NASA's Juno spacecraft made its closest flyby yet of Io, one of Jupiter's largest moons, sending back images of "sharp cliffs, edgy mountain peaks, lakes of pooled lava and even a volcanic plume," reports the New York Times. From the report: The Juno spacecraft, designed to study the origin and evolution of Jupiter, arrived at the planet in 2016. NASA extended the mission in 2021, and the orbiter has since captured photos of the Jovian moons Ganymede, Europa and most recently Io. [...] Juno conducted a number of more distant observations of Io in recent years. Its latest flyby occurred on Dec. 30, when the spacecraft came within 932 miles of the moon. The images captured during this visit were made with an instrument called JunoCam and are in visible wavelengths. They are some of the highest resolution views of Io's global structure. The mission's managers shared six images of Io on the mission's website, and members of the public have since uploaded digitally enhanced versions that highlight features on Io's surface. Mission scientists are already at work analyzing these images, searching for differences across Io's surface to learn how often its volcanoes erupt, how bright and hot those eruptions are and how the resulting lava flows. According to Dr. Bolton, the team will also compare Juno's images to older views of the Jovian moon to determine what has changed on Io over a variety of encounters. And they'll get a second set of data to work with in a month, when Juno completes another close flyby of the explosive world on Feb. 3.

Read more of this story at Slashdot.

Programmer James Lambert has been working on a demake of Valve's Portal puzzle game for the Nintendo N64. After several years of development, Portal 64: The First Slice is now out of beta with two-thirds of the game's test chambers available to play. PC Gamer reports: In the announcement video Lambert goes through some of the new features in the latest build, including a seriously impressive visual rework on the portal gun itself. The video also showcases just how much of Portal's feel this manages to successfully capture, in particular the mind-bending effects of observing rooms and Chell through the portals themselves. I once called this the most impressive homebrew game I've ever seen and, while admittedly the N64 nostalgia helps, I'd stick by that. While this is obviously the first slice (geddit) and there's more to come, it's an incredible achievement in its own right: The first 13 test chambers of the game all present-and-correct. Portal has 19 test chambers, and Portal: Still Alive (which unbelievably has never seen an official PC release) added a further 14, so Lambert's well on his way to completing a vanilla version of Portal 64. You can follow the Portal 64 project on YouTube and download the game here.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Interesting Engineering: Marking a significant advancement in medical science, the world's first partial heart transplant has achieved the expected outcome after over a year of research efforts. Carried out by Duke Health, the patient, a young individual, now exhibits functioning valves and arteries that are growing in tandem with the transplant, as initially expected by the medical team. In spring 2022, doctors carried out the procedure on a baby who needed a new heart valve. Before, they used non-living valves, which didn't grow with the child. This meant the child needed frequent replacements, and the surgeries had a 50 percent chance of being deadly. The new procedure avoids these problems, according to the team. Babies with serious heart valve problems face a tough challenge because there aren't any implants that can grow with them. So, these babies end up needing new implants over and over until they're big enough for an adult-sized valve. It's a problem that doesn't have a solution yet. Duke Health doctors, leading a study published in the Journal of the American Medical Association, discovered that the innovative valve collection method used in the partial heart transplant resulted in two properly functioning valves and arteries that are growing along with the child, resembling natural blood vessels. "This publication is proof that this technology works, this idea works, and can be used to help other children," said Joseph W. Turek, first author of the study and Duke's chief of pediatric cardiac surgery, in a statement. The research also notes that the new procedure requires less immunosuppressant medication, reducing potential long-term side effects. It also facilitates a "domino transplant" method, where one donor heart benefits multiple patients, potentially doubling the number of hearts available for children with heart disease by utilizing previously unused hearts and valves.

Read more of this story at Slashdot.

Generative AI models like Google Bard and GitHub Copilot are increasingly being used in various industries, but users often overlook their limitations, leading to serious errors and inefficiencies. Daniel Stenberg of curl and libcurl highlights a specific problem of AI-generated security reports: when reports are made to look better and to appear to have a point, it takes a longer time to research and eventually discard it. "Every security report has to have a human spend time to look at it and assess what it means," adds Stenberg. "The better the crap, the longer time and the more energy we have to spend on the report until we close it." The Register reports: The curl project offers a bug bounty to security researchers who find and report legitimate vulnerabilities. According to Stenberg, the program has paid out over $70,000 in rewards to date. Of 415 vulnerability reports received, 64 have been confirmed as security flaws and 77 have been deemed informative -- bugs without obvious security implications. So about 66 percent of the reports have been invalid. The issue for Stenberg is that these reports still need to be investigated and that takes developer time. And while those submitting bug reports have begun using AI tools to accelerate the process of finding supposed bugs and writing up reports, those reviewing bug reports still rely on human review. The result of this asymmetry is more plausible-sounding reports, because chatbot models can produce detailed, readable text without regard to accuracy. As Stenberg puts it, AI produces better crap. "A crap report does not help the project at all. It instead takes away developer time and energy from something productive. Partly because security work is considered one of the most important areas so it tends to trump almost everything else." As examples, he cites two reports submitted to HackerOne, a vulnerability reporting community. One claimed to describe Curl CVE-2023-38545 prior to actual disclosure. But Stenberg had to post to the forum to make clear that the bug report was bogus. He said that the report, produced with the help of Google Bard, "reeks of typical AI style hallucinations: it mixes and matches facts and details from old security issues, creating and making up something new that has no connection with reality." [...] Stenberg readily acknowledges that AI assistance can be genuinely helpful. But he argues that having a human in the loop makes the use and outcome of AI tools much better. Even so, he expects the ease and utility of these tools, coupled with the financial incentive of bug bounties, will lead to more shoddy LLM-generated security reports, to the detriment of those on the receiving end.

Read more of this story at Slashdot.

Qualcomm today unveiled a new Snapdragon XR2+ Gen 2 chipset, a single-chip architecture that will likely power Apple Vision Pro competitors from Meta, Samsung, Google and HTC, among others. ZDNet reports: Succeeding last year's XR2 Gen 2, the plus variant brings improved GPU and CPU frequency -- up 15% and 20% respectively, support for 4.3K per eye resolution at 90fps, and the ability for headsets to field 12 or more cameras with on-device AI capabilities. The latter allows equipped models to better track user movements and surrounding objects for more immersive (and harmonious) VR and MR experiences. As for efficiency gains, you'll still be getting the 50% improvement as the previous XR2 Gen 2 when stacked against the XR2 Gen 1 platform. Basically, there's no change on that front. "(Snapdragon XR2+ Gen 2) will take XR productivity and entertainment to the next level by bringing spectacularly clear visuals to use cases such as room-scale screens, life-size overlays and virtual desktops," said Hugo Swart, vice president and general manager of XR, Qualcomm Technologies, Inc, in a Thursday press release. Clearly, the new silicon is aimed at headsets that can do it all -- with feature parity to the $3,500 gorilla in the room, Apple's upcoming Vision Pro headset -- though Qualcomm says it'll be priced accessibly for manufacturers to build hardware around. How affordable will these competing wearables be? Your guess is as good as mine. But considering we've already gotten products like the $500 Meta Quest 3 fielding the slightly less capable XR2 Gen 2 chip, the future of XR may not be as expensive as it seems. The new Snapdragon XR2+ Gen 2 chipset is made in collaboration with Google and Samsung, both of which bring expertise in the Android ecosystem and developing mobile VR devices. The trio had announced plans to develop an XR platform back in February of 2023, likely in reaction to the then-rumored headset by Apple.

Read more of this story at Slashdot.

According to MoffettNathanson, the U.S. pay-TV industry had its worst-ever third quarter after losing about 900,000 subscribers. "That poor result, the research firm added, left the total pay-TV industry shrinking at a record pace of -7.3%, widened from a year-ago decline of -5.9%," reports Light Reading. "It also left pay-TV penetration of occupied households (including vMVPDs) at just 54.8% -- a level last seen in 1989, five years before the debut of DirecTV." From the report: Drilling down on Q3 results, traditional pay-TV providers (cable, telco and satellite) shed 1.97 million subscribers, widened from a loss of 1.94 million in the year-ago quarter. Within that category, US cable lost 1.10 million video subs in Q3, versus a loss of -1.09 million in the year-ago period. Satellite operators (Dish Network and DirecTV) lost 667,000 subs in Q3, versus -567,000 in the year-ago quarter. Telco TV providers lost 198,000 video subs in the period, an improvement when compared to a year-ago loss of -250,000 subs. vMVPDs, meanwhile, added 1.08 million in Q3, down from a year-ago gain of about 1.34 million. Despite those gains, vMVPDs recaptured only 21.7% of traditional pay-TV's subscriber losses in the period, according to MoffettNathanson. Meanwhile, YouTube TV continues to dominate the vMVPD category. MoffettNathanson estimates that YouTube TV added about 350,000 subs in Q3, extending its total to 7 million -- representing 40% of the vMVPD sector's 18 million subscriber total. "Based on our Q3 estimate, YouTube TV has now surpassed Dish Network [6.72 million satellite TV subs at the end of Q3] to become the country's fourth largest MVPD of any kind," Moffett noted. "At the current trajectory, YouTube TV should pass DirecTV for third place in less than a year."

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: ChatGPT is still no House, MD. While the chatty AI bot has previously underwhelmed with its attempts to diagnose challenging medical cases -- with an accuracy rate of 39 percent in an analysis last year -- a study out this week in JAMA Pediatrics suggests the fourth version of the large language model is especially bad with kids. It had an accuracy rate of just 17 percent when diagnosing pediatric medical cases. The low success rate suggests human pediatricians won't be out of jobs any time soon, in case that was a concern. As the authors put it: "[T]his study underscores the invaluable role that clinical experience holds." But it also identifies the critical weaknesses that led to ChatGPT's high error rate and ways to transform it into a useful tool in clinical care. With so much interest and experimentation with AI chatbots, many pediatricians and other doctors see their integration into clinical care as inevitable. [...] For ChatGPT's test, the researchers pasted the relevant text of the medical cases into the prompt, and then two qualified physician-researchers scored the AI-generated answers as correct, incorrect, or "did not fully capture the diagnosis." In the latter case, ChatGPT came up with a clinically related condition that was too broad or unspecific to be considered the correct diagnosis. For instance, ChatGPT diagnosed one child's case as caused by a branchial cleft cyst -- a lump in the neck or below the collarbone -- when the correct diagnosis was Branchio-oto-renal syndrome, a genetic condition that causes the abnormal development of tissue in the neck, and malformations in the ears and kidneys. One of the signs of the condition is the formation of branchial cleft cysts. Overall, ChatGPT got the right answer in just 17 of the 100 cases. It was plainly wrong in 72 cases, and did not fully capture the diagnosis of the remaining 11 cases. Among the 83 wrong diagnoses, 47 (57 percent) were in the same organ system. Among the failures, researchers noted that ChatGPT appeared to struggle with spotting known relationships between conditions that an experienced physician would hopefully pick up on. For example, it didn't make the connection between autism and scurvy (Vitamin C deficiency) in one medical case. Neuropsychiatric conditions, such as autism, can lead to restricted diets, and that in turn can lead to vitamin deficiencies. As such, neuropsychiatric conditions are notable risk factors for the development of vitamin deficiencies in kids living in high-income countries, and clinicians should be on the lookout for them. ChatGPT, meanwhile, came up with the diagnosis of a rare autoimmune condition. Though the chatbot struggled in this test, the researchers suggest it could improve by being specifically and selectively trained on accurate and trustworthy medical literature -- not stuff on the Internet, which can include inaccurate information and misinformation. They also suggest chatbots could improve with more real-time access to medical data, allowing the models to refine their accuracy, described as "tuning."

Read more of this story at Slashdot.

As spotted by X user bedros_p, Google appears to be preparing to introduce a paid upgrade for Bard Advanced, a "new, cutting-edge AI experience" announced in December that gives users access to Google's best models and capabilities. Android Police reports: According to the strings, you will be able to "Try Bard Advanced for 3 months, on us." After that test period, you will likely have to pay up for the service. A defunct link within the code suggests that it may be part of Google One, but it's not clear if Bard Advanced will be added to all tiers or only more expensive ones with more Google Drive storage. It's also possible that it will be an extra new tier in Google One. As a refresher, Google launched its most capable AI model yet in December 2023, called Gemini. The LLM is available in three tiers, including a Nano version capable of running on devices like phones and a Pro version currently powering Bard in the US. There is also a Gemini Ultra which isn't public just yet, but supposedly outperforms other LLMs in almost all metrics. Google says that this is the one that will power Bard Advanced.

Read more of this story at Slashdot.

Michael Larabel reports via Phoronix: AMD engineers are proposing an FPGA Subsystem User-Space Interface to overcome current limitations of the Linux kernel's FPGA manager subsystem. AMD-Xilinx engineers are proposing a new sysfs interface for the FPGA subsystem that allows for more user-space control over FPGAs. The suggested interface would handle FPGA configuration, driver probe/remove, bridges, Device Tree Overlay file support for re-programming an FPGA while the operating system is running, and other capabilities for user-space not currently presented by the mainline kernel. [...] This proposal from AMD hopes to standardize the FPGA subsystem user-space interface in a manner that is suitable for upstreaming into the mainline Linux kernel.

Read more of this story at Slashdot.

An anonymous reader quotes a report from the Los Angeles Times: The former Westside Pavilion, a long shuttered indoor mall, will be transformed into a UCLA biomedical research center aimed at tackling such towering challenges as curing cancer and preventing global pandemics, officials announced Wednesday. The sprawling three-story structure will be known as the UCLA Research Park and will house two multidisciplinary centers focusing on immunology and immunotherapy as well as quantum science and engineering. Establishment of the public-private research center is a coup for Southern California that "will cement California's global, economic, scientific and technical dominance into the 22nd century and beyond," said Gov. Gavin Newsom. The former owners of the mall, Hudson Pacific Properties Inc. and Macerich, said Wednesday that they sold the property to the Regents of the University of California for $700 million. By purchasing the former shopping center, UCLA saved several years of potential toil to build such a facility on campus. UCLA is the most-applied-to university in the nation, but its Westwood home is among the smallest of the nine UC undergraduate campuses, leaving it limited room for growth. The former mall sits on prime real estate in the heart of the Westside at Pico Boulevard and Overland Avenue, about two miles from the UCLA campus. The mall was owned by commercial developers who spent hundreds of millions of dollars to dramatically remake the old shopping center into an office complex intended to appeal to technology firms, which signed some of the biggest office leases in L.A.'s Silicon Beach before the pandemic. Google agreed to become the sole tenant and began paying rent last year yet never moved in. The interior is mostly unfinished, but is ready for UCLA to build out to its specifications in a process Newsom said would take about 40 months. The UCLA Research Park "will serve as a state of the art hub of research and innovation that will bring together academics, corporate partners, government agencies and startups to explore new areas of inquiry and achieve breakthroughs that serve the common good," UCLA Chancellor Gene Block said. In addition to flexible work areas, the former mall's 12-screen multiplex movie theater may be converted into lecture halls or performance spaces offering programming across the arts, humanities, sciences and social sciences, the chancellor's office said. One tenant of the research park will be the new California Institute for Immunology and Immunotherapy.

Read more of this story at Slashdot.

The research publisher Science announced today that all of its journals will begin using commercial software that automates the process of detecting improperly manipulated images. From a report: The move comes many years into our awareness that the transition to digital data and publishing has made it comically easy to commit research fraud by altering images. While the move is a significant first step, it's important to recognize the software's limitations. While it will catch some of the most egregious cases of image manipulation, enterprising fraudsters can easily avoid being caught if they know how the software operates. Which, unfortunately, we feel compelled to describe (and, to be fair, the company that has developed the software does so on its website). Much of the image-based fraud we've seen arises from a dilemma faced by many scientists: It's not a problem to run experiments, but the data they generate often isn't the data you want. Maybe only the controls work, or maybe the experiments produce data that is indistinguishable from controls. For the unethical, this doesn't pose a problem since nobody other than you knows what images come from which samples. It's relatively simple to present images of real data as something they're not. To make this concrete, we can look at data from a procedure called a western blot, which uses antibodies to identify specific proteins from a complex mixture that has been separated according to protein size. Typical western blot data looks like the image at right, with the darkness of the bands representing proteins that are present at different levels in different conditions.

Read more of this story at Slashdot.