An anonymous reader quotes a report from Jalopnik: Experimental submarine the Titan sank in June 2023 while exploring the wreck of the Titanic. The controversial craft imploded while deep beneath the surface of the ocean killing five people onboard, and now a recording of the Titan's final moments has been shared by the National Oceanic and Atmospheric Administration. [...] In the clip, which is available to hear [here], the static sound of the ocean is shattered by a great rumble, which sounds almost like a wave crashing against the beach.

It's this noise that is thought to be the total failure of the Titan, as LBC adds: "It is believed that the noise is the 'acoustic signature' of the sub imploding on 18th June 2023. It was recorded by the National Oceanic and Atmospheric Administration device about 900 miles from where the sub was last seen on radar, south of Newfoundland, Canada, US Coast Guard officials announced. The five crew members who died onboard the sub were British explorer sub were Hamish Harding, 58, British-Pakistani businessman Shahzada Dawood, 48, and his son Suleman, 19, French deep-sea explorer Paul-Henri Nargeolet (known as 'Mr Titanic'), 77, and and co-founder of the submarines owner's company OceanGate, Stockton Rushton, 61."

Longtime Slashdot reader JoeRobe writes: Scientists associated with the Kilometer Cube Neutrino Telescope, or KM3NeT, have reported detection of an ultra-high energy neutrino deep in the Mediterranean sea. The neutrino reportedly had an energy of 120 million billion electron volts (1.2x10^17 eV, or 120 PeV). This is similar to the energy of ping-pong ball traveling ~5 m/s, but all that energy was packed into a single subatomic particle. According to the New York Times, "Here, squeezed into one of the tiniest flecks of matter in our universe, that energy amounted to tens of thousands of times more than what can be achieved by the world's premier particle accelerator, the Large Hadron Collider at CERN."

According to the authors of the Nature paper, the direction of the neutrino "is compatible with the extension of the galactic interstellar medium," but they did not find any catalogued source that would produce such a high energy neutrino, within the Milky Way or from about 40 other galaxies that could be candidates.

Phys.org describes the impressive scale of the KM3NeT detector array: "It is located at 3,450 m depth, about 80 km from the coast of Portopalo di Capo Passero, Sicily. Its 700 m high detection units (DUs) are anchored to the seabed and positioned about 100 m apart. Every DU is equipped with 18 Digital Optical Modules (DOM) each containing 31 photomultipliers (PMTs). In its final configuration, ARCA will comprise 230 DUs. The data collected are transmitted via a submarine cable to the shore station at the INFN Laboratori Nazionali del Sud. The KM3NeT/ORCA (Oscillation Research with Cosmics in the Abyss) detector is optimized to study the fundamental properties of the neutrino itself. It is located at a depth of 2,450 m, about 40 km from the coast of Toulon, France. It will comprise 115 DUs, each 200 m high and spaced by 20 m. The data collected by ORCA are sent to the shore station at La Seyne Sur Mer." "This ultra-high energy neutrino may originate directly from a powerful cosmic accelerator," surmises Phys.org. "Alternatively, it could be the first detection of a cosmogenic neutrino. However, based on this single neutrino it is difficult to conclude its origin."

Apple has released an official Apple TV app for Android phones and tablets that's now available in the Google Play Store. You can download it here. 9to5Google reports: The newest Apple app on Android has a bottom bar with Apple TV+, MLS (Major League Soccer), Downloads for offline viewing, and Search. [...] The video player takes after Apple TV on other platforms, with a portrait mode available. There are convenient shortcuts to activate picture-in-picture, which works inside the app (while browsing) and system-wide, and mute to bring up the system volume bar. Playback is smooth and more stable than other streaming services.

At launch, the Apple TV app lacks Casting support and there do not appear to be new episode notifications. If you're already signed into Apple Music, you have to log in again to Apple TV. Another notable aspect is support for Google Play Billing instead of requiring out-of-app sign-up on another device. This applies to both the Google TV app (and Apple Music) today.

In August 1990, two hikers in Scotland captured photographs of a mysterious diamond-shaped aircraft accompanied by a Harrier jet, but the images and story were suppressed by the Ministry of Defence (MoD) for decades. Was it a prank, a hoax, an optical illusion or something else entirely? The Guardian's Daniel Lavelle reports on "what really happened in Calvine." Here's an excerpt: On a misty evening in August 1990, two men hiking on the moors surrounding Calvine, a pretty hamlet in Perth and Kinross, claimed to have seen a giant diamond-shaped aircraft flying above them. It apparently had no clear means of propulsion and left no smoke plume; it was silent and static, as if frozen in time. Terrified, they hit the ground and scrambled for cover behind a tree. Then a Harrier fighter jet roared into view, circling the diamond as if sizing it up for a scuffle. One of the men snapped a series of photographs just before the bizarre craft shot away vertically and disappeared.

Craig Lindsay was a press officer at the RAF base in Pitreavie Castle in Dunfermline, 50 miles away, when the Daily Record got in touch a few days later. The hikers, who worked as chefs at Fisher's Hotel in Pitlochry, had sent six photos of the diamond to the newspaper and told their story. The Record's picture editor, Andy Allen, sent Lindsay the best of the bunch. Lindsay had never seen such a clear photograph of a supposed UFO, so he forwarded the picture to the Ministry of Defence (MoD), which told him to ask the Record to send the other five photographs and their negatives. The MoD also instructed him to phone the hikers, which he did. One of them told Lindsay the whole story: the diamond, the jet, how it levitated eerily with no sound and accelerated with no obvious propellant. The MoD told Lindsay to leave the case with them. He pushed the diamond to the back of his mind.

That autumn, Lindsay attended a routine meeting in London. On his lunch break, he went for a wander around the MoD's offices and saw something familiar. "There, on the wall in front of me, was a great big poster-size print of the best of them [the photographs]. So, I spoke to the guys that were there and I asked them what their other photographs were like." The ministry's staff placed the other photographs on a windowsill. The snaps showed the Harrier jet moving from the right side of the frame to the left, while the diamond didn't move an inch. He quizzed some of the specialists who had investigated the photos. They told him there was no evidence of a hoax, but they didn't know what the diamond was. "I gradually forgot all about the thing," says Lindsay. "Nothing had appeared from the first inquiry ... I assumed that everything had just been forgotten." The Record didn't run the story, the hikers never spoke publicly about the photos and the images weren't seen by the public for 32 years. "It is the 35th anniversary of what has been described as the best UFO photo ever taken. Now is the time to come forward and tell us what really happened," says Prof David Clarke, a lecturer at Sheffield Hallam University who worked as a reporter in the 1990s.

An anonymous reader quotes a report from Ars Technica: On Monday, researcher Johann Rehberger demonstrated a new way to override prompt injection defenses Google developers have built into Gemini -- specifically, defenses that restrict the invocation of Google Workspace or other sensitive tools when processing untrusted data, such as incoming emails or shared documents. The result of Rehberger's attack is the permanent planting of long-term memories that will be present in all future sessions, opening the potential for the chatbot to act on false information or instructions in perpetuity. [...] The hack Rehberger presented on Monday combines some of these same elements to plant false memories in Gemini Advanced, a premium version of the Google chatbot available through a paid subscription. The researcher described the flow of the new attack as:

1. A user uploads and asks Gemini to summarize a document (this document could come from anywhere and has to be considered untrusted).
2. The document contains hidden instructions that manipulate the summarization process.
3. The summary that Gemini creates includes a covert request to save specific user data if the user responds with certain trigger words (e.g., "yes," "sure," or "no").
4. If the user replies with the trigger word, Gemini is tricked, and it saves the attacker's chosen information to long-term memory.

As the following video shows, Gemini took the bait and now permanently "remembers" the user being a 102-year-old flat earther who believes they inhabit the dystopic simulated world portrayed in The Matrix. Based on lessons learned previously, developers had already trained Gemini to resist indirect prompts instructing it to make changes to an account's long-term memories without explicit directions from the user. By introducing a condition to the instruction that it be performed only after the user says or does some variable X, which they were likely to take anyway, Rehberger easily cleared that safety barrier. Google responded in a statement to Ars: "In this instance, the probability was low because it relied on phishing or otherwise tricking the user into summarizing a malicious document and then invoking the material injected by the attacker. The impact was low because the Gemini memory functionality has limited impact on a user session. As this was not a scalable, specific vector of abuse, we ended up at Low/Low. As always, we appreciate the researcher reaching out to us and reporting this issue."

Rehberger noted that Gemini notifies users of new long-term memory entries, allowing them to detect and remove unauthorized additions. Though, he still questioned Google's assessment, writing: "Memory corruption in computers is pretty bad, and I think the same applies here to LLMs apps. Like the AI might not show a user certain info or not talk about certain things or feed the user misinformation, etc. The good thing is that the memory updates don't happen entirely silently -- the user at least sees a message about it (although many might ignore)."

Today, the KDE Project announced the release of KDE Plasma 6.3, featuring improved fractional scaling, enhanced Night Light color accuracy, better CPU usage monitoring, and various UI and security refinements.

Some of the key features of Plasma 6.3 include:
- Improved fractional scaling with KWin to lead to an all-around better desktop experience with fractional scaling as well as when making use of KWin's zoom effect.
- Screen colors are more accurate with the KDE Night Light feature.
- CPU usage monitoring within the KDE System Monitor is now more accurate and consuming fewer CPU resources.
- KDE will now present a notification when the kernel terminated an app because the system ran out of memory.
- Various improvements to the Discover app, including a security enhancement around sandboxed apps.
- The drawing tablet area of KDE System Settings has been overhauled with new features and refinements.
- Many other enhancements and fixes throughout KDE Plasma 6.3.

You can read the announcement here.

If there was any doubt before, this seals it: YouTube is in the TV business. According to Neal Mohan, YouTube's CEO, TV screens have officially overtaken mobile as the "primary device for YouTube viewing in the U.S." In other words, more people are watching YouTube on TV sets than any other device, at least here in the U.S. From a report: It is, as Mohan writes in his annual letter from the CEO, an indication that "YouTube is the new television."

"But the 'new' television doesn't look like the 'old' television," Mohan writes. "It's interactive and includes things like Shorts (yes, people watch them on TVs), podcasts, and live streams, right alongside the sports, sitcoms and talk shows people already love."

"In-dash advertising is here and Stellantis, the parent company of Jeep, Dodge, Chrysler, and Ram, beat everyone to further enshittification," writes longtime Slashdot reader sinij. "Ads can be seen in this video." From a report: In a move that has left drivers both frustrated and bewildered, Stellantis has introduced full-screen pop-up ads on its infotainment systems. Specifically, Jeep owners have reported being bombarded with advertisements for Mopar's extended warranty service. The kicker? These ads appear every time the vehicle comes to a stop. Imagine pulling up to a red light, checking your GPS for directions, and suddenly, the entire screen is hijacked by an ad. That's the reality for some Stellantis owners. Instead of seamless functionality, drivers are now forced to manually close out of ads just to access basic vehicle functions.

One Jeep 4xe owner recently shared their frustration on an online forum, detailing how these pop-ups disrupt the driving experience. Stellantis, responding through their "JeepCares" representative, confirmed that these ads are part of the contractual agreement with SiriusXM and suggested that users simply tap the "X" to dismiss them. While the company claims to be working on reducing the frequency of these interruptions, the damage to customer trust may already be done.
UPDATE: Jeep Claims 'Software Glitch' Disabled Opting-Out of In-Vehicle Pop-Up Ads in 'a Few' Cases

Today T-Mobile announced what they're calling "the next big thing in wireless" — T-Mobile Starlink. But the real surprise is "The beta is now open for absolutely everyone — yes, even Verizon and AT&T customers — to register for free access until July."

And, as they explained to Americans watching the Super Bowl, "If you can see the sky you're connected." Now in public beta, this breakthrough service, developed in partnership with Starlink, uses straight-out-of-a-sci-fi-movie satellite and mobile communications technology to help keep people connected — even you, Verizon and AT&T customers — in the more than 500,000 square miles of the country unreached by any carrier's earth-bound cell towers. That's nearly the size of two Texases...! The beauty of the service is its simplicity: users don't need to do anything out of the ordinary. When a user's cell phone gets out of range of a cell tower, the phone automatically connects to the T-Mobile Starlink network. No need to manually connect. Messages are sent and received just as they are today on a traditional network, even group texts and reactions. And it works on most smartphones from the last four years. It's not limited to a few smartphones or operating systems...

The beta is free until July at which point T-Mobile Starlink will be included at no extra cost on Go5G Next (including variations like Go5G Next 55+), T-Mobile's best plan. Business customers will also get T-Mobile Starlink at no extra cost on Go5G Business Next, first responder agencies on T-Priority plans and other select premium rate plans. T-Mobile customers on any other plan can add the service for $15/month per line. Through February, T-Mobile customers who have registered for the beta can secure a $10/month per line Early Adopter Discount, 33% off the full price.

AT&T and Verizon customers hate dead zones, too

When your service is amazing and different, you want as many people to try it as possible. T-Mobile is giving AT&T and Verizon customers the opportunity to try out T-Mobile Starlink satellite service on their existing phones... During the beta period, Verizon and AT&T customers can experience T-Mobile Starlink text messaging for free, and once the service launches in July, it will be available for $20/month per line... More details and consumer registration can be found here.

A Vision for Universal Coverage

As T-Mobile and Starlink continue to work towards eliminating mobile deadzones, the companies welcome wireless providers from around the world to join their growing alliance, which aims to provide reciprocal roaming for all participating carriers. So far, KDDI (Japan), Telstra (Australia), Optus (Australia), One NZ (New Zealand), Salt (Switzerland), Entel (Chile & Peru), Rogers (Canada) and Kyivstar (Ukraine) are among the providers that have signed on to join the cause and launch satellite-to-mobile technology. Learn more about the alliance and how providers can join at direct.starlink.com.

The unemployment rate in America's information technology sector "rose from 3.9% in December to 5.7% in January," reports the Wall Street Journal. (Alternate URL here.) Meanwhile last month's overall jobless rate was just 4%, they point out, calling it "the latest sign of how automation and the increasing use of artificial intelligence are having a negative impact on the tech labor market."

Companies began implementing their annual spending cuts in January, and there were layoffs at large tech companies like Meta. But whatever the reason, "The number of unemployed IT workers rose from 98,000 in December to 152,000 last month, according to a report from consulting firm Janco Associates based on data from the U.S. Department of Labor," while the Labor Department said the overall economy added 143,000 jobs.

One management consulting firm offers this explanation: Job losses in tech can be attributed in part to the influence of AI, according to Victor Janulaitis, chief executive of Janco Associates. The emergence of generative AI has produced massive amounts of spending by tech giants on AI infrastructure, but not necessarily new jobs in IT. "Jobs are being eliminated within the IT function which are routine and mundane, such as reporting, clerical administration," Janulaitis said. "As they start looking at AI, they're also looking at reducing the number of programmers, systems designers, hoping that AI is going to be able to provide them some value and have a good rate of return."

Increased corporate investment in AI has shown early signs of leading to future cuts in hiring, a concept some tech leaders are starting to call "cost avoidance." Rather than hiring new workers for tasks that can be more easily automated, some businesses are letting AI take on that work — and reaping potential savings. The latest IT jobs numbers come as unemployment among white-collar workers remains at its highest levels since 2020, according to Cory Stahle, an economist at hiring website Indeed. "What we've really seen, especially in the last year or so, is a bifurcation in opportunities, where white-collar knowledge worker type jobs have had far less employer demand than jobs that are more in-person, skilled labor jobs," Stahle said.
Stahle notes that job postings at Indeed.com for software developers declined 8.5% in January from a year earlier...

"It is now 45+ years since C++ was first conceived," writes 74-year-old C++ creator Bjarne Stroustrup in an article this week for Communications of the ACM. But he complains that many developers "use C++ as if it was still the previous millennium," in an article titled 21st Century C++ that promises "the key concepts on which performant, type safe, and flexible C++ software can be built: resource management, life-time management, error-handling, modularity, and generic programming...

"At the end, I present ways to ensure that code is contemporary, rather than relying on outdated, unsafe, and hard-to-maintain techniques: guidelines and profiles." To help developers focus on effective use of contemporary C++ and avoid outdated "dark corners" of the language, sets of guidelines have been developed. Here I focus on the C++ Core guidelines that I consider the most ambitious... My principal aim is a type-safe and resource-safe use of ISO standard C++. That is:

- Every object is exclusively used according to its definition
- No resource is leaked

This encompasses what people refer to as memory safety and much more. It is not a new goal for C++. Obviously, it cannot be achieved for every use of C++, but by now we have years of experience showing that it can be done for modern code, though so far enforcement has been incomplete... When thinking about C++, it is important to remember that C++ is not just a language but part of an ecosystem consisting of implementations, libraries, tools, teaching, and more.
WG21 (and others) are working on "profiles" to enforce guidelines (though they're "not yet available, except for experimental and partial versions"). But Stroustrup writes that the C++ Core Guidelines "use a strategy known as subset-of-superset." First: extend the language with a few library abstractions: use parts of the standard library and add a tiny library to make use of the guidelines convenient and efficient (the Guidelines Support Library, GSL).
Next: subset: ban the use of low-level, inefficient, and error-prone features.

What we get is "C++ on steroids": Something simple, safe, flexible, and fast; rather than an impoverished subset or something relying on massive run-time checking. Nor do we create a language with novel and/or incompatible features. The result is 100% ISO standard C++. Messy, dangerous, low-level features can still be enabled and used when needed.
Stroustrup writes that the C++ Core Guidelines focus on rules "we hope that everyone eventually could benefit from."

• No uninitialized variables

• No range or nullptr violations

• No resource leaks

• No dangling pointers

• No type violations

• No invalidation

Bjarne Stroustrup answered questions from Slashdot readers in 2014...

The Free Software Foundation "hinted that it would organize an unprecedented virtual memorabilia auction" in March to celebrate this year's 40th anniversary, according to an announcement this week. Those hints "left collectors and free software fans wondering which of the pieces of the FSF's history would be auctioned off."

But Tuesday the FSF "lifted the veil and gave a sneak peak of some of the more prestigious entries in the memorabilia auction." First of all, the memorabilia auction will feature an item that could be especially interesting for art collectors but will certainly also draw the attention of free software fans from all over: the original GNU head drawing by Etienne Suvasa, which became the blueprint for the iconic GNU logo present everywhere in the free software world.

The list of memorabilia for sale also entails some rare and historic hardware, such as a "terminus-est" microcomputer, and an Amiga 3000UX that was used in the FSF's old office at the Massachusetts Institute of Technology (MIT) in the early days of GNU, when these machines were capable of running a GNU-like operating system. Another meaningful item to be auctioned off, and one that collectors will want to keep a keen eye on, is the Internet Hall of Fame medal awarded to founder Richard Stallman. When Stallman was inducted into the Internet Hall of Fame, it was the ultimate recognition of free software's immense impact on the development and advancement of the Internet. This medal is definitely worthy of joining a fine historical collection...! [T]here are several more historic awards, more original GNU artwork, and a legendary katana [as seen in an XKCD comic] that became a lighthearted weapon in the fight for computer user freedom.

The auction is only the opening act to a whole agenda of activities celebrating forty years of free software activism. In May, the FSF invites free software supporters all over the world to gather for local in-person community meetups to network, discuss what people can do next to make the world freer, and celebrate forty years of commitment to software freedom. Then, on the actual birthday of the FSF on October 4, 2025, the organization intends to bring the international free software community to Boston for a celebration featuring keynotes and workshops by prominent personalities of the free software movement.
"The bidding will start as a virtual silent auction on March 17 and run through March 21, with more auction items revealed each day, and will culminate in an virtual live auction on March 23, 2025, 14:00 to 17:00 EDT," according to the announcement.

"Register here to attend the live auction. There's no need to register for the silent auction; you can simply join the bidding on the FSF's LibrePlanet wiki."

Friday on a Washington Post podcast, their columnists discussed the hybrid/remote work trend, asking why it "seems to be reversing". Molly Roberts: Why have some companies decided finally that having offices full of employees is better for them?

Heather Long: It's a loaded question, but I would say, unfortunately, 2025 is the year of operational efficiency, and that's corporate speak for save money at all costs. How do you save money? The easiest way is to get people to quit. What are these return to office mandates, particularly the five day a week in office mandates? We have a lot of data on this now, and it shows people will quit and you don't even have to pay them severance to do it.

Molly Roberts: It's not about productivity for the people who are in the office, then, you think. It's more about just cutting down on the size of the workforce generally.

Heather Long: I do think so. There has been a decent amount of research so far on fully remote, hybrid and fully in office. It's a mixed bag for fully remote. That's why I think if you look at the Fortune 500, only about 16 companies are fully remote, but a lot of them are hybrid. The reason that so much companies are hybrid is because that's the sweet spot. There is no productivity difference between the hybrid schedule and fully in the office five days a week. But what you do see a big difference is employee satisfaction and happiness and employee retention....

I think if what we're talking about is places that have been able to do work from home successfully for the past several years, why are they suddenly in 2025, saying the whole world has changed and we need to come back to the office five days a week? You should definitely be skeptical.
"Who are the first people to leave in these scenarios? It's star employees who know they can get a job elsewhere," Long says (adding later that "There's also quantifiable data that show that, particularly parents, the childcare issues are real.") Long also points out that most of Nvidia's workforce is fully remote — and that housing prices have spiked in some areas where employers are now demanding people return to the office.

But employers also know hiring rates are now low, argues Long, so they're pushing their advantage — possibly out of some misplaced nostalgia. "[T]here's a huge, huge perception difference between what managers, particularly senior leaders in an organization, how effective they think [people were] in offices versus what the rank and file people think. Rank and file people tend to prefer hybrid because they don't want their time wasted."

Their discussion also notes a recent Harvard Business School survey that found that 40% of people would trade 5% or more of their salaries to work from home....

An anonymous reader quotes a report from Ars Technica: In today's issue of Nature, a team at Oxford University describes using quantum teleportation to link two pieces of quantum hardware that were located about 2 meters apart, meaning they could easily have been in different rooms entirely. Once linked, the two pieces of hardware could be treated as a single quantum computer, allowing simple algorithms to be performed that involved operations on both sides of the 2-meter gap. [...] The Oxford team was simply interested in a proof-of-concept, and so used an extremely simplified system. Each end of the 2-meter gap had a single trap holding two ions, one strontium and one calcium. The two atoms could be entangled with each other, getting them to operate as a single unit.

The calcium ion served as a local memory and was used in computations, while the strontium ion served as one of the two ends of the quantum network. An optical cable between the two ion traps allowed photons to entangle the two strontium ions, getting the whole system to operate as a single unit. The key thing about the entanglement processes used here is that a failure to entangle left the system in its original state, meaning that the researchers could simply keep trying until the qubits were entangled. The entanglement event would also lead to a photon that could be measured, allowing the team to know when success had been achieved (this sort of entanglement with a success signal is termed "heralded" by those in the field).

The researchers showed that this setup allowed them to teleport with a specific gate operation (controlled-Z), which can serve as the basis for any other two-qubit gate operation -- any operation you might want to do can be done by using a specific combination of these gates. After performing multiple rounds of these gates, the team found that the typical fidelity was in the area of 70 percent. But they also found that errors typically had nothing to do with the teleportation process and were the product of local operations at one of the two ends of the network. They suspect that using commercial hardware, which has far lower error rates, would improve things dramatically. Finally, they performed a version of Grover's algorithm, which can, with a single query, identify a single item from an arbitrarily large unordered list. The "arbitrary" aspect is set by the number of available qubits; in this case, having only two qubits, the list maxed out at four items. Still, it worked, again with a fidelity of about 70 percent.

While the work was done with trapped ions, almost every type of qubit in development can be controlled with photons, so the general approach is hardware-agnostic. And, given the sophistication of our optical hardware, it should be possible to link multiple chips at various distances, all using hardware that doesn't require the best vacuum or the lowest temperatures we can generate. That said, the error rate of the teleportation steps may still be a problem, even if it was lower than the basic hardware rate in these experiments. The fidelity there was 97 percent, which is lower than the hardware error rates of most qubits and high enough that we couldn't execute too many of these before the probability of errors gets unacceptably high.

The Register's Thomas Claburn reports: Google's overhaul of Chrome's extension architecture continues to pose problems for developers of ad blockers, content filters, and privacy tools. [...] While Google's desire to improve the security, privacy, and performance of the Chrome extension platform is reasonable, its approach -- which focuses on code and permissions more than human oversight -- remains a work-in-progress that has left extension developers frustrated.

Alexei Miagkov, senior staff technology at the Electronic Frontier Foundation, who oversees the organization's Privacy Badger extension, told The Register, "Making extensions under MV3 is much harder than making extensions under MV2. That's just a fact. They made things harder to build and more confusing." Miagkov said with Privacy Badger the problem has been the slowness with which Google addresses gaps in the MV3 platform. "It feels like MV3 is here and the web extensions team at Google is in no rush to fix the frayed ends, to fix what's missing or what's broken still." According to Google's documentation, "There are currently no open issues considered a critical platform gap," and various issues have been addressed through the addition of new API capabilities.

Miagkov described an unresolved problem that means Privacy Badger is unable to strip Google tracking redirects on Google sites. "We can't do it the correct way because when Google engineers design the [chrome.declarativeNetRequest API], they fail to think of this scenario," he said. "We can do a redirect to get rid of the tracking, but it ends up being a broken redirect for a lot of URLs. Basically, if the URL has any kind of query string parameters -- the question mark and anything beyond that -- we will break the link." Miagkov said a Chrome developer relations engineer had helped identify a workaround, but it's not great. Miagkov thinks these problems are of Google's own making -- the company changed the rules and has been slow to write the new ones. "It was completely predictable because they moved the ability to fix things from extensions to themselves," he said. "And now they need to fix things and they're not doing it."

An anonymous reader quotes a report from CyberScoop: Ransomware payments saw a dramatic 35% drop last year compared to 2023, even as the overall frequency of ransomware attacks increased, according to a new report released by blockchain analysis firm Chainalysis. The considerable decline in extortion payments is somewhat surprising, given that other cybersecurity firms have claimed that 2024 saw the most ransomware activity to date. Chainalysis itself warned in its mid-year report that 2024's activity was on pace to reach new heights, but attacks in the second half of the year tailed off. The total amount in payments that Chainalysis tracked in 2024 was $812.55 million, down from 2023's mark of $1.25 billion.

The disruption of major ransomware groups, such as LockBit and ALPHV/BlackCat, were key to the reduction in ransomware payments. Operations spearheaded by agencies like the United Kingdom's National Crime Agency (NCA) and the Federal Bureau of Investigation (FBI) caused significant declines in LockBit activity, while ALPHV/BlackCat essentially rug-pulled its affiliates and disappeared after its attack on Change Healthcare. [...] Additionally, [Chainalysis] says more organizations have become stronger against attacks, with many choosing not to pay a ransom and instead using better cybersecurity practices and backups to recover from these incidents. [...] Chainalysis also says ransomware operators are letting funds sit in wallets, refraining from moving any money out of fear they are being watched by law enforcement.

You can read the full report here.

An anonymous reader quotes a report from Ars Technica: On Tuesday, Hugging Face researchers released an open source AI research agent called "Open Deep Research," created by an in-house team as a challenge 24 hours after the launch of OpenAI's Deep Research feature, which can autonomously browse the web and create research reports. The project seeks to match Deep Research's performance while making the technology freely available to developers. "While powerful LLMs are now freely available in open-source, OpenAI didn't disclose much about the agentic framework underlying Deep Research," writes Hugging Face on its announcement page. "So we decided to embark on a 24-hour mission to reproduce their results and open-source the needed framework along the way!"

Similar to both OpenAI's Deep Research and Google's implementation of its own "Deep Research" using Gemini (first introduced in December -- before OpenAI), Hugging Face's solution adds an "agent" framework to an existing AI model to allow it to perform multi-step tasks, such as collecting information and building the report as it goes along that it presents to the user at the end. The open source clone is already racking up comparable benchmark results. After only a day's work, Hugging Face's Open Deep Research has reached 55.15 percent accuracy on the General AI Assistants (GAIA) benchmark, which tests an AI model's ability to gather and synthesize information from multiple sources. OpenAI's Deep Research scored 67.36 percent accuracy on the same benchmark with a single-pass response (OpenAI's score went up to 72.57 percent when 64 responses were combined using a consensus mechanism).

As Hugging Face points out in its post, GAIA includes complex multi-step questions such as this one: "Which of the fruits shown in the 2008 painting 'Embroidery from Uzbekistan' were served as part of the October 1949 breakfast menu for the ocean liner that was later used as a floating prop for the film 'The Last Voyage'? Give the items as a comma-separated list, ordering them in clockwise order based on their arrangement in the painting starting from the 12 o'clock position. Use the plural form of each fruit." To correctly answer that type of question, the AI agent must seek out multiple disparate sources and assemble them into a coherent answer. Many of the questions in GAIA represent no easy task, even for a human, so they test agentic AI's mettle quite well. Open Deep Research "builds on OpenAI's large language models (such as GPT-4o) or simulated reasoning models (such as o1 and o3-mini) through an API," notes Ars. "But it can also be adapted to open-weights AI models. The novel part here is the agentic structure that holds it all together and allows an AI language model to autonomously complete a research task."

The code has been made public on GitHub.

In 2022, writer and activist Cory Doctorow coined the term "enshittification" to describe the gradual deterioration of a service or product. The term's prevalence has increased to the point that it was the National Dictionary of Australia's word of the year last year. The editors at Ars Technica, having "covered a lot of things that have been enshittified," decided to highlight some of the worst examples the've come across. Here's a summary of each thing mentioned in their report: Smart TVs: Evolved into data-collecting billboards, prioritizing advertising and user tracking over user experience and privacy. Features like convenient input buttons are sacrificed for pushing ads and webOS apps. "This is all likely to get worse as TV companies target software, tracking, and ad sales as ways to monetize customers after their TV purchases -- even at the cost of customer convenience and privacy," writes Scharon Harding. "When budget brands like Roku are selling TV sets at a loss, you know something's up."

Google's Voice Assistant (e.g., Nest Hubs): Functionality has degraded over time, with previously working features becoming unreliable. Users report frequent misunderstandings and unresponsiveness. "I'm fine just saying it now: Google Assistant is worse now than it was soon after it started," writes Kevin Purdy. "Even if Google is turning its entire supertanker toward AI now, it's not clear why 'Start my morning routine,' 'Turn on the garage lights,' and 'Set an alarm for 8 pm' had to suffer."

Portable Document Format (PDF): While initially useful for cross-platform document sharing and preserving formatting, PDFs have become bloated and problematic. Copying text, especially from academic journals, is often garbled or impossible. "Apple, which had given the PDF a reprieve, has now killed its main selling point," writes John Timmer. "Because Apple has added OCR to the MacOS image display system, I can get more reliable results by screenshotting the PDF and then copying the text out of that. This is the true mark of its enshittification: I now wish the journals would just give me a giant PNG."

Televised Sports (specifically cycling and Formula 1): Streaming services have consolidated, leading to significantly increased costs for viewers. Previously affordable and comprehensive options have been replaced by expensive bundles across multiple platforms. "Formula 1 racing has largely gone behind paywalls, and viewership is down significantly over the last 15 years," writes Eric Berger. "Major US sports such as professional and college football had largely been exempt, but even that is now changing, with NFL games being shown on Peacock, Amazon Prime, and Netflix. None of this helps viewers. It enshittifies the experience for us in the name of corporate greed."

Google Search: AI overviews often bury relevant search results under lengthy, sometimes inaccurate AI-generated content. This makes finding specific information, especially primary source documents, more difficult. "Google, like many big tech companies, expects AI to revolutionize search and is seemingly intent on ignoring any criticism of that idea," writes Ashley Belanger.

Email AI Tools (e.g., Gemini in Gmail): Intrusive and difficult to disable, these tools offer questionable value due to their potential for factual inaccuracies. Users report being unable to fully opt-out. "Gmail won't take no for an answer," writes Dan Goodin. "It keeps asking me if I want to use Google's Gemini AI tool to summarize emails or draft responses. As the disclaimer at the bottom of the Gemini tool indicates, I can't count on the output being factual, so no, I definitely don't want it."

Windows: While many complaints about Windows 11 originated with Windows 10, the newer version continues the trend of unwanted features, forced updates, and telemetry data collection. Bugs and performance issues also plague the operating system. "... it sure is easy to resent Windows 11 these days, between the well-documented annoyances, the constant drumbeat of AI stuff (some of it gated to pricey new PCs), and a batch of weird bugs that mostly seem to be related to the under-the-hood overhauls in October's Windows 11 24H2 update," writes Andrew Cunningham. "That list includes broken updates for some users, inoperable scanners, and a few unplayable games. With every release, the list of things you need to do to get rid of and turn off the most annoying stuff gets a little longer."

Web Discourse: The rapid spread of memes, trends, and corporate jargon on social media has led to a homogenization of online communication, making it difficult to distinguish original content and creating a sense of constant noise. "[T]he enshittifcation of social media, particularly due to its speed and virality, has led to millions vying for their moment in the sun, and all I see is a constant glare that makes everything look indistinguishable," writes Jacob May. "No wonder some companies think AI is the future."

The Register's Thomas Claburn reports: Oracle this week asked the US Patent and Trademark Office (USPTO) to partially dismiss a challenge to its JavaScript trademark. The move has been criticized as an attempt to either stall or water down legal action against the database goliath over the programming language's name. Deno Land, the outfit behind the Deno JavaScript runtime, filed a petition with the USPTO back in November in an effort to make the trademarked term available to the JavaScript community. This legal effort is led by Node.js creator and Deno Land CEO Ryan Dahl, summarized on the JavaScript.tm website, and supported by more than 16,000 members of the JavaScript community. It aims to remove the fear of an Oracle lawsuit for using the term "JavaScript" in a conference title or business venture.

"Programmers working with JavaScript have formed innumerable community organizations," the website explains. "These organizations, like the standards bodies, have been forced to painstakingly avoid naming the programming language they are built around -- for example, JSConf. Sadly, without risking a legal trademark challenge against Oracle, there can be no 'JavaScript Conference' nor a 'JavaScript Specification.' The world's most popular programming language cannot even have a conference in its name." [...] In the initial trademark complaint, Deno Land makes three arguments to invalidate Oracle's ownership of "JavaScript." The biz claims that JavaScript has become a generic term; that Oracle committed fraud in 2019 when it applied to renew its trademark; and that Oracle has abandoned its trademark because it does not offer JavaScript products or services.

Oracle's motion on Monday focuses on the dismissal of the fraud claim, while arguing that it expects to prevail on the other two claims, citing corporate use of the trademarked term "in connection with a variety of offerings, including its JavaScript Extension Toolkit as well as developer's guides and educational resources, and also that relevant consumers do not perceive JavaScript as a generic term." The fraud claim follows from Deno Land's assertion that the material Oracle submitted in support of its trademark renewal application has nothing to do with any Oracle product. "Oracle, through its attorney, submitted specimens showing screen captures of the Node.js website, a project created by Ryan Dahl, Petitioner's Chief Executive Officer," the trademark cancellation petition says. "Node.js is not affiliated with Oracle, and the use of screen captures of the 'nodejs.org' website as a specimen did not show any use of the mark by Oracle or on behalf of Oracle."

Oracle contends that in fact it submitted two specimens to the USPTO -- a screenshot from the Node.js website and another from its own Oracle JavaScript Extension Toolkit. And this, among other reasons, invalidates the fraud claim, Big Red's attorneys contend. "Where, as here, Registrant 'provided the USPTO with [two specimens]' at least one of which shows use of the mark in commerce, Petitioner cannot plausibly allege that the inclusion of a second, purportedly defective specimen, was material," Oracle's motion argues, adding that no evidence of fraudulent intent has been presented. Beyond asking the court to toss the fraud claim, Oracle has requested an additional thirty days to respond to the other two claims.

During the fourth quarter of 2024, AMD surpassed Intel in datacenter sales for the first time in history -- despite weaker-than-expected sales of its datacenter GPUs. Tom's Hardware reports: AMD's revenue in Q4 2024 totaled $7.658 billion, up 24% year-over-year. The company's gross margin hit 51%, whereas net income was $482 million. On the year basis, 2024 was AMD's best year ever as the company's revenue reached $25.8 billion, up 14% year-over-year. The company earned net income of $1.641 billion as its gross margin hit 49%. But while the company's annual results are impressive, there is something about Q4 results that AMD should be proud of.

Datacenter business was the company's primary source of earnings, with net revenue reaching record $3.86 billion in Q4, marking a 69% year-over-year (YoY) increase and a 9% quarter-over-quarter (QoQ) rise. Operating income also saw substantial improvement, surging 74% YoY to $1.16 billion. By contrast, Intel's datacenter and AI business unit posted $3.4 billion revenue, while its operating income reached $200 million. But while the quarter marked a milestone for AMD, market analysts expected AMD to sell more of its Instinct MI300-series GPUs for AI and HPC. You can view AMD's 2024 financial results here.