RSS 생중계

An anonymous reader quotes a report from TechCrunch: Google on Tuesday announced the launch of Gemini 2.5 Pro Preview (I/O edition), an updated version of its flagship Gemini 2.5 Pro AI model that the company claims tops a number of widely used benchmarks. Gemini 2.5 Pro Preview (I/O edition) is available via the Gemini API and Google's Vertex AI and AI Studio platforms, and is priced the same as the Gemini 2.5 Pro model it effectively replaces. It's also in Google's Gemini chatbot app for the web and for mobile devices. The model's release comes ahead of Google's annual I/O developer conference (hence the "I/O edition" designation), where Google is expected to unveil a host of models, as well as AI-powered tools and platforms. [...] According to Google, Gemini 2.5 Pro Preview (I/O edition) has "significantly" improved capabilities for coding and building interactive web apps. The model is also better at tasks like code transformation -- that is, modifying a piece of code to achieve a specific goal -- and code editing, the company says. Google says the Gemini 2.5 Pro Preview (I/O edition) leads the WebDev Arena Leaderboard, a benchmark measuring a model's ability to create aesthetically pleasing and functional web apps. It also achieved a score of 84.8% on VideoMME, a popular benchmark designed to evaluate the video analysis capabilities of multi-modal large language models.

Read more of this story at Slashdot.

Linux systems can have large filesystems; trying to keep up with the stream of fanotify filesystem-monitoring notifications for them can be a struggle. Fanotify is one of a few ways to monitor accesses to filesystems provided by the kernel. Song Liu led a discussion on how to improve in-kernel filtering of fanotify events to a joint session of the filesystem and BPF tracks at the 2025 Linux Storage, Filesystem, Memory-Management, and BPF Summit. He wants to combine the best parts of a few different approaches to efficiently filter filesystem events.

Chief executives are exiting their posts at an unprecedented rate as economic volatility and emerging challenges reshape corporate leadership decisions, according to data from executive tracking firm Challenger, Gray & Christmas. Public-company CEO departures reached 373 last year, jumping 24% from 2023 levels. Among U.S. businesses with at least 25 employees, 2,221 chief executives left their positions in 2024, the highest number since Challenger began monitoring departures in 2002. Corporate leaders are citing AI, tariffs, recession fears and scrutiny of diversity initiatives as key stressors driving the exodus. "It's a very difficult time to lead," said Blake Irving, former GoDaddy CEO. "Given all the weird gyrations going on in the economy and with our new administration, it's really hard for even great leaders to find a true north." The trend extends beyond the C-suite, with managers 1.7 times more likely to report high workplace stress than rank-and-file employees, according to a recent McLean & Co. survey of over 200,000 workers.

Read more of this story at Slashdot.

In a combined filesystem and memory-management session at the 2025 Linux Storage, Filesystem, Memory Management, and BPF Summit (LSFMM+BPF), Joanne Koong led a discussion on improving the writeback performance for the Filesystem in Userspace (FUSE) layer. Writeback is how data that is written to the filesystem is actually flushed to the disk; it is the process of writing dirty pages from the page cache to storage. The current FUSE implementation allocates unmovable memory, then copies the dirty data to it before initiating writeback, which is slow; Koong wanted to change that behavior. Since the session, she has posted a patch set that has been applied by FUSE maintainer Miklos Szeredi.

English regulators have approved a new law firm that uses AI instead of lawyers to offer services for as little as $2.67, as the technology continues to disrupt industries from finance to accounting. From a report: Garfield AI, which was founded by a former London litigator and a quantum physicist, is an online tool that allows businesses and individuals such as tradespeople to chase debts owed to them at a substantially lower cost than the average lawyer's fees. Its AI assistant guides claimants through the small claims court process, including creating "polite chaser" letters for $2.67 and filing documents such as claim forms for $67, and can also produce arguments for claimants to use at trial. AI models are increasingly encroaching on legally sensitive tasks in high-paying sectors such as law and finance, potentially undercutting fees in high-volume work. Garfield received approval from the Solicitors Regulation Authority, the legal regulator for England and Wales, in March, in a move the latter hailed as a "landmark moment" for the industry.

Read more of this story at Slashdot.

Microsoft has unveiled a substantial AI-focused update for Windows 11 and Copilot+ PCs, introducing features that leverage neural processing units across the operating system. The update centers on AI-powered helpers across core Windows apps, with an intelligent agent in Settings that can locate and adjust options via natural voice commands. Key additions include expanded Click To Do functionality, allowing users to draft Word content based on screen context, engage Reading Coach, or send details directly to Excel tables. The Photos app gains a relight feature with support for three customizable light sources, while Paint adds object selection and text-to-sticker generation. Snipping Tool will automatically detect and crop prominent screen content, adding text extraction and color picking capabilities. System-level enhancements include an updated Start menu with phone companion integration, AI-powered actions in File Explorer for content summarization, and text generation in Notepad with new formatting options. Most features will debut first on Windows Insider builds for Snapdragon-powered Copilot+ PCs before expanding to systems with AMD or Intel chips. Several tools, including Ask Copilot and Reading Coach, are already available to Insiders.

Read more of this story at Slashdot.

BrianFagioli writes: Fedora Linux is now officially available as a Windows Subsystem for Linux (WSL) distribution! That's right, folks, following prior testing, you can now run Fedora 42 natively inside Windows using WSL. As someone who considers Fedora to be my favorite Linux distribution, this is a pretty exciting development. Installing it is simple enough. Just open up a terminal and type wsl --install FedoraLinux-42 to get started. After that, launch it with wsl -d FedoraLinux-42 and set your username. No password is required by default, and you'll automatically be part of the wheel group, meaning you can use sudo right out of the gate.

Read more of this story at Slashdot.

Reddit's Steve Huffman isn't mincing words about what he found when he came back as CEO in 2015: a company full of idealists who weren't exactly killing themselves with hard work. "We were really idealistic, and that's been good in many ways, but we were also idealistic about not being a business," Huffman said on the "Prof G Pod" podcast. "Wrapped up in some of that idealism was also not working very hard," he added. Huffman sees this as a Silicon Valley disease: "It's almost an entitlement of, 'I work at these companies, but I don't have to work very hard and I'm here for myself.'"

Read more of this story at Slashdot.

Recent college graduates face the worst job market in decades, with unemployment reaching 5.8%, according to recently released New York Federal Reserve data. The "recent-grad gap" - the difference between unemployment rates of young college graduates versus the overall labor force - has hit its lowest point in four decades, indicating college graduates are facing unusual difficulties securing employment. (The New York Federal Reserve said labor conditions for recent college graduates have "deteriorated noticeably" in the past few months.) Even graduates from elite MBA programs are struggling to find work, while law school applications have surged as young people seek shelter from the difficult job market. Economists are attributing the decline to three potential factors: incomplete recovery from pandemic disruptions, diminishing returns on college education, and possibly AI replacing entry-level positions. "When you think about what generative AI can do, it's the kind of things that young college grads have done," said David Deming, a Harvard economist. "They read and synthesize information and data. They produce reports and presentations."

Read more of this story at Slashdot.

Most people in the United States rely on federal science in their daily lives but don't realize it, a new nationwide poll of U.S. adults shows. NPR: The poll was conducted in early April by the Association of Science and Technology Centers, the association for science museums and other educational science centers in the U.S. The poll found that on a weekly basis more than 90% of people use weather forecasts, job market reports, food safety warnings and other information that is based on federal science. But only 10% of respondents are concerned that cuts to federal support for science might impact their access to such information. The Trump administration has made deep budget and personnel cuts to federal agencies that collect weather data and do safety inspections at factories that make food and prescription drugs, among many science-related functions. The association conducted the poll to understand current attitudes about science in the U.S. and inform how their member institutions, which include science museums, aquariums and zoos, can better serve the public.

Read more of this story at Slashdot.

Amazon has updated its Kindle iOS app with a new "Get Book" button that redirects users to complete purchases through their mobile browser, taking advantage of recent App Store rule changes. The update follows Judge Yvonne Gonzalez Rogers' April 30th ruling in Epic Games v. Apple, which bars Apple from collecting a 27% commission on purchases made outside apps or restricting how developers direct users to alternative payment options. Previously, iOS users had to visit Amazon's website through a browser to buy Kindle books -- a workaround implemented after Apple's 2011 rule changes required developers to remove links to external purchasing options. Apple has appealed the ruling but is complying in the interim.

Read more of this story at Slashdot.

Microsoft has instituted a stringent new performance management system that places ousted employees on a two-year rehiring block list and categorizes their departures as "good attrition," Business Insider reported Tuesday, citing internal documents. The company now tracks staff departures it considers beneficial, mirroring Amazon's "unregretted attrition" metric, though no specific targets have been established yet. Microsoft recently terminated 2,000 underperforming employees without severance and implemented a new performance improvement plan (PIP). Employees facing performance issues now must choose between entering the PIP or accepting a "Global Voluntary Separation Agreement" with 16 weeks of pay. Further reading: Microsoft Offers Underperformers Cash To Quit.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Reuters: OpenAI has agreed to buy artificial intelligence-assisted coding tool Windsurf for about $3 billion, Bloomberg News reported on Monday, citing people familiar with the matter. The deal has not yet closed, the report added. Windsurf, formerly known as Codeium, had recently been in talks with investors including General Catalyst and Kleiner Perkins to raise funding at a $3 billion valuation, according to Bloomberg News. The report notes that the deal "would be OpenAI's largest acquisition to date," further complementing ChatGPT's coding capabilities.

Read more of this story at Slashdot.

Security updates have been issued by Fedora (chromium and kappanhang), Red Hat (osbuild-composer and thunderbird), SUSE (chromedriver), and Ubuntu (c-ares, corosync, mysql-8.0, mysql-8.4, openjdk-17, openjdk-21, openjdk-24, openjdk-8, and openjdk-lts).

"Another year, yet another Hugo Awards-adjacent controversy?" writes Gizmodo's Cheryl Eddy, reporting that three key organizers of the 2025 Seattle Worldcon resigned after backlash over the use of ChatGPT to vet program participants. From the report: In a post on Bluesky co-signed by Hugo administrator Nicholas Whyte, deputy Hugo administrator Esther MacCallum-Stewart, and World Science Fiction Society division head Cassidy, the trio announced they were resigning from their roles ahead of the Seattle event, which takes place in August. "We want to reaffirm that no LLMs or generative AI have been used in the Hugo Awards process at any stage," the statement read in part, which might turn the heads of anyone who is a) interested in the Hugos, but b) not up on the latest controversy. However, plenty of people in the community are well aware of what's been going on. A quick journey to the blog File 770 will bring you up to speed, as will a visit to Seattle Worldcon 2025's own site, which on April 30 shared a post clarifying exactly what role AI played in the upcoming event. [...] However, as File 770 pointed out, the damage has apparently already been done: the use of ChatGPT in any capacity in connection to Worldcon created a furor on social media. It also inspired at least one Hugo nominee to remove their book from contention: Yoon Ha Lee, whose Moonstorm was named a Lodestar Award finalist, which honors YA releases. In a May 1 post on Bluesky, the author linked to the April 30 Worldcon blog post noted above, and noted he was withdrawing the title from consideration. Then, in a post shared today responding to File 770's latest post announcing the resignations, the author wrote âoeAll respect and I'm grateful to them for their work, sorry [things] came to this pass." Seattle Worldcon 2025 takes place August 13-17; the Hugo Awards will be handed out August 16.

Read more of this story at Slashdot.

AUTOSEL is a tool that is used to find kernel patches that should be considered for backporting into the stable releases. Sasha Levin has announced a new and completely rewritten version of AUTOSEL for those who would like to play with it.

Unlike the previous version that relied on word statistics and older neural network techniques, AUTOSEL leverages modern large language models and embedding technology to provide significantly more accurate recommendations.

According to Valve insider Tyler McVicker, Half-Life 3 is finally playable from start to finish and could be announced this summer, with a release as soon as winter 2025. Engadget reports: Besides McVicker's hours-long livestream, there have been other recent hints about Valve's progress on its highly anticipated title. In March, Valve concept artist Evgeniy Evstratiy claimed that he was in the room where Valve made Half-Life 3 on CG Voices Podcast. In the same month, another Valve leaker, Gabe Follower, claimed that Half-Life 3 would be the "end of Gordon's adventure," potentially signaling a non-cliffhanger ending to one of gaming's best franchises. Outside of these rumors, internet sleuths discovered code referencing HLX, which is widely thought to be the codename for Half-Life 3, in major updates to Deadlock and Dota 2.

Read more of this story at Slashdot.

The Open Document Format reached its 20th anniversary on May 1, marking two decades since OASIS approved the XML-based standard originally developed by Sun Microsystems from StarOffice code. Even as the format has seen adoption by several governments including the UK, India, and Brazil, plus organizations like NATO, Microsoft Office's proprietary formats remain the de facto standard. Microsoft countered ODF by developing Office Open XML, eventually getting it standardized through Ecma International. "ODF is much more than a technical specification: it is a symbol of freedom of choice, support for interoperability and protection of users from the commercial strategies of Big Tech," said Eliane Domingos, Chair of the Document Foundation, which oversees LibreOffice -- a fork created after Oracle acquired Sun.

Read more of this story at Slashdot.

Riot Games has reduced cheating in Valorant to under 1% of ranked games through its controversial kernel-level anti-cheat system Vanguard, according to the company's anti-cheat director Phillip Koskinas. The system enforces Windows security features like Trusted Platform Module and Secure Boot while preventing code execution in kernel memory. Beyond technical measures, Riot deploys undercover operatives who have infiltrated cheat development communities for years. "We've even gone as far as giving anti-cheat information to establish credibility," Koskinas told TechCrunch, describing how they target even "premium" cheats costing thousands of dollars. Riot faces increasingly sophisticated threats, including direct memory access attacks using specialized PCI Express hardware and screen reader cheats that use separate computers to analyze gameplay and control mouse movements. To combat repeat offenders, Vanguard fingerprints cheaters' hardware. Koskinas admits to deliberately slowing some enforcement: "To keep cheating dumb, we ban slower." The team also employs psychological warfare, publicly discrediting cheat developers and trolling known cheaters to undermine their credibility in gaming communities.

Read more of this story at Slashdot.

An anonymous reader quotes a report from the Associated Press: The European Union launched a drive on Monday to attract scientists and researchers to Europe with offers of grants and new policy plans, after the Trump administration froze U.S. government funding linked to diversity, equity and inclusion initiatives. "A few years ago, no one would have imagined that one of the biggest democracies in the world would cancel research programs under the pretext that the word diversity was in this program," French President Emmanuel Macron said at the "Choose Europe for Science" event in Paris. "No one would have thought that one of the biggest democracies in the world would delete with a stroke the ability of one researcher or another to obtain visas," Macron said. "But here we are." Taking the same stage at the Sorbonne University, European Commission President Ursula von der Leyen said that the EU's executive branch would set up a "super grant" program aimed at offering "a longer-term perspective to the very best" in the field. She said that 500 million euros ($566 million) will be put forward in 2025-2027 "to make Europe a magnet for researchers." It would be injected into the European Research Council, which already has a budget of more than 16 billion euros ($18 billion) for 2021-2027. Von der Leyen said that the 27-nation EU intends "to enshrine freedom of scientific research into law" with a new legal act. As "the threats rise across the world, Europe will not compromise on its principles," she said. Macron said that the French government would also soon make new proposals to beef up investment in science and research. [...] While not mentioning the Trump administration by name, von der Leyen said that it was "a gigantic miscalculation" to undermine free and open research. "We can all agree that science has no passport, no gender, no ethnicity, no political party," she said. "We believe that diversity is an asset of humanity and the lifeblood of science. It is one of the most valuable global assets and it must be protected." Macron said that science and research must not "be based on the diktats of the few." Macron said that Europe "must become a refuge" for scientists and researchers, and he said to those who feel under threat elsewhere: "The message is simple. If you like freedom, come and help us to remain free, to do research here, to help us become better, to invest in our future." Further reading: 75% of Scientists in Nature Poll Weigh Leaving US NASA, Yale, and Stanford Scientists Consider 'Scientific Exile'

Read more of this story at Slashdot.