news aggregator

Point release retuned with new FP8 datatype for better compatibility with homegrown silicon

Chinese AI darling DeepSeek unveiled an update to its flagship large language model that the company claims is already optimized for use with a new generation of homegrown silicon.…

LinkedIn's engagement-driven algorithm systematically elevates shallow, meaningless content over substantive professional discourse, according to a new analysis that highlights how major platforms prioritize user retention metrics over content quality. Entrepreneur and product executive Elliot Smith describes encountering a "seemingly endless stream of posts that are over fluffed, over produced and ultimately say nothing" that he categorizes as "toxic mediocrity." Smith argues the platform's reward system creates a destructive cycle where "comments, likes and other engagement" signal user activity to LinkedIn's algorithm, which then promotes similar vapid content. "LinkedIn wants you on LinkedIn," Smith writes, noting the Microsoft-owned platform correlates engagement with ad clicks and premium conversions. Smith recommends professionals focus on substantive work rather than platform gaming, arguing "nothing you post there is going to change your career."

Read more of this story at Slashdot.

Pro tip: When taking revenge, don't use your real name

A US court sentenced a former developer at power management biz Eaton to four years in prison after he installed malware on the company’s servers.…

Starting September 1st, Russia will require all smartphones and tablets sold in the country to come with MAX, a state-backed messaging app seen as a rival to WhatsApp and Telegram. Critics say the app could be used to track users. Reuters reports: The Russian government said in a statement that MAX, which will be integrated with government services, would be on a list of mandatory pre-installed apps on all "gadgets," including mobile phones and tablets, sold in Russia from September 1. State media says accusations from Kremlin critics that MAX is a spying app are false and that it has fewer permissions to access user data than rivals WhatsApp and Telegram. It will also be mandatory that from September 1, Russia's domestic app store, RuStore, which is pre-installed on all Android devices, will be pre-installed on Apple devices. A Russian-language TV app called LIME HD TV, which allows people to watch state TV channels for free, will be pre-installed on all smart TVs sold in Russia from January 1, the government added. [...] MAX said this week that 18 million users had downloaded its app, parts of which are still in a testing phase. Russia's interior ministry said on Wednesday that MAX was safer than foreign rivals, but that it had arrested a suspect in the first fraud case using the new messenger.

Read more of this story at Slashdot.

Because savvy terrorists always use public internet services to plan their mischief, right?

Anthropic says it has scanned an undisclosed portion of conversations with its Claude AI model to catch concerning inquiries about nuclear weapons.…

Reviewers are more likely to approve a manuscript if their own work is cited in subsequent versions than are reviewers who are not cited, according to an analysis of 18,400 articles from four open-access publications. From a report: The study, which is yet to be peer reviewed, was posted online as a preprint earlier this month. The study was inspired by anecdotes from authors who cited articles only because reviewers asked them to, says study author Adrian Barnett, who researches peer review and meta-research at Queensland University of Technology in Brisbane, Australia. Sometimes, these requests are fine, he says. But if reviewers ask for too many citations or the reason to cite their work is not justified, the peer-review process can become transactional, says Barnett. Citations increase a researcher's h-index, a metric reflecting the impact of their publications.

Read more of this story at Slashdot.

Google's Android 16 QPR2 beta 1 is rolling out with new customization features, including the ability to force dark mode and icon themes on apps that don't support them. The update also adds enhanced parental controls, better data migration, PDF editing, and Bluetooth audio sharing, with a full release expected in December. The Verge reports: The beta includes a new dark theme option that will "intelligently invert the UI of apps that appear light despite users having selected the dark theme" when enabled, according to Google's announcement, forcibly making apps that don't natively support the feature to appear darker. Google says this is "largely intended as an accessibility feature" for users with low vision or photosensitivity, and will also automatically darken app splash screens and adjust status bar colors to match the darker theming. Another feature will allow users to forcibly apply themed icon colors to apps that don't natively support them. Android's icon theming currently only works if app developers have provided a monochrome version of their app icon that can be adjusted, which is annoying for users who want to apply a consistent aesthetic across their entire home page. Auto-themed app icons spare developers from adding this capability manually, removing the hassle for users to customize their phone's theme. The full list of features in the QPR2 beta 1 update can be found on the Android developers' blog.

Read more of this story at Slashdot.

Better late than never after SharePoint assault?

Microsoft has reportedly stopped giving Chinese companies proof-of-concept exploit code for soon-to-be-disclosed vulnerabilities following last month's SharePoint zero-day attacks, which appear to be related to a leak in Redmond's early-bug-notification program.…

An anonymous reader quotes a report from the New York Times: Jay Blahnik was a fitness superstar with a book and nearly two decades of work with Nike before he was hired in 2013 to work on the Apple Watch. He became known inside Apple as the creator of the watch's signature fitness feature: three circular bands that people could complete through the day by exercising, standing and burning calories. Marketed with the tagline "Close Your Rings," the concept helped galvanize sales of Apple's first breakout product after Steve Jobs's death. But along the way, Mr. Blahnik created a toxic work environment (Warning: source may be paywalled; alternative source), said nine current and former employees who worked with or for Mr. Blahnik and spoke about personnel issues on the condition of anonymity. They said Mr. Blahnik, 57, who leads a roughly 100-person division as vice president for fitness technologies, could be verbally abusive, manipulative and inappropriate. His behavior contributed to decisions by more than 10 workers to seek extended mental health or medical leaves of absence since 2022, about 10 percent of the team, these people said. When confronted with Mr. Blahnik's behavior, Apple moved to protect him after an internal investigation. The company settled one complaint alleging sexual harassment by Mr. Blahnik and is fighting a lawsuit by an employee, Mandana Mofidi, who said he had bullied her. Mr. Blahnik stayed in his job after company officials said their investigation had found no evidence of wrongdoing, according to interviews and Ms. Mofidi's lawsuit, which she filed against Mr. Blahnik and Apple last year in Los Angeles County Superior Court. The tension inside Mr. Blahnik's division speaks to workplace dysfunction at the heart of one of Apple's signature health initiatives. These employees said the company was more willing to protect a star executive than address the concerns of rank-and-file workers.

Read more of this story at Slashdot.

Underground forums now recruiting English-speaking social engineers

English speakers adept at social engineering are a hot commodity in the cybercrime job market.…

Google is rolling out AI Mode in Search to 180 countries, expanding beyond the US, UK, and India, with even more countries being added soon. The search giant is also expanding its AI Mode's agentic capabilities, so you can now use natural language to find restaurant reservations. Engadget reports: Google says you can ask about getting a dinner reservation with conditions such as group size, date, location and your preference of cuisine, all of which be taken into consideration when AI Mode pulls in its results from across the web. Suggestions will be presented in list form with the available reservation slots. It'll also provide a link to the booking page you need. Google also plans to add local service appointments and event ticketing capabilities soon, with Ticketmaster and StubHub among its partners. AI Mode leverages Google's web-browsing AI agent Project Mariner' its direct partners on Search and resources like Knowledge Graph and Google Maps when prompted to find you somewhere to eat. It has partnered with the likes of OpenTable, Resy and Tock to incorporate as many restaurants as possible and streamline the booking process. Right now, this feature is exclusive to those subscribed to the wildly expensive Google AI Ultra plan in the US, and can be accessed through its Labs platform. If you opt into the AI Mode experiment it can also remember your previous conversations and searches to give you results that more closely match your preferences.

Read more of this story at Slashdot.

Google’s Gemini-powered tools tripped up by image-scaling prompt injection

Security researchers with Trail of Bits have found that Google Gemini CLI and other production AI systems can be deceived by image scaling attacks, a well-known adversarial challenge for machine learning systems.…

Light pollution extends the active hours of birds worldwide by nearly an hour, fundamentally altering the natural rhythms of over 500 species across multiple continents. Analysis of 60 million bird call recordings reveals that artificial lighting causes birds to begin vocalizing 20 minutes earlier each morning and continue 30 minutes later each evening. The effect varies significantly by species anatomy and behavior. Birds with proportionally larger eyes experience the strongest responses, with American Robins in highly illuminated areas beginning their songs up to two hours before sunrise. Species nesting in tree cavities or enclosed spaces show minimal changes, as opaque barriers shield them from artificial light. The research, drawing from BirdWeather devices deployed globally by birdwatchers, is the largest analysis of light pollution's biological impacts to date.

Read more of this story at Slashdot.

Denmark's PostNord will end centuries of letter delivery as digitalization and high postage costs make physical mail nearly obsolete. The BBC reports: The steep decline in letter volumes has been driven largely by digitalization, and PostNord announced in March that it will cease letter services at the end of the year. It will bring to an end four centuries of letter deliveries by the state-owned operation. A third of its workforce is being let go, as it sheds 2,200 positions in its loss-making letter arm. Instead it will focus on its profitable parcel business, creating 700 new roles. "Danes hardly receive any letters anymore. It's been going down for years and years," says Kim Pedersen, chief of PostNord Denmark. "They're receiving one letter a month on average, it's not a lot." "On the contrary, Danes love to shop online," he adds. "Global e-commerce is growing significantly, and we are moving with it." Fifteen years ago, PostNord operated several enormous letter-sorting facilities, but now there's just one on the western outskirts of Copenhagen. Since 2000, the volume of letters the business handles has declined by more than 90%, from around 1.4 billion to 110 million last year, and it continues to fall rapidly. As PostNord prepares to cease letter deliveries, 1,500 of its red post boxes are being removed from Danish streets. However, few locals in the capital appear to use them much. From email and cashless mobile payments, to digital health cards carried by smartphone, there's an app for almost everything in Denmark - and it's one of the world's most digitalized nations, second only to South Korea, according to the OECD's 2023 Digital Government Index. The Danish government has embraced a "digital by default" policy, and for more than a decade correspondence with the public has been carried out electronically. "We are facing this natural evolution of a digitalized society, earlier than maybe some other countries," Mr Pedersen explains. "In Denmark, we are maybe five or 10 years ahead." The high cost of sending a letter in Denmark is also a contributing factor behind its decline. In 2024, a new law opened up the postal market to private competition and took away its exemption from the country's 25% rate of VAT, so the price of a PostNord stamp jumped to 29 Danish krone ($4.55) per letter. "That made [volumes] drop even further faster," Mr Pedersen points out. The report notes that private firm DAO will take over nationwide letter deliveries in Denmark after PostNord exits. However, concerns remain that elderly citizens and rural residents may struggle with fewer post boxes and reduced service quality. Both the advocacy group DaneAge and the 3F Postal Union warn the transition could disproportionately affect vulnerable populations.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: As banks around the world prepare to replace many thousands of workers with AI, Australia's biggest bank is scrambling to rehire 45 workers after allegedly lying about chatbots besting staff by handling higher call volumes. In a statement Thursday flagged by Bloomberg, Australia's main financial services union, the Finance Sector Union (FSU), claimed a "massive win" for 45 union members whom the Commonwealth Bank of Australia (CBA) had replaced with an AI-powered "voice bot." The FSU noted that some of these workers had been with CBA for decades. Those workers in particular were shocked when CBA announced last month that their jobs had become redundant. At that time, CBA claimed that launching the chatbot supposedly "led to a reduction in call volumes" by 2,000 a week, FSU said. But "this was an outright lie," fired workers told FSU. Instead, call volumes had been increasing at the time they were dismissed, with CBA supposedly "scrambling" -- offering staff overtime and redirecting management to join workers answering phones to keep up. To uncover the truth, FSU escalated the dispute to a fair work tribunal, where the union accused CBA of failing to explain how workers' roles were ruled redundant. The union also alleged that CBA was hiring for similar roles in India, Bloomberg noted, which made it appear that CBA had perhaps used the chatbot to cover up a shady pivot to outsource jobs. While the dispute was being weighed, CBA admitted that "they didn't properly consider that an increase in calls" happening while staff was being fired "would continue over a number of months," FSU said. "This error meant the roles were not redundant," CBA confirmed at the tribunal. Now, CBA has apologized to the fired workers. A spokesperson told Bloomberg that they can choose to come back to their prior roles, seek another position, or leave the firm with an exit payment. "We have apologized to the employees concerned and acknowledge we should have been more thorough in our assessment of the roles required," CBA's spokesperson told Bloomberg. The FSU said that "the damage has already been done." These employees "have had to endure the stress and worry of facing redundancy" and were "suddenly confronted with the prospect of being unable to pay their bills." FSU warned that CBA's flip-flopping on AI serves as a "stark reminder to all of us that we can never trust employers to do the right thing by workers, and change can happen at any time and impact any one of us."

Read more of this story at Slashdot.

Cisco, the San Jose-based technology giant, has announced another round of layoffs affecting Bay Area workers, marking a familiar pattern of reporting skyrocketing revenue followed by drastic job cuts. From a report: According to Aug. 13 WARN filings with California's Employment Development Department, the company will eliminate 221 positions across its Milpitas and San Francisco offices. WARN documents are generally required by the state in the event of mass layoffs. Employees were notified of the layoffs on Aug. 14 and their terminations will be effective Oct. 13. The most cuts, affecting 157 jobs, largely in software engineering roles, were at Cisco's Milpitas office at 560 McCarthy Blvd. Cisco's San Francisco office at 500 Terry A. Francois Blvd. will cut 64 positions, according to the filing. The filings came the same day Cisco released its fourth-quarter earnings, which reported $14.7 billion in revenue, an 8% increase from the same quarter last year. Revenue for the 2025 fiscal year was $56.7 billion, up 5% from the previous year.

Read more of this story at Slashdot.

Bill would let US President commission white hat hackers to go after foreign threats, seize assets on the online seas

It's been more than 200 years since the United States issued a letter of marque allowing privateers to attack the vessels of foreign nations, but those letters may return to empower cyber operators if a bill introduced in Congress actually manages to pass. …

Gaming handhelds are becoming the industry's new battleground as Microsoft launches its ROG Xbox Ally devices October 16, chasing Nintendo Switch 2's record-breaking 5.8 million units sold in seven weeks. The ASUS-manufactured handhelds run full Windows 11 with a gaming-optimized interface, accessing Xbox Game Pass, Steam, Battle.net, and other PC storefronts without platform lockdown. Two models arrive at launch: the standard Xbox Ally with AMD Ryzen Z2 A processor and the premium Xbox Ally X featuring Ryzen AI Z2 Extreme, 24GB RAM, 1TB storage, and 80Wh battery. Microsoft's Handheld Compatibility Program pre-verifies thousands of PC games for portable play. Pricing remains unannounced.

Read more of this story at Slashdot.

If anyone’s gonna lock in Uncle Sam’s business, it'd better be us!

It's now safe to say the gang's all here when it comes to big generative AI model makers signing dollar discount deals with Uncle Sam. Google has joined Anthropic and OpenAI, inking questionable short-term discount terms for government agencies. …

Phone manufacturers rarely acknowledge that IP68 water resistance degrades over time, but Google has broken that silence with advertising disclaimers for its Pixel 10 Pro Fold. The fine print explicitly warns that water and dust protection "will diminish or be lost over time due to normal wear and tear, device repair, disassembly or damage." The company further notes that liquid damage voids warranties despite IP68 certification at manufacture.

Read more of this story at Slashdot.