Leading AI models can fix broken code, but they're nowhere near ready to replace human software engineers, according to extensive testing [PDF] by OpenAI researchers. The company's latest study put AI models and systems through their paces on real-world programming tasks, with even the most advanced models solving only a quarter of typical engineering challenges. The research team created a test called SWE-Lancer, drawing from 1,488 actual software fixes made to Expensify's codebase, representing $1 million worth of freelance engineering work. When faced with these everyday programming tasks, the best AI model â" Claude 3.5 Sonnet -- managed to complete just 26.2% of hands-on coding tasks and 44.9% of technical management decisions. Though the AI systems proved adept at quickly finding relevant code sections, they stumbled when it came to understanding how different parts of software interact. The models often suggested surface-level fixes without grasping the deeper implications of their changes. The research, to be sure, used a set of complex methodologies to test the AI coding abilities. Instead of relying on simplified programming puzzles, OpenAI's benchmark uses complete software engineering tasks that range from quick $50 bug fixes to complex $32,000 feature implementations. Each solution was verified through rigorous end-to-end testing that simulated real user interactions, the researchers said.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Bloomberg: Niantic, the company behind the 2016 hit Pokemon Go, is in talks to sell its video-game business to Saudi Arabia-owned Scopely, according to several people familiar with the discussions. A deal could be announced in coming weeks. The price being discussed is about $3.5 billion, according to one of the people. Any agreement would involve the Pokemon title as well as other mobile games, according to the people, who asked not to be identified because the discussions are private. There's no assurance an agreement will be reached.

Read more of this story at Slashdot.

Microsoft is reminding IT administrators that WSUS driver synchronization will be deprecated on April 18, 2025, urging them to transition to cloud-based update solutions like Windows Autopatch, Azure Update Manager, and Microsoft Intune. "For on-premises contexts, drivers will be available on the Microsoft Update catalog, but you won't be able to import them into WSUS," the company said in a Windows message center update on Tuesday. "You'll need to use any of the available alternative solutions, such as Device Driver Packages, or transition to cloud-based driver services for your organization, such as Microsoft Intune and Windows Autopatch." BleepingComputer reports: This reminder follows two other warnings issued since June 2024, announcing the deprecation of WSUS driver synchronization and encouraging customers to adopt Redmond's newer cloud-based driver services. The company also revealed in September 2024 that WSUS had been deprecated, but Microsoft added that it plans to keep publishing updates through the channel and maintain all existing capabilities. This announcement came after WSUS was listed on August 13 as one of the "features removed or no longer developed starting with Windows Server 2025." "Specifically, this means that we are no longer investing in new capabilities, nor are we accepting new feature requests for WSUS," Microsoft's Nir Froimovici said at the time. "However, we are preserving current functionality and will continue to publish updates through the WSUS channel. We will also support any content already published through the WSUS channel."

Read more of this story at Slashdot.

Late Tuesday afternoon, the White House confirmed that it would not proceed with laying off more than 1,000 probationary employees at NASA. "NASA had sought exemptions for all of these employees, who comprise about 6 percent of NASA's workforce," notes Ars Technica. "The Trump administration has sought to fire federal employees at several federal agencies who are in the 'probationary' period of their employment. This includes new hires within the last one or two years or long-time employees who have moved into or been promoted into a new position." From the report: It was not immediately clear why. A NASA spokesperson in Washington, DC, offered no comment on the updated guidance. Two sources indicated that it was plausible that private astronaut Jared Isaacman, whom President Trump has nominated to lead the space agency, asked for the cuts to be put on hold. Although this could not be confirmed, it seems reasonable that Isaacman would want to retain some control over where cuts at the agency are made. Firing all probationary employees -- which is the most expedient way to reduce the size of government -- is a blunt instrument. It whacks new hires that the agency may have recruited for key positions, as well as high performers who earned promotions. The reprieve in these terminations does not necessarily signal that NASA will escape significant budget or employment cuts in the coming months. The administration could still seek to terminate probationary employees. In addition, Ars reported earlier that directors at the agency's field centers have been told to prepare options for a "significant" reduction in force in the coming months. The scope of these cuts has not been defined, and it's likely they would need to be negotiated with Congress.

Read more of this story at Slashdot.

Greg was fighting with an academic CMS, and discovered that a file called write_helper.js was included on every page. It contained this single function:

function document_write(s) { document.write(s); }

Now, setting aside the fact that document.write is one of those "you probably shouldn't use this" functions, and is deprecated, one has to wonder what the point of this function is. Did someone really not like object-oriented style code? Did someone break the "." on their keyboard and just wanted to not have to copy/paste existing "."s?

It's the kind of function you expect to see that someone wrote but that isn't invoked anywhere, and you'd almost be correct. This function, in a file included on every page, is called once and only once.

More like the wrong helper, if we're being honest.

[Advertisement] Picking up NuGet is easy. Getting good at it takes time. Download our guide to learn the best practice of NuGet for the Enterprise.

An anonymous reader quotes a report from Ars Technica: Astronomers have detected over 5,800 confirmed exoplanets. One extreme class is ultra-hot Jupiters, of particular interest because they can provide a unique window into planetary atmospheric dynamics. According to a new paper published in the journal Nature, astronomers have mapped the 3D structure of the layered atmosphere of one such ultra-hot Jupiter-size exoplanet, revealing powerful winds that create intricate weather patterns across that atmosphere. A companion paper published in the journal Astronomy and Astrophysics (PDF) reported on the unexpected identification of titanium in the exoplanet's atmosphere as well. [...] This latest research relied on observational data collected by the European South Observatory's (ESO) Very Large Telescope, specifically, a spectroscopic instrument called ESPRESSO that can process light collected from the four largest VLT telescope units into one signal. The target exoplanet, WASP-121b -- aka Tylos -- is located in the Puppis constellation about 900 light-years from Earth. One year on Tylos is equivalent to just 30 hours on Earth, thanks to the exoplanet's close proximity to its host star. Since one side is always facing the star, it is always scorching, while the exoplanet's other side is significantly colder. Those extreme temperature contrasts make it challenging to figure out how energy is distributed in the atmospheric system, and mapping out the 3D structure can help, particularly with determining the vertical circulation patterns that are not easily replicated in our current crop of global circulation models, per the authors. For their analysis, they combined archival ESPRESSO data collected on November 30, 2018, with new data collected on September 23, 2023. They focused on three distinct chemical signatures to probe the deep atmosphere (iron), mid-atmosphere (sodium), and shallow atmosphere (hydrogen). "What we found was surprising: A jet stream rotates material around the planet's equator, while a separate flow at lower levels of the atmosphere moves gas from the hot side to the cooler side. This kind of climate has never been seen before on any planet," said Julia Victoria Seidel of the European Southern Observatory (ESO) in Chile and the Observatoire de la Cote d'Azur in France. "This planet's atmosphere behaves in ways that challenge our understanding of how weather works -- not just on Earth, but on all planets. It feels like something out of science fiction."

Read more of this story at Slashdot.

Nvidia has confirmed on its forums that the RTX 50 series GPUs no longer support 32-bit PhysX. Tom's Hardware reports: As far as we know, there are no 64-bit games with integrated PhysX technology, thus terminating the tech entirely on RTX 50 series GPUs and newer. RTX 40 series and older will still be able to run 32-bit CUDA applications and thus PhysX, but regardless, the technology is now officially retired, starting with Blackwell. [...] The only way now to run PhysX on RTX 50 series GPUs (or newer) is to install a secondary RTX 40 series or older graphics card and slave it to PhysX duty in the Nvidia control panel. As far as we are aware, Nvidia has not disabled this sort of functionality. But the writing is on the wall for PhysX, and we doubt there will be any future games that attempt to use the API.

Read more of this story at Slashdot.

Valve removed the game PirateFi from Steam after discovering it was laced with the Vidar infostealer malware, designed to steal sensitive user data such as passwords, cookies, cryptocurrency wallets, and more. TechCrunch reports: Marius Genheimer, a researcher who analyzed the malware and works at SECUINFRA Falcon Team, told TechCrunch that judging by the command and control servers associated with the malware and its configuration, "we suspect that PirateFi was just one of multiple tactics used to distribute Vidar payloads en masse." "It is highly likely that it never was a legitimate, running game that was altered after first publication," said Genheimer. In other words, PirateFi was designed to spread malware. Genheimer and colleagues also found that PirateFi was built by modifying an existing game template called Easy Survival RPG, which bills itself as a game-making app that "gives you everything you need to develop your own singleplayer or multiplayer" game. The game maker costs between $399 and $1,099 to license. This explains how the hackers were able to ship a functioning video game with their malware with little effort. According to Genheimer, the Vidar infostealing malware is capable of stealing and exfiltrating several types of data from the computers it infects, including: passwords from the web browser autofill feature, session cookies that can be used to log in as someone without needing their password, web browser history, cryptocurrency wallet details, screenshots, and two-factor codes from certain token generators, as well as other files on the person's computer.

Read more of this story at Slashdot.

Meta's WhatsApp messaging service has surpassed 45 million users, earning the designation of a "Very Large Online Platform" under the EU's Digital Services Act. Bloomberg reports: WhatsApp's open channels, which are feeds affiliated with news outlets or public figures that under the DSA are comparable to a social network, averaged about 46.8 million monthly average users in the second half of 2024, Meta said in a filing on Feb. 14 that hasn't previously been reported. [...] The DSA content moderation rulebook imposes stricter requirements on very large online platforms, defined as those whose EU-based monthly active users exceed 45 million. Users of WhatsApp's core messaging feature do not count toward the designation under the DSA. The commission would still need to rule that WhatsApp should be included in the more regulated tier. Under the DSA, very large online platforms must carry out risk assessments on the spread of illegal or harmful content, and put in place a mitigation strategy. Fines under the DSA can reach as much as 6% of a company's annual global sales. The DSA requires platforms to disclose user numbers every six months. Messaging service Telegram also published an update this week, saying that monthly EU users of its public channels are "significantly fewer than 45 million."

Read more of this story at Slashdot.

Google has gained permission to sell its e-books and audiobooks directly to customers through its iOS app, Google Play Books. From a report: While iOS apps today can offer access to content previously purchased elsewhere, like e-books bought via a website, developers have to request a specific exception to link their iOS app's users to the company's own website to make purchases. According to a brief post on Google's blog, users will now be able to click on a new "Get book" button in the Google Play Books iOS app which will take them to the Google Play website to complete their e-book or audiobook purchase. From there, users will be able to see their recently opened book listings and complete a purchase using their Google Account and saved payment information. By processing the transaction on its own website, Google can avoid paying Apple a commission (generally 30%) on in-app purchases of digital content.

Read more of this story at Slashdot.

Former OpenAI CTO Mira Murati has launched Thinking Machines Lab with several leaders from OpenAI on board, including John Schulman, Barrett Zoph, and Jonathan Lachman. Their mission is "to make AI systems more widely understood, customizable, and generally capable," with a commitment to publishing technical research and code. The Verge reports: In a press release shared with The Verge, the company suggests that it's building products that help humans work with AI, rather than fully autonomous systems. "We're building a future where everyone has access to the knowledge and tools to make AI work for their unique needs and goals," says the press release.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Bloomberg: HP will acquire assets from Humane, the maker of a wearable Ai Pin introduced in late 2023, for $116 million. The deal will include the majority of Humane's employees in addition to its software platform and intellectual property, the company said Tuesday. It will not include Humane's Ai pin device business, which will be wound down, an HP spokesperson said. Humane's team, including founders Imran Chaudhri and Bethany Bongiorno, will form a new division at HP to help integrate artificial intelligence into the company's personal computers, printers and connected conference rooms, said Tuan Tran, who leads HP's AI initiatives. Chaudhri and Bongiorno were design and software engineers at Apple before founding the startup. [...] Tran said he was particularly impressed with aspects of Humane's design, such as the ability to orchestrate AI models running both on-device and in the cloud. The deal is expected to close at the end of the month, HP said. "There will be a time and place for pure AI devices," Tran said. "But there is going to be AI in all our devices -- that's how we can help our business customers be more productive."

Read more of this story at Slashdot.

Christoph Hellwig continues to voice strong opposition to Rust in the Linux kernel, arguing that its introduction creates fragmentation, unclear language guidelines, and additional burdens on maintainers. He also says Linus Torvalds has privately stated he will override objections to Rust code, effectively making its adoption inevitable. Phoronix's Michael Larabel has the latest: The latest on Hellwig's perspective of Rust code within the Linux kernel is below. Some interesting insight from a dissenting view. The thread in full can be found on the Rust for Linux mailing list. [Here's an excerpt from the thread:] "I don't think having a web page in any form is useful. If you want it to be valid it has to be in the kernel tree and widely agreed on. It also states factually incorrect information. E.g. 'Some subsystems may decide they do not want to have Rust code for the time being, typically for bandwidth reasons. This is fine and expected.' while Linus in private said that he absolutely is going to merge Rust code over a maintainers objection. (He did so in private in case you are looking for a reference). So as of now, as a Linux developer or maintainer you must deal with Rust if you want to or not. [...] Right now the rules is Linus can force you whatever he wants (it's his project obviously) and I think he needs to spell that out including the expectations for contributors very clearly."

Read more of this story at Slashdot.

An anonymous reader quotes a report from Gizmodo: Protector, an app that lets you book armed goons the same way you'd call for an Uber, is having a viral moment. The app started doing the rounds on social media after consultant Nikita Bier posted about it on X. Protector lets the user book armed guards on demand. Right now it's only available in NYC and LA. According to its marketing, every guard is either "active duty or retired law enforcement and military." Every booking comes with a motorcade and users get to select the number of Escalades that'll be joining them as well as the uniforms their hired goons will wear. Protector is currently "#7 in Travel" on Apple's App Store. It's not available for people who use Android devices. [...] The marketing for Protector, which lives on its X account, is surreal. A series of robust and barrel-chested men in ill-fitting black suits deliver their credentials to the camera while sitting in front of a black background. They're all operators. They describe careers in SWAT teams and being deployed to war zones. They show vanity shots of themselves kitted out in operator gear. All of them have a red lapel pin bearing the symbol of Protector. If the late UnitedHealthcare CEO had used Protector, he might still be alive today, suggests Protector in its marketing materials. A video on X shows "several fantasy versions of the assassination where a Protector is on hand to prevent the assassin from killing the CEO," reports Gizmodo. The app is a product from parent company Protector Security Solutions, which was founded by Nick Sarath, a former product designer at Meta.

Read more of this story at Slashdot.

An anonymous reader shares a report: U.S. personal injury law firm Morgan & Morgan sent an urgent email this month to its more than 1,000 lawyers: Artificial intelligence can invent fake case law, and using made-up information in a court filing could get you fired. A federal judge in Wyoming had just threatened to sanction two lawyers at the firm who included fictitious case citations in a lawsuit against Walmart. One of the lawyers admitted in court filings last week that he used an AI program that "hallucinated" the cases and apologized for what he called an inadvertent mistake. AI's penchant for generating legal fiction in case filings has led courts around the country to question or discipline lawyers in at least seven cases over the last two years, and created a new high-tech headache for litigants and judges, Reuters found. The Walmart case stands out because it involves a well-known law firm and a big corporate defendant. But examples like it have cropped up in all kinds of lawsuits since chatbots like ChatGPT ushered in the AI era, highlighting a new litigation risk.

Read more of this story at Slashdot.

Public interest groups on Tuesday asked the full 6th U.S. Circuit Court of Appeals to reconsider a ruling that the Federal Communications Commission lacked legal authority to reinstate landmark net neutrality rules. From a report: The decision by a three-judge panel blocked the FCC under then President Joe Biden that had sought to reinstate the open internet rules implemented in 2015 but later repealed by the agency under President Donald Trump. The groups -- Free Press, Public Knowledge, Open Technology Institute and the Benton Institute for Broadband & Society -- argue the appeals court decision conflicts with an earlier decision by another court. The groups said the case centers on the FCC's decades-long effort to prevent broadband internet providers "from abusing their gatekeeping power, in furtherance of the providers' economic or political interests, to constrain their users' access to third-party websites."

Read more of this story at Slashdot.

Lloyds Banking Group is assessing the skills of thousands of technology staffers in the UK to determine whether they can keep working at the bank once it upgrades its technology [alternative source]. Bloomberg: The British lender, which is carrying out a multiyear overhaul of its systems, put these workers on notice this month that they are at risk of losing their jobs and will be required to reapply for new positions across the bank, according to people familiar with the matter. In a company town hall last week, executives informed those staffers that they were in the process of assessing their technical skills based on a test they took last year to determine where, if anywhere, they can be placed within Lloyds, the people said, asking not to be named discussing non-public information.

Read more of this story at Slashdot.

A new report released by Cisco finds that 97% of CEOs surveyed are planning AI integration. Similarly, 92% of companies recently surveyed by McKinsey plan to invest more in generative AI over the next three years. Fortune: To that end, many companies are seeking tech-savvy finance talent, according to a new report by software company Datarails. The researchers analyzed 6,000 job listings within the CFO's office -- CFO, controller, financial planning and analysis (FP&A), and accountant -- advertised on job search websites including LinkedIn, Glassdoor, Indeed, Job2Careers, and ZipRecruiter. Of the 1,000 job listings for CFOs in January 2025, 27% included AI in the job description. This compares to 8% mentions of AI in 1,000 CFO job listings at the same time last year. Take, for example, Peaks Healthcare Consulting which required a CFO candidate to "continuously learn and integrate AI to improve financial processes and decision making," Datarails notes in the report. Regarding FP&A professionals, in January 2025, 35% of analyst roles mentioned AI competency as a requirement, compared to 14% in January 2024, according to the report.

Read more of this story at Slashdot.

Chinese AI startup DeepSeek has updated its business registry information with key changes to personnel and operational scope, signaling a shift towards monetizing its cost-efficient-yet-powerful large language models. From a report: The Hangzhou-based firm's updated business scope includes "internet information services," according to business registry service Tianyancha. The move is the first sign of DeepSeek's desire to monetise its popular technology, according to Zhang Yi, founder and chief analyst at consultancy iiMedia. With eyes on developing a business model, DeepSeek intends to shift away from being purely focused on research and development, Zhang added. "The move reflects that for a company like DeepSeek, which managed to accumulate technology and develop a product, monetisation is becoming a necessary next step," Zhang said. DeepSeek's previous business scope said it engages in engineering and AI software development, among others, hinting at a more research-driven approach.

Read more of this story at Slashdot.

An anonymous reader shares a report: Later this month, Intuitive Machines, the private company behind the first commercial lander that touched down on the moon, will launch a second lunar mission from NASA's Kennedy Space Center. The plan is to deploy a lander, a rover, and hopper to explore a site near the lunar south pole that could harbor water ice, and to put a communications satellite on lunar orbit. But the mission will also bring something that's never been installed on the moon or anywhere else in space before -- a fully functional 4G cellular network. Point-to-point radio communications, which need a clear line of sight between transmitting and receiving antennas, have always been a backbone of both surface communications and the link back to Earth, starting with the Apollo program. Using point-to-point radio in space wasn't much of an issue in the past because there never have been that many points to connect. Usually, it was just a single spacecraft, a lander, or a rover talking to Earth. And they didn't need to send much data either. "They were based on [ultra high frequency] or [very high frequency] technologies connecting a small number of devices with relatively low data throughput," says Thierry Klein, president of Nokia Bell Labs Solutions Research, which was contracted by NASA to design a cellular network for the moon back in 2020.

Read more of this story at Slashdot.