SoylentNews

upstart writes:

GNOME 49 arrives this week, and it's packed with features and polish you'll love:

GNOME is a Linux desktop environment that you either love or hate. I've used GNOME and GNOME-based desktops for years and have always fallen on the side of "love. " With each new release, I always find a feature or two (or a bit of extra polish) that makes me smile.

The release of GNOME 49 is no different. Although there might not be any game-changing features for this release (I'm guessing the developers are holding out until the big 5-0), there's just enough to make it a worthy upgrade from 48.

GNOME 49 will officially be released on Sept. 17 and will hopefully make it to your distribution of choice soon after.

If you're curious as to what's coming, read on.

With the release of GNOME 49, the X11 session has been officially disabled. That doesn't mean your distribution will leave it disabled, as the GNOME team has made it possible for distro maintainers to enable X11 support.

Enjoy X11 while you can, because the team plans on stripping all X11 code from the desktop for the 50th release.

GNOME Shell is now a Wayland-only desktop environment. For those who use applications that have yet to add Wayland support, fret not, as Xwayland will continue to work, so those apps will still run.

upstart writes:

China Bans Nvidia's AI Chips:

Nvidia just can't seem to catch a break.

After beating a ban on China sales imposed by the Trump administration, the tech giant is now facing a ban on its products by Chinese regulators.

Chinese internet regulator Cyberspace Administration of China (CAC) ordered top Chinese tech companies like Bytedance and Alibaba to end their testing and orders of Nvidia's RTX Pro 6000D chips, the Financial Times reported on Wednesday citing people with knowledge of the matter.

Nvidia unveiled the RTX Pro 6000D chip, a lower tech chip designed just to be sold in China under compliance with American export control rules, earlier this year. The chips were initially designed to fill the void left by the then-banned H20 chips, another lower tech China-only chip. The H20s were recently reapproved for sale by the U.S. but orders have not yet begun shipping out. China is also advocating for the U.S. government to approve the sale of higher tech Blackwell chips to China.

The move comes as the latest breaking point of a fraying relationship between China and Nvidia.

Last month, Chinese authorities started questioning and cautioning industry titans like Tencent over their purchases of Nvidia's H20 chips, according to Reuters.

"We can only be in service of a market if a country wants us to be," Nvidia CEO Jensen Huang said at a press conference on Wednesday in response to a question about the China CAC. "I'm disappointed with what I see but they have larger agendas to work out between China and the United States. And I'm patient about it. We'll continue to be supportive of the Chinese government and Chinese companies as they wish."Nvidia has been caught up in the middle of a trade storm between the two countries.

The Biden administration was first to enforce export restrictions on Nvidia chips sold to China, in an effort to curb the entry of high-tech chips into China off of national security and competitive fears. The restrictions expanded for a while before they were relaxed again under Trump after Beijing landed a big blow to domestic AI confidence earlier this year with Deepseek's R1, an AI model that rivaled the best of American companies offerings using lower cost chips, inadvertently showing Americans that Chinese innovation did not require the top Nvidia chips.

The trade dispute has since gone beyond just chips to include discussions over TikTok and rare earth metals, of which China controls roughly 90% of the world's supply.

China's AI industry was largely dependent on American chipmaker Nvidia's chips but that might all be changing soon. After Trump's blanket ban earlier this year choked off the Chinese industry from access to Nvidia chips, Chinese chip development has ramped up. China chip stocks have experienced a major boom so big that the Beijing-based company Cambricon had to warn investors recently.

"The message is now loud and clear," an unnamed Chinese tech executive told the FT. "Earlier, people had hopes of renewed Nvidia supply if the geopolitical situation improves. Now it's all hands on deck to build the domestic system."

Although no chip, Chinese-made or otherwise, has been considered up to par with Nvidia's offerings, the Financial Times article paints a different picture.

After talking to top Chinese tech companies Huawei, Cambricon, Alibaba, and Baidu about their chip development, Beijing has concluded that Chinese AI processors are now comparable or even better than the downgraded Nvidia products that are allowed into China, according to the FT, citing a person with knowledge of the matter.

"The top-level consensus now is there's going to be enough domestic supply to meet demand without having to buy Nvidia chips," an industry insider told the FT.

The Chinese chip industry is having a good day. Alibaba shares rose after the company secured a prominent Chinese customer, state-owned telecommunications operator China Unicom, for its new AI chips. Baidu's shares in Hong Kong jumped by the most they have in more than three years after analysts voiced confidence in the search engine operator's chip venture.

Original Submission

upstart writes:

Modular nuclear reactors sound great, but won't be ready any time soon:

The UK government has announced plans to build more than a dozen small nuclear reactors across the country, ushering in what it calls a new "golden age" for nuclear power. One of the ultimate goals is to help the country to finally divest from Russian energy within three years – but do tiny nuclear reactors make engineering and commercial sense, and can they even be built?

Ahead of a 16 September London visit by US President Trump, the US and UK announced a partnership between British firm Centrica and US start-up X-Energy to build 12 small modular nuclear reactors to power data centres, plus a "micro modular nuclear power plant" at DP World's London Gateway port built by US start-up Last Energy.

However, no dates were given for the beginning of any of the projects, and the Department for Energy Security and Net Zero did not respond to New Scientist's request for more detail.

The announcement fits a trend of smaller nuclear reactors. Bruno Merk at the University of Liverpool in the UK says Rosatom, Russia's state nuclear energy organisation, recently finished building a batch of small reactors for a highly specific use in nuclear-powered icebreaker ships. Crucially, they then continued building more, showing either that there is demand from somewhere, or that Rosatom is taking a risk and building them as a commercial demonstration in the hope of selling more despite a raft of energy sanctions imposed after its invasion of Ukraine.

China, too, has built a Linglong One small nuclear reactor, but it is not clear whether it will yet be a commercially viable product. And giant technology firms like Amazon, Google and Microsoft are investing in these sorts of nuclear technologies, too.

David Dye at Imperial College London says tiny reactors make sense for remote military installations or Arctic sites, but is sceptical about using tiny nuclear reactors to power these tech giants' needs. He says it is far easier to build data centres near a ready supply of energy instead.

"If you're a tech visionary multibillionaire and you want to believe...and you've made your billion, what is it to chuck $50 million at this cool technology?" says Dye.
"This is very rich men, or clubs of very rich men, giving a few crumbs off the table to this technology they've always loved the idea of, without really looking too carefully."

One motivation could be oversight, says Michael Bluck at Imperial College London. "If you're a data centre, you need to be on 99.995 per cent of the time," says Bluck. "That means you really want to be in control of that electricity. You get first dibs on that electricity."

Bluck says there is no engineering or scientific reason we can't build tiny nuclear reactors, and build them fast. He points out the first experimental reactors were small, and many devices of a similar size operate in universities and military submarines around the world still.

"Size is not the issue. It's the modularity, it's the building it on a production line, it's the standardisation of components. It's really practical. It's standard engineering," says Bluck.

upstart writes:

'Beyond EUV' chipmaking tech pushes Soft X-Ray lithography closer to challenging Hyper-NA EUV:

Researchers at Johns Hopkins University have unveiled a new approach to chipmaking that uses lasers with a 6.5nm ~ 6.7nm wavelength — also known as Soft X-rays — that could increase the resolution of lithography tools to 5nm and below, reports Cosmos, citing a paper published in Nature.

The scientists call their method 'beyond-EUV' — suggesting that their technology could replace industry-standard EUV lithography — but the researchers admit they are currently years away from building even an experimental B-EUV tool.

Soft X-Rays can challenge Hyper-NA. On paper

The most advanced chips nowadays are made using EUV lithography, which operates at a wavelength of 13.5 nm and can produce features as small as 13nm (Low-NA EUV of 0.33 numerical aperture), 8nm (High-NA EUV of 0.55 NA), or even 4nm ~ 5nm (Hyper-NA EUV on 0.7 – 0.75 NA) at the cost of extreme complexity of the lithography systems that have very advanced optics that cost hundreds of millions of dollars.

By using a shorter wavelength, researchers from Johns Hopkins University can get an intrinsic resolution boost even with lenses with moderate NA. However, they face many challenges with B-EUV.

Firstly, B‑EUV light sources are not yet ready. Various researchers have tried multiple methods of generating 6.7 nm wavelength radiation (e.g., gadolinium laser-produced plasma), but there is no industry-standard approach. Secondly, these shorter wavelengths — due to their high photon energy — interact poorly with traditional photoresist materials used in chipmaking. Thirdly, because 6.5nm ~ 6.7nm wavelength light is absorbed rather than reflected by pretty much everything, multilayer-coated mirrors for this type of radiation haven't been produced before.

Finally, these lithography tools must be designed from scratch, and currently, there is no ecosystem to support the designs with components and consumables. To sum up, building a B-EUV machine (or Soft X-ray machine?) requires breakthroughs in light sources, projection mirrors, resists, and even consumables like pellicles or photomasks.

The team discovered that metals like zinc are able to absorb B-EUV light and emit electrons, which then trigger chemical reactions in organic compounds called imidazoles. These reactions make it possible to etch very fine patterns onto semiconductor wafers.

Interestingly, while zinc performs poorly with traditional 13.5nm EUV light, it becomes highly effective at shorter wavelengths, highlighting how important it is to match the material with the right wavelength.

To apply these metal–organic compounds to silicon wafers, the researchers developed a technique called chemical liquid deposition (CLD). This method creates thin, mirror-like layers of a material called aZIF (amorphous zeolitic imidazolate frameworks), growing at a rate of 1nm per second. CLD also allows for fast testing of different metal–imidazole combinations, making it easier to discover the best pairings for different lithography wavelengths. While zinc is well suited for B-EUV, the team noted that other metals might perform better at different wavelengths, offering flexibility for future chipmaking technologies.

This approach gives manufacturers a toolbox of at least 10 metal elements and hundreds of organic ligands to create custom resists tailored to specific lithography platforms, the researchers disclosed.

Although the researchers did not solve the full stack of B-EUV challenges (e.g., source power, masks), they advanced one of the most critical bottlenecks: finding resist materials that can work with 6nm wavelength light. They created the CLD process to apply thin, uniform films of amorphous zeolitic imidazolate frameworks (aZIFs) onto silicon wafers. They experimentally showed that certain metals (like zinc) can absorb Soft X-ray light and emit electrons that trigger chemical reactions in imidazole-based resists.

There are plenty of challenges to be solved with B-EUV, and the technology doesn't have a clear path to the mass market. However, the CLD process can be used pretty widely, both in semiconductor and non-semiconductor applications.

Journal: Spin-on deposition of amorphous zeolitic imidazolate framework films for lithography applications by Yurun Miao, Shunyi Zheng, Kayley E. Waltz, Mueed Ahmad, Xinpei Zhou, Yegui Zhou, Heting Wang, et al.

Original Submission

upstart writes:

"We're excited that Northrop is ready to deliver this incredibly beneficial increase in capacity" :

What happens when you use a SpaceX Falcon 9 rocket to launch Northrop Grumman's Cygnus supply ship? A record-setting resupply mission to the International Space Station.

The first flight of Northrop's upgraded Cygnus spacecraft, called Cygnus XL, is on its way to the international research lab after launching Sunday evening from Cape Canaveral Space Force Station, Florida. This mission, known as NG-23, is set to arrive at the ISS early Wednesday with 10,827 pounds (4,911 kilograms) of cargo to sustain the lab and its seven-person crew.

By a sizable margin, this is the heaviest cargo load transported to the ISS by a commercial resupply mission. NASA astronaut Jonny Kim will use the space station's Canadian-built robotic arm to capture the cargo ship on Wednesday, then place it on an attachment port for crew members to open hatches and start unpacking the goodies inside.

The Cygnus XL spacecraft looks a lot like Northrop's previous missions to the station. It has a service module manufactured at the company's factory in Northern Virginia. This segment of the spacecraft provides power, propulsion, and other necessities to keep Cygnus operating in orbit.

The most prominent features of the Cygnus cargo freighter are its circular, fan-like solar arrays and an aluminum cylinder called the pressurized cargo module that bears some resemblance to a keg of beer. This is the element that distinguishes the Cygnus XL from earlier versions of the Cygnus supply ship.

The cargo module is 5.2 feet (1.6 meters) longer on the Cygnus XL. The full spacecraft is roughly the size of two Apollo command modules, according to Ryan Tintner, vice president of civil space systems at Northrop Grumman. Put another way, the volume of the cargo section is equivalent to two-and-a-half minivans.

"The most notable thing on this mission is we are debuting the Cygnus XL configuration of the spacecraft," Tintner said. "It's got 33 percent more capacity than the prior Cygnus spacecraft had. Obviously, more may sound like better, but it's really critical because we can deliver significantly more science, as well as we're able to deliver a lot more cargo per launch, really trying to drive down the cost per kilogram to NASA."

China Rules That Nvidia Violated its Antitrust Laws

upstart writes:

China rules that Nvidia violated its antitrust laws:

A Chinese regulator has found Nvidia violated the country's antitrust law, in a preliminary finding against the world's most valuable chipmaker.

Nvidia had failed to fully comply with provisions outlined when it acquired Mellanox Technologies, an Israeli-US supplier of networking products, China's State Administration for Market Regulation (SAMR) said on Monday. Beijing conditionally approved the US chipmaker's acquisition of Mellanox in 2020.

Monday's statement came as US and Chinese officials prepared for more talks in Madrid over trade, with a tariff truce between the world's two largest economies set to expire in November.

SAMR reached its conclusion weeks before Monday's announcement, according to two people with knowledge of the matter, adding that the regulator had released the statement now to give China greater leverage in the trade talks.

The regulator started the anti-monopoly investigation in December, a week after the US unveiled tougher export controls on advanced high-bandwidth memory chips and chipmaking equipment to the country.

[...] The preliminary findings against the chipmaker could result in fines of between 1 percent and 10 percent of the company's previous year's sales. Regulators can also force the company to change business practices that are considered in violation of antitrust laws.

Over recent years, Nvidia has become a global market leader in artificial intelligence chips, with its graphics processing units becoming crucial in developing leading AI models.

That has also meant that Nvidia has increasingly been caught up in the trade tensions between Washington and Beijing.

[...] Nvidia chief Jensen Huang, who has made frequent visits to China in a signal of his commitment to a crucial overseas market, has previously criticized the US curbs as a "failure" that has spurred Chinese rivals to accelerate development of their own products.

upstart writes:

There's no such thing as free laundry:

An unknown hacker has broken into smart washing machines that accept digital payments, leaving over a thousand students without laundry service in Amsterdam. According to Dutch publication Folia, the hacker(s) disabled the payment system on the appliances, so students based in the Spinozacampus housing complex could get their clothes cleaned for free. However, this did not last long, as Duwo, the management company behind the service, didn't want to get stuck sorting the stains of the unpaid laundry bills.

"Because we purchase the machines ourselves, we need the income to be able to continue offering laundry services to our residents at affordable prices," its spokesperson told Folia. While it might seem like a small amount for a company to shoulder the cost of laundry, it could soon add up if it's going to shoulder the cost of providing clean clothes to Spinozacampus' 1,250 residents.

The hack was first discovered in mid-July, but it wasn't until recently that the company disabled the machines. Although there are still 10 other analog washing machines students can use, Folia reports that these are almost always out of order, with one student claiming that only one machine works for all the students. This has even got them worried about the risk of a lice outbreak because of the limited availability of laundry machines, digital or otherwise. Thankfully, there's another residence building a little over 200 meters or 650 feet away that has more washing machines, allowing for shorter queues.

In the meantime, Duwo is slowly switching back to non-digital appliances, with the company expecting to receive five analog washing machines in a few days. It's also reported that other buildings and housing associations are moving away from IoT washing machines.

As for the hacker, they could face up to a year in prison if they are caught, with the sentence going up to six years if it is proven that they did it for monetary gain. Nevertheless, ethical hacker Sijmen Ruwhof told Folia that finding the culprit is costly and time-consuming, so it might not be worth it for Duwo to follow up on the case. And although Ruwhof suspects a professional hacker is behind the washing machine attack, he also conceded that there are a lot of bright students on campus who would be capable of executing the breach.

"There are lots of bright minds on that campus who also know how to program. It gives you a huge kick when you hack into a washing machine like that," Sijmen said. He also added, "If I were a student and saw those digital washing machines, as a hacker, I would be getting the itch, too."

Original Submission

upstart writes:

A new stealthy radio device could help give people the edge, letting them fly drones without detection:

A stealthy military radio that hides communications in background noise is extremely difficult to jam or locate, meaning that it could allow drone pilots to operate without detection.

Electronic warfare has entered an intense new phase as drones increasingly dominate the battlefield. In the ongoing war between Russia and Ukraine, both sides use jammers to block drone-control signals. They also trace radio signals to target enemy drone operators with artillery strikes.

Now, US-based start-up Rampart Communications has designed a radio with two levels of protection that make the signal extremely hard to detect. Its StrataWave radio encrypts the signal and spreads it across the radio spectrum, rather than broadcasting on a single frequency, making radio emissions quieter and harder to detect.

Similar techniques have been used before, but StrataWave goes an extra step. While spreading the signal across the radio spectrum makes it harder to intercept, it doesn't hide the fact that a radio broadcast is taking place. To do that, StrataWave scrambles the entire broadcast to hide the very presence of a radio signal in background noise.

The first level of protection is like writing a letter in code and tearing it into large pieces – even if an adversary can't read your letter, they can at least see you have written one. The second level is more like grinding the letter to dust.

"Without the correct encryption key and algorithm, the signal will appear as noise to any other receiver," says Aaron Correa at Rampart.

[...] Electronic warfare is a game of cat and mouse, with every development met by a new counter. In Ukraine, drones are updated every few weeks to stay ahead of jammers. Rampart says adversaries will effectively be starting from scratch when trying to detect or jam StrataWave.

Thomas Withington, an electronic warfare expert at the Royal United Services Institute (RUSI), a defence think tank in the UK, says this is unlikely to be the final move in the game of radios versus jammers. "Radio frequency engineers will tell you that every new system works brilliantly – until it doesn't," he says.

Withington notes that cognitive radio systems using AI and large amounts of data are getting ever better at finding hidden signals in noise. But it may take a while to crack StrataWave. "This type of system will certainly give you a temporary advantage, and that may be all you need," he says.

Original Submission

canopic jug writes:

Famed developer Poul-Henning Kamp (phk) has posted an update on the status and future of the project currently known as Varnish Cache. And, after 20 years of being a go-to component in WWW infrastructure, it will change its name to The Vinyl Cache Project with version 8.0.0 being the last under the old name. The software project will be sheperded under the new name by a Danish association formed for that specific purpose.

We will instead form a voluntary association, a "Forening", under the laws of Denmark, with bylaws that set out what the goal is (develop, maintain and distribute the software), who gets to make the decisions (a governing board appointed by the members), who can become members (anybody but subject to approval by the members) and that the association cannot ever hold or handle any money.

The commented bylaws of the association will be ratified by the founders and made public this autumn, and the first general assembly will be on Monday February 23rd 2026 - hopefully with many membership applications to approve - more about that when we publish the bylaws.

We will also, at the same time, reluctantly change the name of the project.

Varnish Cache is a very fast web application accelerator, or caching HTTP reverse proxy. It runs in front of web servers and the output of the server is cached there, subject to specific caching criteria, to save overloading the back end.

Previously:
(2022) The New Yorker on NTP Software Maintenance
(2022) PHK on Surveillance Which Is Too Cheap to Meter
(2018) Transparency Versus Liability in Hardware

Original Submission

upstart writes:

FAA found factory violations, says Boeing sought approval for unairworthy planes:

The Federal Aviation Administration on Friday proposed fines of $3.1 million against Boeing for various safety violations related to the January 2024 door plug blowout and what the FAA called "interference with safety officials' independence."

An FAA statement said the proposed fine covers "safety violations that occurred from September 2023 through February 2024," and is the "maximum statutory civil penalty authority consistent with law." Boeing, which reported $22.7 billion in revenue and a net loss of $612 million last quarter, has 30 days to file a response with the agency.

"The FAA identified hundreds of quality system violations at Boeing's 737 factory in Renton, Washington, and Boeing subcontractor Spirit AeroSystems' 737 factory in Wichita, Kansas. Additionally, Boeing presented two unairworthy aircraft to the FAA for airworthiness certificates and failed to follow its quality system," the FAA said.

The FAA said that a Boeing safety official faced pressure to sign off on an aircraft that did not meet standards. The employee is part of the Boeing ODA [Organization Designation Authorization] unit that performs functions the FAA delegated to the company.

The FAA said it "found that a non-ODA Boeing employee pressured a Boeing ODA unit member to sign off on a Boeing 737-MAX airplane so Boeing could meet its delivery schedule, even though the ODA member determined the aircraft did not comply with applicable standards." Boeing's ODA process has faced criticism for years. A 2021 Inspector General report found that "the Boeing ODA process and structure do not ensure ODA personnel are adequately independent."

[...] Separately, Boeing in July 2024 agreed to plead guilty to a criminal charge of defrauding the FAA and pay a $243.6 million fine for violating a 2021 deferred prosecution agreement with the government. However, the Boeing plea deal was rejected by a federal judge in December 2024.

upstart writes:

When it comes to US AI rules, there's too many cooks:

The US government wants AI in every corner of government, but the unstoppable force of new tech is running into the immovable object of bureaucracy - a growing mass of AI rules.

It's been well established in the first year of Trump's second presidency that AI is a priority for the administration. Even prior to Trump taking office, government generative AI use cases had surged, growing ninefold between 2023 and 2024. In recent months, agencies have cut numerous deals with most leading AI companies under the General Services Administration's Trump-driven OneGov contracting strategy. These agreements give federal agencies access to leading AI models for $1 or less per agency for the first year, suggesting that the Trump team is keen on acting fast.

But given the nature of government, it's not surprising to hear from the auditors at the Government Accountability Office (GAO) that agencies face a large, fragmented set of AI requirements.

According to a report published on Tuesday, the GAO identified 94 separate "AI-related government-wide requirements" that agencies have to adhere to - and those rules aren't centralized under a single management body.

AI rules and requirements come from ten separate executive-branch oversight and advisory groups, including the Office of Management and Budget, Office of Science and Technology Policy, Department of Commerce, GSA, and National Science Foundation. Those groups help set and police requirements drawn from five AI-related laws, six executive orders, and three guidance documents, the GAO said.

In short, there are a lot of rules and regulations surrounding federal government AI use to account for, making for a tricky - and possibly shifting - deployment path for agencies to navigate.

The GAO declined to take a stance on whether there were too many AI regulations for federal agencies to account for, or whether a central AI regulation for federal use is necessary to streamline operations, but its own prior work suggests that there's a rather severe AI regulatory burden placed on agencies.

It noted that agencies were struggling with GenAI deployment. Rather than chalk it up to chasing use cases that don't exist, those agencies pointed to familiar hurdles - a lack of computing resources, concerns over bias and hallucinations, and - you guessed it - too many rules.

Of the 12 agencies the GAO spoke with about AI implementation for that report, 10 said that existing federal AI policy either didn't account for all the obstacles an agency could face when implementing AI or, conversely, that federal AI policy "could present obstacles to the adoption of generative AI."

"AI is rapidly growing and holds substantial promise for improving the government," GAO's director of IT and cybersecurity Kevin Walsh told The Register in an email. "But AI technologies also pose risks."

Walsh told us that AI can substantially improve operations at federal agencies, but can also be misused to enable cyberattacks, commit fraud, and deanonymize data.

"The rules that govern AI will be critical in our attempts to ensure AI is used for good," Walsh told us while declining to take a position on whether the current fragmented, cumbersome AI regulatory regime was an appropriate one.

That said, the GAO has been making recommendations to improve federal AI oversight as far back as four years ago, when it first published a "framework to help managers ensure accountability and the responsible use of AI in government programs and processes," according to its latest report. In 2023, it issued a second report on how federal agencies were complying with AI rules, but few of the compliance recommendations it made have been acted upon as of yet.

"We made 35 recommendations to 19 agencies ... to fully implement federal AI requirements," the GAO said. "As of July 2025, three agencies had implemented four recommendations."

Original Submission

upstart writes:

Google cut managers by 35%: Inside Pichai's layoffs overhaul:

Google has cut 35% of its managers, focusing on those leading teams with fewer than three people. The move, announced during an all-hands meeting on August 27, 2025, has jolted workers across the globe. The recent Google management layoffs is part of CEO Sundar Pichai's bold move that focuses on efficiency, reshaping the tech darling's hierarchy amid ongoing restructuring plans. Pichai's ongoing Google layoffs not only reshape the company's structure but also push for leaner operations.

What's interesting is that the Google job cuts will once again help the giant double down on its AI and cost efficiency moves.

The recent Google layoffs target roles seen as unnecessary, particularly managers overseeing small teams. Brian Welle, Google's VP of People Analytics and Performance, shared the details: "We now have 35% fewer managers, with fewer direct reports than a year ago."

Welle added that Google aims to reduce its leadership ranks, i.e. managers, directors, and vice presidents, to a smaller share of the workforce over time. So, why did Google fire managers?

The ongoing layoffs at Google won't just cut managers' roles. Many affected managers have now been shifted to individual contributor roles, thereby retaining their expertise within the company.

Pichai has been clear about the reasoning behind these Google layoffs in 2025. "We need to be more efficient as we grow, so we don't just throw more people at every problem," he said during the meeting. The CEO's approach marks a significant shift from Google's past, where rapid hiring fueled growth.

These Google layoffs build on earlier job cuts. This includes the 6% workforce reductions at Google in 2023, and targeted layoffs in teams like Android and Pixel. With a nod to rival Meta's policies, Pichai jokingly remarked, "Maybe I should try running the company with all of Meta's policies," but clarified that Google's existing leave options are sufficient.

To soften the blow in the aftermath of Google layoffs, the giant has introduced a Voluntary Exist Program (VEP) in January 2025 for U.S. employees in areas like search, marketing, hardware, and people operations. Fiona Cicconi, Google's chief people office, called the VEP a success. "It's been quite effective," with 3% to 5% of eligible employees taking the offer, often for personal reasons like family or breaks from work.

Pichai praised the program's flexibility, "I'm glad it's worked out well, it gives people agency."

News is breaking that Robert Redford has died at his home age 89.

upstart writes:

Real-Time Observation of Magnet Switching in a Single Atom:

Nuclear spins stay magnetically stable because they're great at ignoring their surroundings. But to read or change their state, they need just a little interaction with the outside world. That's why knowing and controlling their atomic neighborhood is crucial for quantum tech.

Until now, we could read single nuclear spins, but their environments were a mystery. Enter Scanning tunneling microscopy (STM) + electron spin resonance (ESR): a powerful duo that lets scientists zoom in and listen to nuclear spins at the atomic level, thanks to hyperfine interactions.

In a breakthrough from Delft University, scientists used an STM to spy on a single titanium atom's nuclear spin, like catching its magnetic heartbeat in real time. By tapping into the atom's electrons, they watched the spin flip back and forth, live.

The twist? That tiny spin stayed stable for several seconds, an eternity in quantum terms. This opens the door to better control over atomic-scale bits, nudging us closer to ultra-precise quantum technologies.

A new way to control atomic nuclei as "qubits"

A scanning tunneling microscope (STM) is like a super-sharp needle that can "feel" individual atoms on a surface and create images with incredible detail. But what it actually senses are the electrons swirling around the atom's nucleus.

Both electrons and the nucleus act like tiny magnets, each with a property called spin. Scientists figured out how to detect the spin of a single electron using an STM about ten years ago.

Now, a team at TU Delft, led by Professor Sander Otte, asked a bold question: Can we go deeper and read the spin of the nucleus itself, in real time?

Otte explains, "The general idea had been demonstrated a few years ago, making use of the so-called hyperfine interaction between electron and nuclear spins. However, these early measurements were too slow to capture the motion of the nuclear spin over time."

Evert Stolte, first author of the study, said, "We were able to show that this switching corresponds to the nuclear spin flipping from one quantum state to another, and back again."

They found that the nuclear spin in the atom stays stable for about five seconds before flipping, much longer than most quantum systems. In comparison, the electron spin in the same atom lasts only about 100 nanoseconds, which is millions of times shorter.

Because the researchers could measure the nuclear spin faster than it changed, and without disturbing it, they achieved what's called single-shot readout. This means they could catch the spin's state in one go, like snapping a photo before it moves.

This breakthrough makes it possible to control nuclear spins more precisely, opening up new experiments. In the long run, it could help build powerful tools for quantum simulation and atomic-scale sensing.

Journal Reference:
Stolte, Evert W., Lee, Jinwon, Vennema, Hester G., et al. Single-shot readout of the nuclear spin of an on-surface atom [open], Nature Communications (DOI: 10.1038/s41467-025-63232-5)

Original Submission

upstart writes:

Pentagon begins deploying new satellite network to link sensors with shooters:

The first 21 satellites in a constellation that could become a cornerstone for the Pentagon's Golden Dome missile defense shield successfully launched from California Wednesday aboard a SpaceX Falcon 9 rocket.

The Falcon 9 took off from Vandenberg Space Force Base, California, at 7:12 am PDT (10:12 am EDT; 14:12 UTC) and headed south over the Pacific Ocean, heading for an orbit over the poles before releasing the 21 military-owned satellites to begin several weeks of activations and checkouts.

These 21 satellites will boost themselves to a final orbit at an altitude of roughly 600 miles (1,000 kilometers). The Pentagon plans to launch 133 more satellites over the next nine months to complete the build-out of the Space Development Agency's first-generation, or Tranche 1, constellation of missile tracking and data relay satellites.

"We had a great launch today for the Space Development Agency, putting this array of space vehicles into orbit in support of their revolutionary new architecture," said Col. Ryan Hiserote, system program director for the Space Force's assured access to space launch execution division.

Military officials have worked for six years to reach this moment. The Space Development Agency (SDA) was established during the first Trump administration, which made plans for an initial set of demonstration satellites that launched a couple of years ago. In 2022, the Pentagon awarded contracts for the first 154 operational spacecraft. The first batch of 21 data relay satellites built by Colorado-based York Space Systems is what went up Wednesday.

"Back in 2019, when the SDA was stood up, it was to do two things. One was to make sure that we can do beyond line of sight targeting, and the other was to pace the threat, the emerging threat, in the missile warning and missile tracking domain. That's what the focus has been," said GP Sandhoo, the SDA's acting director.

Historically, the military communications and missile warning networks have used a handful of large, expensive satellites in geosynchronous orbit some 22,000 miles (36,000 kilometers) above the Earth. This architecture was devised during the Cold War, and is optimized for nuclear conflict and intercontinental ballistic missiles.

For example, the military's ultra-hardened Advanced Extremely High Frequency satellites in geosynchronous orbit are designed to operate through an electromagnetic pulse and nuclear scintillation. The Space Force's missile warning satellites are also in geosynchronous orbit, with infrared sensors tuned to detect the heat plume of a missile launch.

The problem? Those satellites cost more than $1 billion a pop. They're also vulnerable to attack from a foreign adversary. Pentagon officials say the SDA's satellite constellation, officially called the Proliferated Warfighter Space Architecture, is tailored to detect and track more modern threats, such as smaller missiles and hypersonic weapons carrying conventional warheads. It's easier for these missiles to evade the eyes of older early warning satellites.

What's more, the SDA's fleet in low-Earth orbit will have numerous satellites. Losing one or several satellites to an attack would not degrade the constellation's overall capability. The SDA's new relay satellites cost between $14 and $15 million each, according to Sandhoo. The total cost of the first tranche of 154 operational satellites totals approximately $3.1 billion.

These satellites will not only detect and track ballistic and hypersonic missile launches. They will also transmit signals between US forces using an existing encrypted tactical data link network known as Link 16. This UHF system is used by NATO and other US allies to allow military aircraft, ships, and land forces to share tactical information through text messages, pictures, data, and voice communication in near real-time, according to the SDA's website.

Up to now, Link 16 radios were ubiquitous on fighter jets, helicopters, naval vessels, and missile batteries. But they had a severe limitation. Link 16 was only able to close a radio link with a clear line of sight. The Space Development Agency's satellites will change that, providing direct-to-weapon connectivity from sensors to shooters on Earth's surface, in the air, and in space.

The relay satellites, which the SDA calls the transport layer, are also equipped with Ka-band and laser communication terminals for higher bandwidth connectivity.