Early analysis of the near-Earth asteroid Bennu has revealed unexpected evidence of magnesium-sodium phosphate, suggesting Bennu might have originated from a primitive ocean world. Space.com reports: On Earth, magnesium-sodium phosphate can be found in certain minerals and geological formations, as well as within living organisms where it is present in various biochemical processes and is a component of bone and teeth. According to a NASA press release, however, its presence on Bennu surprised the research team because it wasn't seen in the OSIRIS-REx probe's remote sensing data prior to sample collection. The team says its presence "hints that the asteroid could have splintered off from a long-gone, tiny, primitive ocean world." "The presence and state of phosphates, along with other elements and compounds on Bennu, suggest a watery past for the asteroid," said Lauretta. "Bennu potentially could have once been part of a wetter world. Although, this hypothesis requires further investigation."

The OSIRIS-REx spacecraft obtained a sample of Bennu's regolith on October 20, 2020 using its Touch-and-Go Sample Acquisition Mechanism (TAGSAM), which comprises a specialized sampler head situated on an articulated arm. Bennu is a small B-type asteroid, which are relatively uncommon carbonaceous asteroids. "[Bennu] was selected as the mission target in part because telescopic observations indicated a primitive, carbonaceous composition and water-bearing minerals," stated the team in their paper. [...] Further analysis on the samples revealed the prevailing component of the regolith sample is magnesium-bearing phyllosilicates, primarily serpentine and smectite -- types of rock typically found at mid-ocean ridges on Earth. A comparison of these serpentinites with their terrestrial counterparts provides possible insights into Bennu's geological past. "Offering clues about the aqueous environment in which they originated," wrote the team.

While Bennu's surface may have been altered by water over time, it still preserves some of the ancient characteristics scientists believe were present during the early solar system's days. Bennu's surface materials still contain some original features from the cloud of gas and dust from which our solar system's planets formed -- known as the protoplanetary disk. The team's study also confirmed the asteroid is rich in carbon, nitrogen and some organic compounds -- all of which, in addition to the magnesium phosphate, are essential components for life as we know it on Earth.

BrianFagioli writes: Canonical has introduced a new service enabling the creation of custom distroless Docker images under its "Everything LTS" program. This initiative allows customers to include any open-source software in their Docker images, regardless of whether it is packaged in Ubuntu, with a security maintenance commitment of up to 12 years. [...] This expansion of the Ubuntu Pro offering incorporates numerous new open-source components, especially current AI/ML tools, maintained directly from the source rather than as traditional 'deb' packages. This approach aims to minimize the attack surface of containers, thereby enhancing security and aiding compliance with various regulatory standards such as FIPS, FedRAMP, EU Cyber Resilience Act, FCC U.S. Cyber Trust Mark, and DISA-STIG.

Microsoft has received a thumbs-up from iFixit, with a provisional 8 out of 10 for repairability on its latest Surface Pro and Laptop devices. From a report: Despite some issues with software recovery, the devices have been built for hardware repairability. It is quite the turnaround from the days of the first iteration of the Surface Laptop, in which the iFixit team was forced to use a scalpel to get into the device. "This is definitely not going back together without a roll of duct tape," the team observed during the 2017 teardown. In comparison, the team described Microsoft's latest laptop as "an astonishingly repair friendly device."

Where once there might have been glue or fragile clips, there are now screws and even QR codes linking to the service manuals (made available on release day, according to iFixit). Stripping the device is a breeze, assuming the correct tools are used. Microsoft has helpfully provided "Wayfinders" to indicate the type and quantity of screws being used to secure components, meaning that a repairer could even do without the online guides when pulling the hardware apart.

In 2022, Red Hat announced plans to extend RHEL to the automotive industry through Red Hat In-Vehicle Operating System (providing automakers with an open and functionally-safe platform). And this week Red Hat announced it achieved ISO 26262 ASIL-B certification from exida for the Linux math library (libm.so glibc) — a fundamental component of that Red Hat In-Vehicle Operating System.

From Red Hat's announcement: This milestone underscores Red Hat's pioneering role in obtaining continuous and comprehensive Safety Element out of Context certification for Linux in automotive... This certification demonstrates that the engineering of the math library components individually and as a whole meet or exceed stringent functional safety standards, ensuring substantial reliability and performance for the automotive industry. The certification of the math library is a significant milestone that strengthens the confidence in Linux as a viable platform of choice for safety related automotive applications of the future...

By working with the broader open source community, Red Hat can make use of the rigorous testing and analysis performed by Linux maintainers, collaborating across upstream communities to deliver open standards-based solutions. This approach enhances long-term maintainability and limits vendor lock-in, providing greater transparency and performance. Red Hat In-Vehicle Operating System is poised to offer a safety certified Linux-based operating system capable of concurrently supporting multiple safety and non-safety related applications in a single instance. These applications include advanced driver-assistance systems (ADAS), digital cockpit, infotainment, body control, telematics, artificial intelligence (AI) models and more. Red Hat is also working with key industry leaders to deliver pre-tested, pre-integrated software solutions, accelerating the route to market for SDV concepts.
"Red Hat is fully committed to attaining continuous and comprehensive safety certification of Linux natively for automotive applications," according to the announcement, "and has the industry's largest pool of Linux maintainers and contributors committed to this initiative..."

Or, as Network World puts it, "The phrase 'open source for the open road' is now being used to describe the inevitable fit between the character of Linux and the need for highly customizable code in all sorts of automotive equipment."

The Information: Apple has told at least one supplier that it has suspended work on its next high-end Vision headset, an employee at a manufacturer that makes key components for the Vision Pro said. The pullback comes as analysts and supply chain partners have flagged slowing sales of the $3,500 device. The company is still working on releasing a more affordable Vision product with fewer features before the end of 2025, the person involved in its supply chain and a person involved in the manufacturing of the headsets said.

Apple originally planned to divide its Vision line into two models, similar to the standard and Pro versions of the iPhone, according to people involved in its supply chain and former Apple employees who worked on the devices. Apple's decision to halt work on the next version of its high-end headset is the latest example of the company reshuffling priorities. Apple has ramped up work on AI-powered features while paring back money-losing projects like its self-driving car, which it canceled earlier this year after spending nearly a decade on development. Augmented reality is one of Apple's biggest bets. The company aims to eventually replace the iPhone with lightweight glasses, and the Vision Pro is the first step in building consumer and developer interest in that effort.

A Zillow listing for a $2.4 million house in a Dallas suburb is grabbing attention for its 5,786-square-foot data center with immersion cooling tanks, massive server racks, and two separate power grids. Tom's Hardware reports: With a brick exterior, cute paving, and mini-McMansion arch stylings, the building certainly looks to be a residential home for the archetypal Texas family. Prospective home-buyers will thus be disappointed by the 0 bedroom, 1 bathroom setup, which becomes a warehouse-feeling office from the first step inside where you are met with a glass-shielded reception desk in a white-brick corridor. The "Crypto Collective" branding betrays the former life of the unit, which served admirably as a crypto mining base.

The purchase of the "upgraded turnkey Tier 2 Data Center" will include all of its cooling and power infrastructure. Three Engineered Fluids "SLICTanks," single-phase liquid immersion cooling tanks for use with dielectric coolant, will come with pumps and a 500kW dry cooler. The tanks are currently filled with at least 80 mining computers visible from the photos, though the SLICTanks can be configured to fit more machines. Also visible in proximity to the cooling array is a deep row of classic server racks and a staggering amount of networking.

The listing advertises a host of potential uses for future customers, from "AI services, cloud hosting, traditional data center, servers or even Bitcoin Mining". Also packed into the 5,786 square feet of real estate is two separate power grids, 5 HVAC units, a hefty amount of four levels of warehouse-style storage aisles, a lounge/office space, and a fully-paved backyard. In other good news, its future corporate residents will not have an HOA to deal with, and will only be 20 minutes outside of the heart of Dallas, sitting just out of earshot of two major highways.

"SpaceX has received FCC clearance to operate a mysterious 'wireless module' device," PC Magazine reported earlier this week, speculating that the device "might be a new Starlink router." On Tuesday, the FCC issued an equipment authorization for the device, which uses the 2.4GHz and 5GHz Wi-Fi radio bands. A document in SpaceX's filing also says it features antennas along with Wi-Fi chips apparently from MediaTek. Another document calls the device by the codename "UTW-231," and defines it as a "wireless router" supporting IEEE 802.11b/g/n/ax for Wi-Fi 6 speeds up to 1,300Mbps. But perhaps the most interesting part is an image SpaceX attached, which suggests the router is relatively small and can fit in a person's open hand.... SpaceX CEO Elon Musk has said the "Starlink mini" dish is slated to arrive later this year and that it's small enough to fit in a backpack...

On Wednesday, PCMag also spotted the official Starlink.com site referencing the name "Mini" in a specification page for the satellite internet system.
Today saw some interesting speculation on the unoffical "Starlink Hardware" blog (written by Noah Clarke, who has a degree in electronics). Clarke guesses the product "will be aimed at portable use cases, such as camping, RV's, vans, hiking... designed to be easy to store, transport, and deploy". But he also notes Starlink updated their app today, with a new shopping page showing what he believes the upcoming product will look like. ("Very similar to the Standard dish, just smaller. It has a similar shape, and even a kickstand.") If you go into developer mode and play around with the Mini network settings, you notice something interesting. There is no separate router. Devices are connected to the dish itself... I'm guessing that, in order to make the Mini as portable as possible, Starlink decided it was best to simplify the system and limit the number of components.

There are more Wifi details that have been revealed, and that is mesh compatibility. For those of you that might be interested in using the Mini at home, or for larger events where you need additional Wifi coverage, the Mini's built-in router will be compatible with Starlink mesh. You'll be able to wirelessly pair another Starlink router to the Mini.

"Titanium that was distributed with fake documentation has been found in commercial Boeing and Airbus jets," reports CNN. America's Federal Aviation Administration is now investigating whether those components pose a safety hazard to the public," along with the manufacturers of the aircraft and supplier Spirit AeroSystems.

"A parts supplier found small holes in the material from corrosion," the New York Times reported Friday: Boeing and Airbus both said their tests of affected materials so far had shown no signs of problems.

Boeing said it directly purchased most of the titanium used in its plane production, so most of its supply was unaffected. "This industrywide issue affects some shipments of titanium received by a limited set of suppliers, and tests performed to date have indicated that the correct titanium alloy was used," Boeing said in a statement. "To ensure compliance, we are removing any affected parts on airplanes prior to delivery. Our analysis shows the in-service fleet can continue to fly safely."

Liam Proven reports via The Register: The latest version of the systemd init system is out, with the openly confrontational tag line: "Available soon in your nearest distro, now with 42 percent less Unix philosophy." As Lennart Poettering's announcement points out, this is the first version of systemd whose version number is a nine-bit value. Version 256, as usual, brings in a broad assortment of new features, but also turns off some older features that are now considered deprecated. For instance, it won't run under cgroups version 1 unless forced.

Around since 2008, cgroups is a Linux kernel containerization mechanism originally donated by Google, as The Reg noted a decade ago. Cgroups v2 was merged in 2016 so this isn't a radical change. System V service scripts are now deprecated too, as is the SystemdOptions EFI variable. Additionally, there are some new commands and options. Some are relatively minor, such as the new systemd-vpick binary, which can automatically select the latest member of versioned directories. Before any OpenVMS admirers get excited, no, Linux does not now support versions on files or directories. Instead, this is a fresh option that uses a formalized versioning system involving: "... paths whose trailing components have the .v/ suffix, pointing to a directory. These components will then automatically look for suitable files inside the directory, do a version comparison and open the newest file found (by version)."

The latest function, which The Reg FOSS desk suspects will ruffle some feathers, is a whole new command, run0, which effectively replaces the sudo command as used in Apple's macOS and in Ubuntu ever since the first release. Agent P introduced the new command in a Mastodon thread. He says that the key benefit is that run0 doesn't need setuid, a basic POSIX function, which, to quote its Linux manual page, "sets the effective user ID of the calling process." [...] Another new command is importctl, which handles importing and exporting both block-level and file-system-level disk images. And there's a new type of system service called a capsule, and "a small new service manager" called systemd-ssh-generator, which lets VMs and containers accept SSH connections so long as systemd can find the sshd binary -- even if no networking is available. The release notes are available here.

cusco writes: Indian space startup Agnikul used a 3-D printer from German company EOS to print an engine out of inconel, a high-performance nickel-chromium alloy, in one solid piece over the course of roughly 72 hours. While other companies like Relativity Space and Rocket Lab are using 3-D printers extensively, Agnikul's engine is unique in being printed in one go, rather than as multiple components that need to be stitched together. This approach significantly speeds up manufacturing time.

The single-engine technology demonstration rocket produced 6 kilonewtons of thrust and reached an altitude of 6.5 kilometers before splashing down into the ocean. The launch vehicle used was about 6 meters tall with a single engine, making it roughly equivalent to the second stage of the company's planned commercial product, Agnibaan. Agnibaan will be a two-stage rocket, 18 meters tall, featuring eight engines in total, and capable of carrying a 300-kilogram payload to an altitude of around 700 km. The company believes that their 3D printing approach opens the door to providing low-cost, "on-demand" launch services to operators of small satellites.

IEEE Spectrum adds: Assembling the rest of the rocket and integrating the engine took roughly two weeks. The company says that opens the door to providing low-cost, "on-demand" launch services to operators of small satellites, which otherwise need to wait for a ride share on a bigger rocket. The big challenge now will be going from a single engine to a cluster of seven on Agnibaan's first stage, says cofounder and CEO Srinath Ravichandran. This raises all kinds of challenges, from balancing thrust across the engines at lift-off to managing engine plume interactions when the engines gimbal to alter the trajectory. "But these are problems that people have figured out," he says. "We believe that we should just be able to fine-tune it for our mission and go." The company is currently building facilities to carry out ground tests of engine clusters, says Ravichandran, and is targeting its first orbital launch for this time next year.

Intel has begun ferrying around 20 "super loads" across Ohio for the construction of its new $28 billion Ohio One Campus. The extensive planning and coordination required for these shipments are expected to cause road closures and delays during the nine days of transport. Tom's Hardware reports: Intel's new campus coming to New Albany, OH, is in heavy construction, and around 20 super loads are being ferried across Ohio's roads by the Ohio Department of Transportation after arriving at a port of the Ohio River via barge. Four of these loads, including the one hitting the road now, weigh around 900,000 pounds -- that's 400 metric tons, or 76 elephants. The super loads were first planned for February but were delayed due to the immense planning workload. Large crowds are estimated to accumulate on the route, potentially slowing it even further.

Intel's 916,000-pound shipment is a "cold box," a self-standing air-processor structure that facilitates the cryogenic technology needed to fabricate semiconductors. The box is 23 feet tall, 20 feet wide, and 280 feet long, nearly the length of a football field. The immense scale of the cold box necessitates a transit process that moves at a "parade pace" of 5-10 miles per hour. Intel is taking over southern Ohio's roads for the next several weeks and months as it builds its new Ohio One Campus, a $28 billion project to create a 1,000-acre campus with two chip factories and room for more. Calling it the new "Silicon Heartland," the project will be the first leading-edge semiconductor fab in the American Midwest, and once operational, will get to work on the "Angstrom era" of Intel processes, 20A and beyond. The Ohio Department of Transportation has shared a timetable for how long this process will take.

An anonymous reader shares a report: Intel's fastest processors have included hyperthreading, a technique that lets more than one thread run on a single CPU core, for over 20 years -- and it's used by AMD (which calls it "simultaneous multi-threading") as well. But you won't see a little "HT" on the Intel sticker for any Lunar Lake laptops, because none of them use it. Hyperthreading will be disabled on all Lunar Lake CPU cores, including both performance and efficiency cores. Why? The reason is complicated, but basically it's no longer needed. The performance cores or P-Cores on the new Lunar Lake series are 14 percent faster than the same cores on the previous-gen Meteor Lake CPUs, even with the multi-thread-processing of hyperthreading disabled.

Turning on the feature would come at too high a power cost, and Lunar Lake is all about boosting performance while keeping laptops in this generation thin, light, and long-lasting. That means maximizing single-thread performance -- the most relevant to users who are typically focusing on one task at a time, as is often the case for laptops -- in terms of surface area, to improve overall performance per watt. Getting rid of the physical components necessary for hyperthreading just makes sense in that context.

Palmer Luckey, founder of Oculus VR and Anduril, has launched the ModRetro Chromatic, a retro gaming handheld that plays original Game Boy and Game Boy Color cartridges. The device boasts a magnesium alloy case, sapphire crystal screen, and a pixel-perfect IPS display. It comes bundled with a licensed copy of Tetris and supports Link Cable multiplayer. Luckey suggested that the Chromatic is the most authentic way to play Game Boy games, with custom components designed to provide the ultimate gaming experience. The handheld is available for preorder and will ship during the 2024 holiday season.

"A Swiss biocomputing startup has launched an online platform that provides remote access to 16 human brain organoids," reports Tom's Hardware: FinalSpark claims its Neuroplatform is the world's first online platform delivering access to biological neurons in vitro. Moreover, bioprocessors like this "consume a million times less power than traditional digital processors," the company says. FinalSpark says its Neuroplatform is capable of learning and processing information, and due to its low power consumption, it could reduce the environmental impacts of computing. In a recent research paper about its developments, FinalSpakr claims that training a single LLM like GPT-3 required approximately 10GWh — about 6,000 times greater energy consumption than the average European citizen uses in a whole year. Such energy expenditure could be massively cut following the successful deployment of bioprocessors.

The operation of the Neuroplatform currently relies on an architecture that can be classified as wetware: the mixing of hardware, software, and biology. The main innovation delivered by the Neuroplatform is through the use of four Multi-Electrode Arrays (MEAs) housing the living tissue — organoids, which are 3D cell masses of brain tissue...interfaced by eight electrodes used for both stimulation and recording... FinalSpark has given access to its remote computing platform to nine institutions to help spur bioprocessing research and development. With such institutions' collaboration, it hopes to create the world's first living processor.
FinalSpark was founded in 2014, according to Wikipedia's page on wetware computing. "While a wetware computer is still largely conceptual, there has been limited success with construction and prototyping, which has acted as a proof of the concept's realistic application to computing in the future."

Thanks to long-time Slashdot reader Artem S. Tashkinov for sharing the article.

Last than two weeks after it was announced, "Windows enthusiasts have managed to crack Microsoft's flagship AI-powered Recall feature to run on unsupported hardware," reports The Verge. From the report: Recall leverages local AI models on new Copilot Plus PCs to run in the background and take snapshots of anything you've done or seen on your PC. You then get a timeline you can scrub through and the ability to search for photos, documents, conversations, or anything else on your PC. Microsoft positioned Recall as needing the very latest neural processing units (NPU) on new PCs, but you can actually get it running on older Arm-powered hardware.

Windows watcher Albacore has created a tool called Amperage, which enables Recall on devices that have an older Qualcomm Snapdragon chip, Microsoft's SQ processors, or an Ampere chipset. You need to have the latest Windows 11 24H2 update installed on one of these Windows on Arm devices, and then the tool will unlock and enable Recall. [...] You can technically unlock Recall on x86 devices, but the app won't do much until Microsoft publishes the x64 AI components required to get it up and running. Rumors suggest both AMD and Intel are close to announcing Copilot Plus PCs, so Microsoft's AI components for those machines may well appear soon. I managed to get Recall running on an x64 Windows 11 virtual machine earlier today just to test out the initial first-run experience.

"While general-purpose languages like Python, C++, and Java remain popular in AI development," writes VentureBeat, "the resurgence of AI-first languages signifies a recognition that AI's unique demands require specialized languages tailored to the domain's specific needs... designed from the ground up to address the specific needs of AI development." Bend, created by Higher Order Company, aims to provide a flexible and intuitive programming model for AI, with features like automatic differentiation and seamless integration with popular AI frameworks. Mojo, developed by Modular AI, focuses on high performance, scalability, and ease of use for building and deploying AI applications. Swift for TensorFlow, an extension of the Swift programming language, combines the high-level syntax and ease of use of Swift with the power of TensorFlow's machine learning capabilities...

At the heart of Mojo's design is its focus on seamless integration with AI hardware, such as GPUs running CUDA and other accelerators. Mojo enables developers to harness the full potential of specialized AI hardware without getting bogged down in low-level details. One of Mojo's key advantages is its interoperability with the existing Python ecosystem. Unlike languages like Rust, Zig or Nim, which can have steep learning curves, Mojo allows developers to write code that seamlessly integrates with Python libraries and frameworks. Developers can continue to use their favorite Python tools and packages while benefiting from Mojo's performance enhancements... It supports static typing, which can help catch errors early in development and enable more efficient compilation... Mojo also incorporates an ownership system and borrow checker similar to Rust, ensuring memory safety and preventing common programming errors. Additionally, Mojo offers memory management with pointers, giving developers fine-grained control over memory allocation and deallocation...

Mojo is conceptually lower-level than some other emerging AI languages like Bend, which compiles modern high-level language features to native multithreading on Apple Silicon or NVIDIA GPUs. Mojo offers fine-grained control over parallelism, making it particularly well-suited for hand-coding modern neural network accelerations. By providing developers with direct control over the mapping of computations onto the hardware, Mojo enables the creation of highly optimized AI implementations.

According to Mojo's creator, Modular, the language has already garnered an impressive user base of over 175,000 developers and 50,000 organizations since it was made generally available last August. Despite its impressive performance and potential, Mojo's adoption might have stalled initially due to its proprietary status. However, Modular recently decided to open-source Mojo's core components under a customized version of the Apache 2 license. This move will likely accelerate Mojo's adoption and foster a more vibrant ecosystem of collaboration and innovation, similar to how open source has been a key factor in the success of languages like Python.

Developers can now explore Mojo's inner workings, contribute to its development, and learn from its implementation. This collaborative approach will likely lead to faster bug fixes, performance improvements and the addition of new features, ultimately making Mojo more versatile and powerful.
The article also notes other languages "trying to become the go-to choice for AI development" by providing high-performance execution on parallel hardware. Unlike low-level beasts like CUDA and Metal, Bend feels more like Python and Haskell, offering fast object allocations, higher-order functions with full closure support, unrestricted recursion and even continuations. It runs on massively parallel hardware like GPUs, delivering near-linear speedup based on core count with zero explicit parallel annotations — no thread spawning, no locks, mutexes or atomics. Powered by the HVM2 runtime, Bend exploits parallelism wherever it can, making it the Swiss Army knife for AI — a tool for every occasion...

The resurgence of AI-focused programming languages like Mojo, Bend, Swift for TensorFlow, JAX and others marks the beginning of a new era in AI development. As the demand for more efficient, expressive, and hardware-optimized tools grows, we expect to see a proliferation of languages and frameworks that cater specifically to the unique needs of AI. These languages will leverage modern programming paradigms, strong type systems, and deep integration with specialized hardware to enable developers to build more sophisticated AI applications with unprecedented performance. The rise of AI-focused languages will likely spur a new wave of innovation in the interplay between AI, language design and hardware development. As language designers work closely with AI researchers and hardware vendors to optimize performance and expressiveness, we will likely see the emergence of novel architectures and accelerators designed with these languages and AI workloads in mind. This close relationship between AI, language, and hardware will be crucial in unlocking the full potential of artificial intelligence, enabling breakthroughs in fields like autonomous systems, natural language processing, computer vision, and more.

The future of AI development and computing itself are being reshaped by the languages and tools we create today.
In 2017 Modular AI's founder Chris Lattner (creator of the Swift and LLVM) answered questions from Slashdot readers.

Western Digital has unveiled new 6TB external hard drives -- "the first new capacity point for this hard drive drive form factor in about seven years," reports Tom's Hardware. "There is a catch, though: the HDD is slow and will unlikely fit into any mobile PCs, so it looks like it will exclusively serve portable and specialized storage products." From the report: Western Digital's 6TB 2.5-inch HDD is currently used for the latest versions of the company's My Passport, Black P10, and G-Drive ArmorATD external storage devices and is not available separately. All of these drives (excluding the already very thick G-Drive ArmorATD) are thicker than their 5 TB predecessors, which may suggest that in a bid to increase the HDD's capacity, the manufacturer simply installed another platter and made the whole drive thicker instead of developing new platters with a higher areal density.

While this is a legitimate way to expand the capacity of a hard drive, it is necessary to note that 5TB 2.5-inch HDDs already feature a 15-mm z-height, which is the highest standard z-height for 2.5-inch form-factor storage devices. As a result, these 6TB 2.5-inch drives will unlikely fit into any desktop PC. When it comes to specifications of the latest My Passport, Black P10, and G-Drive ArmorATD external HDDs, Western Digital only discloses that they offer up to 130 MB/s read speed (just like their predecessors), feature a USB 3.2 Gen 1 (up to 5 GT/s) interface using either a modern USB Type-C or Micro USB Type-B connector and do not require an external power adapter.

At Google I/O today, Google introduced a "next-generation AI assistant" called Project Astra that can "make sense of what your phone's camera sees," reports Wired. It follows yesterday's launch of GPT-4o, a new AI model from OpenAI that can quickly respond to prompts via voice and talk about what it 'sees' through a smartphone camera or on a computer screen. It "also uses a more humanlike voice and emotionally expressive tone, simulating emotions like surprise and even flirtatiousness," notes Wired. From the report: In response to spoken commands, Astra was able to make sense of objects and scenes as viewed through the devices' cameras, and converse about them in natural language. It identified a computer speaker and answered questions about its components, recognized a London neighborhood from the view out of an office window, read and analyzed code from a computer screen, composed a limerick about some pencils, and recalled where a person had left a pair of glasses. [...] Google says Project Astra will be made available through a new interface called Gemini Live later this year. [Demis Hassabis, the executive leading the company's effort to reestablish leadership inÂAI] said that the company is still testing several prototype smart glasses and has yet to make a decision on whether to launch any of them.

Hassabis believes that imbuing AI models with a deeper understanding of the physical world will be key to further progress in AI, and to making systems like Project Astra more robust. Other frontiers of AI, including Google DeepMind's work on game-playing AI programs could help, he says. Hassabis and others hope such work could be revolutionary for robotics, an area that Google is also investing in. "A multimodal universal agent assistant is on the sort of track to artificial general intelligence," Hassabis said in reference to a hoped-for but largely undefined future point where machines can do anything and everything that a human mind can. "This is not AGI or anything, but it's the beginning of something."

Red Hat Enterprise Linux 9.4 has been released. But also released is Rocky Linux 9.4, reports 9to5Linux: Rocky Linux 9.4 also adds openSUSE's KIWI next-generation appliance builder as a new image build workflow and process for building images that are feature complete with the old images... Under the hood, Rocky Linux 9.4 includes the same updated components from the upstream Red Hat Enterprise Linux 9.4
This week also saw the release of Alma Linux 9.4 stable (the "forever-free enterprise Linux distribution... binary compatible with RHEL.") The Register points out that while Alma Linux is "still supporting some aging hardware that the official RHEL 9.4 drops, what's new is largely the same in them both."

And last week also saw the launch of the AlmaLinux High-Performance Computing and AI Special Interest Group (SIG). HPCWire reports: "AlmaLinux's status as a community-driven enterprise Linux holds incredible promise for the future of HPC and AI," said Hayden Barnes, SIG leader and Senior Open Source Community Manager for AI Software at HPE. "Its transparency and stability empowers researchers, developers and organizations to collaborate, customize and optimize their computing environments, fostering a culture of innovation and accelerating breakthroughs in scientific research and cutting-edge AI/ML."
And this week, InfoWorld reported: Red Hat has launched Red Hat Enterprise Linux AI (RHEL AI), described as a foundation model platform that allows users to more seamlessly develop and deploy generative AI models. Announced May 7 and available now as a developer preview, RHEL AI includes the Granite family of open-source large language models (LLMs) from IBM, InstructLab model alignment tools based on the LAB (Large-Scale Alignment for Chatbots) methodology, and a community-driven approach to model development through the InstructLab project, Red Hat said.

This year Dell moved to soldered RAM for its XPS 14 and 16, writes Digital Trends, which "makes it impossible to upgrade, or even repair."

"This was a big change from the past, where the XPS 15 and 17 were both celebrated for their upgradability." Of course, Dell isn't the first to make the transition. In fact, they're one of the last, which is what makes the decision so much tougher to swallow. Where soldered RAM was previously limited to just MacBooks and ultrabooks, it's now affecting most high-performance laptops for gaming as well. Even the fantastic ROG Zephyrus G14 moved to soldered memory this year.
After two months of research, the article's author acknowledges "there are tangible benefits to companies using soldered RAM, and all the people I spoke to while writing this agree that they outweigh the downsides, but how that applies to the end-user is a bit more complicated." If there's one thing and one thing only that soldered RAM is indisputably good for, it's saving space. [Haval Othman, a senior director of experience engineering at HP] explained the benefits, saying: "If battery life, mobility, form factor (thin and light), and power efficiency are my priority among other design choices, then my mind immediately goes to soldered RAM; because that's where soldered RAM can be beneficial and power-efficient, which will lead to longer battery life. Plus, it's going to give me more space on the motherboard, so I can design the product thinner and lighter. [...] If we want a thin product, the trade-off is soldering more of the devices onto the board."

This tracks. In a laptop, there's only so much space that can be used for components, and that free space grows smaller by the year to make ultrabooks possible. They're an industrywide trend that was first popularized by Apple, and the rest of the laptop manufacturing world quickly caught on. Each year, laptops are released thinner and lighter, and that means having to squeeze the components together in new, innovative ways... Soldering the memory down onto the motherboard means that it can be attached almost anywhere within the laptop instead of being slotted into a specific part of it. It effectively makes the laptop thinner by cutting back on the space that the RAM module takes up. The space saved by soldering memory can be used for other things, such as a bigger battery....

All three companies that I spoke to stress the form factor much more than any tangible cost benefits... Stuart Gill, director of global media relations, campaigns, and corporate content [said] "Both soldered and socketed RAM designs are now quite mature. As a result, we see no impact on the manufacturing process and, therefore, the cost to the consumer."
SO-DIMM chips also have "relatively limited bandwidth," according to HP's Othman, "while when you solder the memory chips onto the board, you can build it for a much wider bandwidth."

But the article ends by looking to the future. "The good news is that SO-DIMM memory might eventually be replaced by the CAMM2 standard." Recently approved by JEDEC, CAMM2 is said to be significantly thinner, and it'll be available both in soldered and non-soldered variants. Using CAMM2 will allow laptops to stack up to 128GB of RAM, and the frequencies are said to be going up, too. CAMM2 can also activate dual-channel memory with just a single module.