a new voice,” Pentera's offensive security services team leader Dvir Avraham told The Register. “We acknowledge the huge trust in AI models – everybody uses them,” he said in a phone interview. “We used this trust to manipulate the victim, like under the hood, the victim didn't see it coming.” It also prompted Avraham to check his own platforms. “I became a little bit paranoid,” he told us. “I'm not allowing any command to run without me examining it twice.” In a report set to publish Wednesday, and shared in advance exclusively with The Register, Avraham and research technical lead Reef Spektor detailed the attack and what it means for organizations using agentic AI tools with local code-execution access. It began with a red-team assignment on a third-party platform that aggregates customer email inboxes into a single management interface. Avraham and Spektor won’t name the platform, or tell us exactly how they gained access to it. They used this compromised inbox – and told us any compromised inbox would work – to get into the victim’s Claude account. As the duo noted, breaking into an email inbox in real life – via a third-party management platform, phishing link, social engineering password reset, or even using AI agents – isn’t too difficult. “AI agents today have access to connectors and to direct MCPs into inboxes,” Spektor added. In addition to this prerequisite (compromised inbox), the attack chain also requires the victim to have Claude Desktop installed. Anthropic’s desktop app works across macOS, Windows, and Linux systems. It provides the same AI chat for conversations as claude.ai, and it also syncs across all devices and sessions tied to the user’s account. “We asked ourselves, can we leverage the sync behavior to infect other sessions and devices? (hint: yes!),” the red teamers wrote in the Wednesday report. Back to the AI Stone Age As of January, the desktop app also includes Cowork for longer agentic tasks, and Code for software development. So, for example, a user can send Claude a task from their phone and instruct it to work on their computer. As Anthropic says: “Anything you can do on your computer, Claude can do. Open apps, fill spreadsheets, navigate your browser. No setup, no passwords handed off.” The Cowork feature now makes Pentera Labs’ attack scenario even easier. However, when the security analysts were doing this research in November 2025, “back in the Stone Age in terms of AI, you didn't have Cowork or Claude Code, so we needed a way to actually execute commands because we wanted to take over the machine,” Avraham said. For this part, they took a keen interest in Claude Desktop’s personalization features. These are account-wide settings that tell the AI agent the user’s preferred approach and general communication instructions, along with more specific project instructions, such as guidelines for a particular workflow, or defined roles Claude should adopt within a project. The red teamers developed a base64-encoded prompt that instructed Claude to check for command-capable tools on the developer’s machine and execute the command if available, or produce a fake error message if not, prompting the user to download a tool that will execute the attacker’s commands. Then they pasted the prompt into the victim’s personal preferences on Claude, and this prompt syncs across all of the user’s devices. This ensures that the next time the user opens Claude Desktop and types in a chat, the poisoned instructions are loaded into their preferences and will silently run behind the scenes. The user thinks they are simply interacting with Claude as usual. They don’t see Claude checking to see what extensions and tools are installed. If the user already has Desktop Commander or a similar MCP connector or extension installed, the poisoned instructions tell Claude to use it. This allows the attacker, via Claude, to execute a stealthy reverse shell or other malicious code. “And from there it's full compromise of the machine,” Avraham said. Phishing - but without the email However, if there aren’t any command-capable tools installed, then Claude becomes what the researchers describe as a “phishing layer.” (They also noted that if they had performed this research more recently, not back in November, the Claude Cowork feature would have eliminated this entire tool enumeration and phishing phase because Cowork can execute commands on a user’s behalf.) The injected prompt instructs Claude to present a realistic-looking error as soon as the victim asks the chatbot a question. This includes a realistic error code, a link that purports to be a fix, and step-by-step instructions. “This message tells the victim: ‘please download this,’ and we took links from the actual Anthropic site, with known emojis that the AI loves,” Avraham said. Because the error message looks real and people usually trust their AI assistant, they will likely click on the link and execute the attacker-controlled command. “From here, the attacker has full command execution – reverse shells, data exfiltration, credential harvesting, whatever the objective calls for,” the duo wrote. “In our case, we had Claude curl a remote server we controlled on every interaction, fetching and executing whatever bash commands we served back. We could rotate those commands server side at will, effectively turning Claude into a persistent, stealthy C2 agent that the victim themselves kept feeding.” In this specific case, the target was a developer who had credentials and access to several internal systems. After compromising the dev’s workstation – which gave the red teamers a foothold into the organization – they moved laterally across the company using various attack vectors that they declined to tell us about, citing customer privacy and proprietary methods. But, Spektor added, developers make for an “excellent starting point for an attacker,” because of their access to secrets including API keys, tokens, and cloud credentials, which allows intruders to move from a single workstation into the larger organization’s cloud environment. From there, they’ve got free rein to steal source code and other sensitive data, or poison internal git repositories, and cause all sorts of pain for enterprises as we've seen play out multiple times across several recent attacks. Feature, not a bug The team reported their findings to Anthropic back in November, and the AI company essentially said it’s Claude Desktop working as intended – a feature, not a bug. “After reviewing your submission, we've determined this doesn't represent a security vulnerability that falls within our program scope,” Anthropic said. “Our current threat model treats personal preferences, skills, and MCP connectors as features that can execute code through Claude Desktop by design. While we recognize these features can be leveraged to execute arbitrary code when manipulated, this represents expected functionality rather than a security vulnerability in our infrastructure.” The Register reached out to Anthropic for comment and did not receive any response. The red teamers, however, have some suggestions to keep your organization safer from rogue AI agents. First, for anyone using agents or chatbots: pay close attention to what the AI can do on your machine, and don’t blindly follow install prompts or error messages. “If you can, run it on a sandbox and not on your personal computer,” Spektor said. Security teams should treat AI desktop apps as “privileged software” as they can execute code, read files, and interact with local tools. “Monitor for changes of AI assistant configurations and synced settings,” the researchers wrote. “Restrict which extensions and tools can be installed alongside AI apps.” And finally, red teams should add AI desktop apps to their assessment toolbox, Avraham and Spektor noted: “There's a real attack surface here that most engagements don’t cover yet.” ®

Full text

If the number of satellites in Earth's orbit exceeds 100,000, humanity may lose its ability to study the universe from the planet's surface.

That's the conclusion of a study conducted by astronomers from the European Southern Observatory (ESO) which warns that if existing plans to deploy a million orbiting data centers and tens of thousands sun-reflecting mirrors were to come to fruition, the world's most cutting-edge astronomical telescopes may as well be mothballed.

"We can reach conditions where basically, there is no point in operating the telescopes anymore because all the data will be corrupted. All. 100 percent," Olivier Hainaut, the director of operations at ESO and lead author of the study, told Space.com.

Hainaut used computer modelling to understand the effect of varying numbers of satellites of different brightness levels on astronomical observations. The modelling showed that if 100,000 satellites were to orbit the planet and all were barely visible to the naked eye, astronomy could cope. If those satellites were brighter, however, around magnitude 7 or below in astronomical terms, astronomical research would become more difficult and costly.

Satellites affect the sky in two ways. Firstly, the sunlight they reflect increases the overall brightness of the sky , creating light pollution. Second, brighter satellites also create streaks in telescope images that mar observations.

"If you increase the light pollution, it means that you will see fewer natural stars and you will see more of these satellites," said Hainaut . "For telescopes that means increasing exposure times. If you have a 10 percent increase in light pollution, you have to increase all the exposure times by 10 percent. It scales directly. For a 100 percent increase in light pollution, you have to increase all the exposures by 100 percent."

The exposure time increases mean that less science gets done and every observation becomes more expensive. The International Astronomical Union says that an increase in light pollution by more than 10 percent compared to natural dark sky conditions is an astronomy killer.

As light pollution has spread with urban development over the past two centuries, astronomers have increasingly been retreating into ever more remote locations. Many of the world's most expensive telescopes, including the Vera C. Rubin Observatory and the ESO's Very Large Telescope and Extremely Large Telescope , are located in Chile's Atacama Desert where the night sky is still nearly perfectly dark .

But while it is possible to retreat from the city lights, there will be no escape from satellite light pollution, Hainaut warns. You may be visiting a tiny village in Africa, camping in the Australian outback, or on an expedition to Antarctica or the Amazon rain forest, and your sky would still be brightened by the satellites.

This diagram shows the number of satellites that would be visible above ESO's Very Large Telescope (VLT) if SpaceX launches their planned constellation of 1 million satellites. (Image credit: ESO/O. Hainaut) "What they propose would make our observations close to impossible"

What is worse, if plans to launch thousands of sun-reflecting mirrors, as proposed by the U.S. company Reflect Orbital, were to come to fruition, the sky would transform completely.

Headquartered in Hawthorne, California, Reflect Orbital's vision is to deliver light on demand to solar power plants at night and to illuminate warzones and areas struck by natural disasters. The company has applied to the Federal Communications Commission (FCC) to launch a demonstration space mirror into orbit later this year.

The satellite, called Eärendil-1, is 59 by 59 feet (18 by 18 meters) in size, and should be the first in a constellation of 50,000, if things were to go according to Reflect Orbital's plan. "Reflect Orbital is really bad," said Hainaut.

"What they propose would make our observations close to impossible.
These are super bright satellites."

Astronomers have calculated that each Reflect Orbital space mirror would be brighter than the full moon if observed from the area where its beam is aiming. But the satellites would be visible regardless of where their beams are aiming to everybody around the world.

"Even outside the beam, the satellite will appear brighter than the planet Venus, which is the brightest object in the night sky after the moon," said Hainaut. "If they were to launch 50,000 of these space mirrors, there would be many hundred or even a few thousand of these super bright objects visible to observers anywhere on Earth."

An image illustrating how sunlight scattered by Reflect Orbital’s space mirrors would increase the overall brightness of the sky above ESO’s Very Large Telescope (VLT). (Image credit: ESO/O. Hainaut) Today, in areas with relatively low levels of light pollution, one can only see a few hundred bright stars in the sky. That means there would be more satellites than stars visible in the sky anywhere in the world with the full Reflect Orbital constellation in orbit. The constellation would also brighten the night sky by up to 300 percent, Hanuit calculated.

"If you increase the light pollution, it means that you will see fewer natural stars," said Hainaut. "And you will see more of these satellites."

Hainaut said that SpaceX's planned orbital data centers, despite the fact that they feature 230-foot-wide (70 m) solar panels, would be much dimmer and about as visible as Starlink satellites .

"From the available information, we see that these satellites have been optimized to minimize the impact as seen from the ground," Hainaut said. "The reflective surfaces are tilted away from Earth and the satellite itself is very narrow, pointing to Earth with its small end."

Still, overall, satellites operated by all operators around the world should remain below 100,000 satellites combined if astronomy is not to suffer, scientists caution. SpaceX is currently awaiting FCC's decision on its application to launch one million orbital data centers.

Currently, some 14,000 satellites orbit the planet.

NGC 457, the Owl Cluster in Cassiopeia, in a stack of images showing the total number of satellite trails recorded over 36 minutes of total exposure in October 2022. (Image credit: Alan Dyer/VW Pics/Universal Images Group via Getty Images) "For astronomy, this would obviously be catastrophic."

Robert Massey, the Deputy Executive Director of the Royal Astronomical Society, said Hainaut's findings were "not hugely surprising."

"For astronomy, this would obviously be catastrophic," Massey told Space.com. "It's very difficult to imagine how you could mitigate that on this scale. But I am also concerned about the public impact. The public has not signed up for having an entirely transformed sky."

Massey pointed out that based on the international law governing space activities, it is perfectly legal for a U.S. organization to be single-handedly deciding on something that would impact the entire world.

"If it's agreed by the FCC, this will be deeply regrettable," Massey said. "This will say that we are in a world where large corporations can determine the view of the sky above our heads, just as they can transform the environment on Earth. But the transformation of the environment on Earth is subject to pretty tight regulations."

The Trump administration has been taking steps to reduce the burden for satellite operators to prove their projects will have no negative environmental impacts. Currently, no environmental review has to be undertaken by the FCC or private companies before satellite applications go for approval.

Betty Kioko, an institutional affairs adviser at ESO, said the United Nations Outer Space Treaty , signed in 1967, states that the responsibility for space launches is with the nation states where those space objects are registered. She, however, added that the Treaty requires states to use space "for the common good of humankind."

"We now have to wait for the FCC to decide, because ultimately, the Outer Space Treaty was written at a time before we envisioned access to space by private entities."

ESO is among hundreds of organizations from around the world that have filed objections to the SpaceX and Reflect Orbital applications.

The study has been published in the journal Astronomy & Astrophysics.