Kylie Robison reports via The Verge: Hugging Face, one of the biggest names in machine learning, is committing $10 million in free shared GPUs to help developers create new AI technologies. The goal is to help small developers, academics, and startups counter the centralization of AI advancements. [...] Delangue is concerned about AI startups' ability to compete with the tech giants. Most significant advancements in artificial intelligence -- like GPT-4, the algorithms behind Google Search, and Tesla's Full Self-Driving system -- remain hidden within the confines of major tech companies. Not only are these corporations financially incentivized to keep their models proprietary, but with billions of dollars at their disposal for computational resources, they can compound those gains and race ahead of competitors, making it impossible for startups to keep up. Hugging Face aims to make state-of-the-art AI technologies accessible to everyone, not just the tech giants. [...]

Access to compute poses a significant challenge to constructing large language models, often favoring companies like OpenAI and Anthropic, which secure deals with cloud providers for substantial computing resources. Hugging Face aims to level the playing field by donating these shared GPUs to the community through a new program called ZeroGPU. The shared GPUs are accessible to multiple users or applications concurrently, eliminating the need for each user or application to have a dedicated GPU. ZeroGPU will be available via Hugging Face's Spaces, a hosting platform for publishing apps, which has over 300,000 AI demos created so far on CPU or paid GPU, according to the company.

Access to the shared GPUs is determined by usage, so if a portion of the GPU capacity is not actively utilized, that capacity becomes available for use by someone else. This makes them cost-effective, energy-efficient, and ideal for community-wide utilization. ZeroGPU uses Nvidia A100 GPU devices to power this operation -- which offer about half the computation speed of the popular and more expensive H100s. "It's very difficult to get enough GPUs from the main cloud providers, and the way to get them -- which is creating a high barrier to entry -- is to commit on very big numbers for long periods of times," Delangue said. Typically, a company would commit to a cloud provider like Amazon Web Services for one or more years to secure GPU resources. This arrangement disadvantages small companies, indie developers, and academics who build on a small scale and can't predict if their projects will gain traction. Regardless of usage, they still have to pay for the GPUs. "It's also a prediction nightmare to know how many GPUs and what kind of budget you need," Delangue said.

The Japanese space agency (JAXA) and NASA plan to launch a satellite made of wood in 2024. The Independent reports: The high durability of wood in space was recently tested and confirmed at the International Space Station (ISS) by an international group of scientists led by those from Kyoto University. Their experiments showed wood samples tested at the ISS for durability underwent minimal deterioration and maintained good stability. Preliminary inspection, including strength tests and crystal structural analyses, of the wood samples was also done once they were brought back to Earth from the ISS by Japanese astronaut Koichi Wakata.

Despite extreme conditions in space, including temperature changes and exposure to intense cosmic rays and dangerous solar particles for 10 months, tests found no changes in the samples, such as cracking, warping, peeling or surface damage, according to a recent Kyoto University statement. The retrieved wood specimens were tested and showed no deformation after space exposure and also did not undergo any mass change before and after space exposure, scientists said.

The international research group has determined that the satellite LignoSat, slated to be jointly launched in 2024 by Nasa and Japan's space agency Jaxa, will likely use Magnolia wood -- "Hoonoki" in Japanese. Magnolia, researchers said, has relatively high workability, dimensional stability and overall strength, making its properties ideal for the mission. Wood also has some benefits compared to complex alloys used in space vehicles, as it is environmentally friendly, easier to produce and can be disposed off better at the end of a satellite's life. Such wooden satellites may also be designed to completely burn up on re-entry into the atmosphere and even if small fragments did survive, they may decompose easily.

"NASA's DAVINCI mission to Venus is scheduled for launch in 2029," reports Gizmodo, adding that a new paper "details this upcoming journey, a daring mission that could shed new light on the scorching hot planet's mysterious, and potentially habitable, past." Upon its arrival at the second planet from the Sun, the probe will plunge through Venus' atmosphere, ingesting its gases for approximately one hour before landing on the planet's surface, according to the paper published in The Planetary Science Journal. DAVINCI is designed to act as a flying chemistry lab, and it will use its built-in instruments to analyze Venus's atmosphere, temperatures, pressure and wind speed, while taking a few photos of its trip through planetary hell...

If it survives the atmospheric entry, the probe will — hopefully — land in the Alpha Regio mountains, which are roughly the size of Texas, according to the researchers behind the new paper. Under ideal conditions, the probe will operate for 17 to 18 minutes once it sticks the landing, but it isn't really required to operate on Venus since all the precious data will have already been collected during its atmospheric plunge.
Digital Trends calls Venus "a frontier in planetary science about which very little is known" — then explains why that is. The biggest issue for any potential mission to Venus is the heat, as the surface temperatures can be as high as 900 degrees Fahrenheit (475 degrees Celsius). That's hot enough to melt lead, and it wreaks havoc with electronics... The pressure at the surface is around 95 bars, or nearly 100 times the atmospheric pressure on Earth's surface, so engineering a probe for this kind of environment is kind of like building a submarine... To keep the probe active for as long as possible, it is spherical and covered in a thick titanium shell to withstand the pressure and insulate against the heat. Then there's more insulation inside this shell, made of special materials including astroquartz, a type of fiber made from fused quartz... It's then filled with carbon dioxide gas to protect the high-voltage electronics from sparking and to stop any Earth gases from leaking in during launch....

The descent sphere will also have a camera that will be snapping high-contrast images of the surface, which can then be built up into 3D maps. For a camera to operate from inside a metal sphere, though, you need a window. And glass isn't a great material for dealing with intensely high-pressure environments. That's why DAVINCI's window will be made not of glass but sapphire... "Our final images will have 10-centimeter resolution," said the team's principal investigator, Jim Garvin. "That's the scale you'd see looking out across your living room...."

Researchers know that the clouds of Venus have drops of sulfuric acid in them — and sulfuric acid eats through materials. It's a particular concern for the Kevlar lanyard that will attach the descent sphere to the parachute. So to test whether the lanyard can withstand the acidic environment, the engineers don't just suspend it in a few drops of acid — they coat the entire surface in acid, then test the lanyard's pull strength to make sure it can survive long enough to take the probe through the atmosphere even in the worst possible case.
SciTechDaily notes DAVINCI "is the first mission to study Venus using both spacecraft flybys and a descent probe....

"It will also provide the first descent imaging of the mountainous highlands of Venus while mapping their rock composition and surface relief at scales not possible from orbit."

"AMD and Intel have both proposed better ways of doing interrupt and exception handling the last few months," reports LinuxReviews.org.

Then they share this analysis Linus Torvalds posted on the Real World Technologies forum: "The AMD version is essentially "Fix known bugs in the exception handling definition".

The Intel version is basically "Yeah, the protected mode 80286 exception handling was bad, then 386 made it odder with the 32-bit extensions, and then syscall/sysenter made everything worse, and then the x86-64 extensions introduced even more problems. So let's add a mode bit where all the crap goes away".

In contrast, the AMD one is basically a minimal effort to fix actual fundamental problems with all that legacy-induced crap that are nasty to work around and that have caused issues...

Both are valid on their own, and they are actually fairly independent. Honestly, the AMD paper looks like a quick "we haven't even finished thinking all the details through, but we know these parts were broken, so we might as well release this".

I don't know how long it has been brewing, but judging by the "TBD" things in that paper, I think it's a "early rough draft"."
In the article (shared by long-time Slashdot reader xiando), LinuxReviews.org summarizes the state of the conversation today: Torvalds went on to say that while AMD's proposed "quick fix" would be easier to implement for him and others operating system vendors, it's not ideal in the long run. Intel's proposal throws the entire existing interrupt descriptor table (IDT) delivery system under the bus so it can be replaced with what they call a new "FRED event delivery" system. Torvalds believes this is a better long-term solution...

While the pros and cons of Intel and AMD's respective proposals for interrupt and event handling in future processors are worthy of discussion, it's in reality mostly up to Intel. They are the bigger and more powerful corporation. It is more likely than not that future processors from Intel will use their proposed Flexible Return and Event Delivery system. Their next generation processors won't, it will take years not months before consumer CPUs have the FRED technology. Remember, the above-mentioned technical document was published earlier this month [in March]. Things do not magically go from the drawing-board to store-shelves overnight.

Intel isn't going to just hand the FRED technology over to AMD and help them implement it. We will likely see both move forward with their own proposals. Intel will have FRED and AMD will have Supervisor Entry Extensions until AMD, inevitably, adopts FRED or some form of it years down the line.
They also note that Torvalds took issue with a poster arguing that microkernels are more secure than monolithic kernels like Linux. "Bah, you're just parroting the usual party line that had absolutely no basis in reality and when you look into the details, doesn't actually hold up.

It's all theory and handwaving and just repeating the same old FUD that was never actually really relevant."

An anonymous reader quotes a report from Mac Rumors: Earlier this week, Japanese publication Nikkan Kogyo Shimbun reported that Sony will supply Apple with OLED microdisplays for its widely rumored AR/VR glasses, as spotted by Mac Otakara. The report has since been corroborated by display industry analyst Ross Young, who said multiple sources have informed him that Apple is indeed planning to use Sony's microdisplay technology for its head-mounted accessory. According to FRAMOS, a supplier of embedded vision technologies, Sony's OLED microdisplays are small, cutting-edge displays with an ultra-fast response rate, ultra-high contrast, a wide color gamut for precise color reproduction, high luminance, low reflectance, and other benefits that would be ideal for Apple's glasses. Sony's microdisplays also have integrated drivers for a thin and light design, and power-saving modes are available for longer battery life.

Young said the glasses will use a 0.5-inch display with a 1,280x960 resolution, and these specs appear to correspond with Sony's ECX337A component. According to Sony's website, this microdisplay in particular has a max brightness of 1,000 nits, an ultra-high contrast of 100,000:1, and an ultra-fast response rate of 0.01 ms or less. The high contrast provided by Sony's microdisplays allows an additional information layer to appear seamlessly, and not as an overlay. "This information is simply added to the background for a 'real AR' experience," according to FRAMOS. According to the Nikkan Kogyo Shimbun, Apple plans to release its AR/VR glasses in 2021, but analyst Ming-Chi Kuo does not expect a release until 2022 at the earliest. Young also believes that the glasses will be introduced in the first half of 2022.

An anonymous reader shares a fascinating true-crime story from Pittsburgh. Smithsonian magazine reports: Like nuclear power plants and sensitive computer networks, the safest rare book collections are protected by what is known as "defense in depth" — a series of small, overlapping measures designed to thwart a thief who might be able to overcome a single deterrent. The Oliver Room, home to the Carnegie Library of Pittsburgh's rare books and archives, was something close to the platonic ideal of this concept. Greg Priore, manager of the room starting in 1992, designed it that way.

The room has a single point of entry, and only a few people had keys to it. When anyone, employee or patron, entered the collection, Priore wanted to know. The room had limited daytime hours, and all guests were required to sign in and leave personal items, like jackets and bags, in a locker outside. Activity in the room was under constant camera surveillance. In addition, the Oliver Room had Priore himself. His desk sat at a spot that commanded the room and the table where patrons worked. When a patron returned a book, he checked that it was still intact. Security for special collections simply does not get much better than that of the Oliver Room.

In the spring of 2017, then, the library's administration was surprised to find out that many of the room's holdings were gone. It wasn't just that a few items were missing. It was the most extensive theft from an American library in at least a century, the value of the stolen objects estimated to be $8 million...
Perpetrating a daring 25-year heist, the thief "stole nearly everything of significant monetary value," the magazine reports. So who done it? Just about the only thing that keeps an insider from stealing from special collections is conscience. Security measures may thwart outside thieves, but if someone wants to steal from the collection he stewards, there is little to stop him. Getting books and maps and lithographs out the door is not much harder than simply taking them from the shelves...
The perpetrator was ultimately sentenced to three years' house arrest and 12 years' probation, the article reports, while his fence received four years' house arrest and 12 years' probation.

"After the sentences were made public, Carole Kamin, a member of the board of the Carnegie Natural History Museum, wrote to the Pittsburgh Post-Gazette that supporters of local nonprofits 'were appalled at the unbelievably light sentences.'"

Walmart uses "Everseen" AI technology in thousands of its stores "to prevent shoplifting at registers and self-checkout kiosks," reports Wired.

But some Walmart workers claim that instead it's often failed to stop actual instances of stealing, misidentified innocuous behavior as theft -- and made it harder for them to social distance: The workers said they had been upset about Walmart's use of Everseen for years and claimed colleagues had raised concerns about the technology to managers but were rebuked. They decided to speak to the press, they said, after a June 2019 Business Insider article reported Walmart's partnership with Everseen publicly for the first time. The story described how Everseen uses AI to analyze footage from surveillance cameras installed in the ceiling and can detect issues in real time, such as when a customer places an item in their bag without scanning it. When the system spots something, it automatically alerts store associates...

In interviews, the workers, whose jobs include knowledge of Walmart's loss-prevention programs, said their top concern with Everseen was false positives at self-checkout. The employees believe that the tech frequently misinterprets innocent behavior as potential shoplifting, which frustrates customers and store associates, and leads to longer lines. "It's like a noisy tech, a fake AI that just pretends to safeguard," said one worker.

The coronavirus pandemic has given their concerns more urgency. One Concerned Home Office Associate said they worry false positives could be causing Walmart workers to break social-distancing guidelines unnecessarily. When Everseen flags an issue, a store associate needs to intervene and determine whether shoplifting or another problem is taking place. In an internal communication from April obtained by WIRED, a corporate Walmart manager expressed strong concern that workers were being put at risk by the additional contact necessitated by false positives and asked whether the Everseen system should be turned off to protect customers and workers.

Before COVID-19, "it wasn't ideal, it was a poor customer experience," the worker said. "AI is now creating a public health risk."

szczys writes: Laser cutters are awesome, but you have to bring your mechanical engineering A-game if you want to build resilient stuff using laser-cut parts. Joshua Vasquez has been building up his bag of tricks using Delrin and a laser cutter to build with techniques like press-fitting, threading, snap-fits, etc. that aren't possible or are non-ideal with the laser-cutting steadfasts of plywood and acrylic. Delrin (PDF) won't shatter like acrylic, and it has more give to it, so even the less precise entry-level lasers can cut joints that will have a snug fit.

MojoKid writes "The power efficiency of NVIDA's Maxwell architecture make it ideal for mobile applications, so today's announcement by NVIDIA of a new top-to-bottom line-up of mobile GPUs—most of them featuring the Maxwell architecture—should come as no surprise. Though a couple of Kepler and even Fermi-based GPUs still exist in NVIDIA's new line-up, the heart of the product stack leverages Maxwell. The entry-level parts in the GeForce 800M series consist of the GeForce GT 820M, 830M, and 840M. The 820M is a Fermi-based GPU, but the 830M and 840M are new chips that leverage Maxwell. The meat of the GeForce GTX 800M series consist of Kepler-based GPUs, though Maxwell is employed in the more mainstream parts. NVIDIA is claiming the GeForce GTX 880M will be fastest mobile GPU available, but the entire GTX line-up will offer significantly higher performance then any integrated graphics solution. The GeForce GTX 860M and 850M are essentially identical to the desktop GeForce GTX 750 Ti, save for different frequencies and memory configurations. There are a number of notebooks featuring NVIDIA's GeForce 800M series GPUs coming down the pipeline from companies like Alienware, Asus, Gigabyte, Lenovo, MSI and Razer, though others are sure the follow suit. Some of the machines will be available immediately."

Hugh Pickens writes writes "Henry Alford writes that in an ideal world, we would all use Google to be better friends by having better recall and to research our new friends and acquaintances to get to know them better. 'It's perfectly natural and almost always appropriate,' says social anthropologist Kate Fox. 'Obviously, one is always going to have to be discreet when talking about what you've found. But our brains haven't changed since the Stone Age, and humans are designed to live in small groups in which everyone knows one another. Googling is an attempt to recreate a primeval, preindustrial pattern of interaction.' But the devil is in the details. If we tell a new friend that we've read her LinkedIn entry or her wedding announcement, it probably won't be perceived as trespassing, as long we bear no ulterior motives. If we happen to reveal that we've also read her long-ago abandoned blog about her cat, we're more likely to be seen as chronically bored than menacing. 'I'm so baffled by this idea that we're not supposed to Google people,' says Dean Olsher. 'Why would there be a line? Like everyone else is allowed to know something but I'm not?' But doesn't taking the google shortcut to a primeval, preindustrial pattern of recognition sometimes rob encounters of their inherent mystery or even get us in trouble? Tina Jordan, an executive in book publishing who has the same name as a former girlfriend of Hugh Hefner, says, 'I typically tell any blind dates before I meet them that they probably shouldn't Google my name, otherwise they'll be sorely disappointed when they meet me.'"

Monday you had a chance to ask Linus Torvalds any question you wanted. We sent him a dozen of the highest rated and below you'll see what he has to say about computers, programming, books, and copyrights. He also talks about what he would have done differently with Linux if he had to do it all over again. Hint: it rhymes with nothing.

Bennett Haselton is back with a thought provoking essay about not just an incident of Internet censorship on an American university campus, but a proposed method of propagating news, so that relevant stories aren't buried as easily by chance or time. Bennett writes: "The real scandal in the story of Arizona State University blocking students' access to the Change.org website, is not just that it happened, but that the block persisted for two months without being mentioned in the media. As a card-carrying member of the 'outrage grapevine,' I surely think we need a way to respond faster." Read on for the rest.

An anonymous reader writes "So, it's that wonderful time of year again. Instead of buying the latest, greatest whiz-bang, overpriced fad toy of the year, I thought I might try my hand at corrupting my nieces (ages 12 and 14) in a nerdier direction with some sort of introductory robotics kit. They have no programming experience, and part of my idea is that it would be encouraging for them to see interactions of their code that they write with the real-world by being able to control some actuators and read sensors. The first thing that comes to mind is Lego Mindstorms, but I find them a bit on the pricey side of things. My budget falls between $40 and $100, and the ideal kit would focus more on the software side than on soldering together circuits. I'd be looking for a kit that provides an easy to learn API and development tools that will work with a standard Windows PC. I don't mind spending a few afternoons helping them out with the basics, but I'd like for them to be able to be able to explore on their own after grasping the initial concepts. Has anybody gotten their younger relations into programming through robotics, and what kits might you recommend?"

Reader Justblair recommends his blog entry detailing how he made a hard drive silencer for a pittance. "This article demonstrates a very easy-to-make hard drive silencer that not only outperforms most commercially available devices, but is cheaper to implement as well. Requiring very little in fabrication skills, it is an ideal addition to a media PC or HTPC. It may even suit you if your head is aching after many hours of being whined at by your hard drive."

It's that time of year again. Last year's response to our Game of the Year post was so enthusiastic that I thought it would be worthwhile to give it another go. So, once again, some of the Slashdot folks have come together to offer up our 'games of the year'. Scuttlemonkey, Scott Collins, Chris Brown, CmdrTaco, and myself have all put together quick blurbs about the games we couldn't get enough of this year. When you're through reading those, it's your turn to speak up. What was the game you couldn't put down? The next-gen consoles really came into their own this year; was it one of those games, or something for the PC? In your opinion, what was the best game of the year?

You submitted questions for Håkon Wium Lie on June 20. Today we have his answers, not only to the (+5 moderated) questions we sent him, but to a bunch of others he thought would also be interesting to answer.

A few months back we went and redesigned Slashdot with fancy new CSS templates. The idea was that with a new clean CSS framework under the skin, we could more easily redesign the look & feel of the site. At that time I mentioned that we wanted to have a contest to redesign Slashdot. Well that time has come. Read on for the rules, instructions, and timeline. Oh, and did I mention that the top prize is a new laptop?

Charles Fishman, senior writer for Fast Company magazine has recently published a book entitled The Man Who Said No To Wal-mart. It's an excellent book (Yes, I've read it) that talks about the intersection of making good stuff, the commodization of products, and the changing world that we work in; not exactly high tech, but tech nonetheless.

Joe Kauzlarich writes "There's an ever-growing number of fun niche books seeping onto the mathematics bookshelves, that, while not essential, are almost always guaranteed to leave the reader with a fuller taste of the subject at hand and an appetite to learn more. Mario Livio's The Equation That Couldn't Be Solved is a modest semi-classic of pop-math literature, focusing on the central concepts of group theory, the subject that turned mathematics on its head a century and a half ago and has ever since been one of the delights of studying higher mathematics." Read on for the rest of Joe's review.

Bruce Lawson submits the review below of Stuart Langridge's Excellent guide to creating dynamic web pages; scalable and sensible., writing "Don't be put off by the title: the DHTML here bears no resemblance to the stupid Web tricks of the late 90s that allowed animated unicorns to follow your cursor or silly Powerpoint-like transitions between Web pages." Read on for the rest.