An anonymous reader shares a report: "Early Access" was once a novel, quirky thing, giving a select set of Steam PC games a way to involve enthusiastic fans in pre-alpha-level play-testing and feedback. Now loads of games launch in various forms of Early Access, in a wide variety of readiness. It's been a boon for games like Baldur's Gate 3, which came a long way across years of Early Access. Early Access, and the "Advanced Access" provided for complete games by major publishers for "Deluxe Editions" and the like, has also been a boon to freeloaders.

Craven types could play a game for hours and hours, then demand a refund within the standard two hours of play, 14 days after the purchase window of the game's "official" release. Steam-maker Valve has noticed and, as of Tuesday night, updated its refund policy. "Playtime acquired during the Advanced Access period will now count towards the Steam refund period," reads the update. In other words: Playtime is playtime now, so if you've played more than two hours of a game in any state, you don't get a refund. That closes at least one way that people could, with time-crunched effort, play and enjoy games for free in either Early or Advanced access.

Just as Google, Samsung and Microsoft continue to push their efforts with generative AI on PCs and mobile devices, Apple is moving to join the party with OpenELM, a new family of open source large language models (LLMs) that can run entirely on a single device rather than having to connect to cloud servers. From a report: Released a few hours ago on AI code community Hugging Face, OpenELM consists of small models designed to perform efficiently at text generation tasks. There are eight OpenELM models in total -- four pre-trained and four instruction-tuned -- covering different parameter sizes between 270 million and 3 billion parameters (referring to the connections between artificial neurons in an LLM, and more parameters typically denote greater performance and more capabilities, though not always).

[...] Apple is offering the weights of its OpenELM models under what it deems a "sample code license," along with different checkpoints from training, stats on how the models perform as well as instructions for pre-training, evaluation, instruction tuning and parameter-efficient fine tuning. The sample code license does not prohibit commercial usage or modification, only mandating that "if you redistribute the Apple Software in its entirety and without modifications, you must retain this notice and the following text and disclaimers in all such redistributions of the Apple Software." The company further notes that the models "are made available without any safety guarantees. Consequently, there exists the possibility of these models producing outputs that are inaccurate, harmful, biased, or objectionable in response to user prompts."

Today, Framework is the modular repairable laptop company. Tomorrow, it wants to be a consumer electronics company, period. From a report: That's one of the biggest reasons it just raised another $18 million in funding -- it wants to expand beyond the laptop into "additional product categories." Framework CEO Nirav Patel tells me that has always been the plan. The company originally had other viable ideas beyond laptops, too. "We chose to take on the notebook space first," he says, partly because Framework knew it could bootstrap its ambitions by catering to the PC builders and tinkerers and Linux enthusiasts left behind by big OEMs -- and partly because it wanted to go big or go home.

If Framework could succeed in laptops, he thought, it would be able to build almost anything. After five years building laptops, what might Framework add to the portfolio? Patel won't say -- I only get the barest hints, no matter how many different ways I ask. He won't even say if they'll make less or more of a splash than laptops. Framework might choose an "equally difficult" category or might instead try something "a bit smaller and simpler to execute, streamlined now that we have all this infrastructure."

Edward Zitron, citing emails released as part of the Department of Justice's antitrust case against Google, writes about Prabhakar Raghavan: And Raghavan -- a manager, hired by Sundar Pichai, a former McKinsey man and a manager by trade -- is an example of everything wrong with the tech industry. Despite his history as a true computer scientist with actual academic credentials, Raghavan chose to bulldoze actual workers and replace them with toadies that would make Google more profitable and less useful to the world at large. Since Prabhakar took the reins in 2020, Google Search has dramatically declined, with the numerous "core" search updates allegedly made to improve the quality of results having an adverse effect, increasing the prevalence of spammy, search engine optimized content.

It's because the people running the tech industry are no longer those that built it. Larry Page and Sergey Brin left Google in December 2019 (the same year as the Code Yellow fiasco), and while they remain as controlling shareholders, they clearly don't give a shit about what "Google" means anymore. Prabhakar Raghavan is a manager, and his career, from what I can tell, is mostly made up of "did some stuff at IBM, failed to make Yahoo anything of note, and fucked up Google so badly that every news outlet has run a story about how bad it is." This is the result of taking technology out of the hands of real builders and handing it to managers at a time when "management" is synonymous with "staying as far away from actual work as possible." And when you're a do-nothing looking to profit as much as possible, you only care about growth. You're not a user, you're a parasite, and it's these parasites that have dominated and are draining the tech industry of its value.

Raghavan's story is unique, insofar as the damage he's managed to inflict (or, if we're being exceptionally charitable, failed to avoid in the case of Yahoo) on two industry-defining companies, and the fact that he did it without being a CEO or founder. Perhaps more remarkable, he's achieved this while maintaining a certain degree of anonymity. Everyone knows who Musk and Zuckerberg are, but Raghavan's known only in his corner of the Internet. Or at least he was. Now Raghavan has told those working on search that their "new operating reality" is one with less resources and less time to deliver things. Rot Master Raghavan is here to squeeze as much as he can from the corpse of a product he beat to death with his bare hands. Raghavan is a hall-of-fame rot economist, and one of the many managerial types that have caused immeasurable damage to the Internet in the name of growth and "shareholder value." And I believe these uber-managers - these ultra-pencil-pushers and growth-hounds - are the forces destroying tech's ability to innovate.

An anonymous reader shares a report: It used to be that you could pay for a retail version of Windows 11 and expect it to be ad-free, but those days are apparently finito. The latest update to Windows 11 (KB5036980) comes out this week and includes ads for apps in the "recommended" section of the Start Menu, one of the most oft-used parts of the OS. "The Recommended section of the Start menu will show some Microsoft Store apps," according to the release notes. "These apps come from a small set of curated developers." The app suggestions are enabled by default, but you can restore your previously pristine Windows experience if you've installed the update, fortunately. To do so, go into Settings and select Personalization > Start and switch the "Show recommendations for tips, app promotions and more" toggle to "off."

An anonymous reader shares a report: Diamonds may be forever but they are also seriously on sale. Natural rough diamond prices have collapsed 26 per cent in the past couple of years. Tepid US and Chinese demand for diamond jewellery hasn't helped. But most ring fingers point at the increasing popularity of cheaper laboratory grown diamonds (LGD). This fracturing of the diamond market is set to last. After a brief pandemic-era boom in diamond jewellery, miners are battling to whittle down oversupply of gems. Anglo-American's De Beers, along with Russia's Alrosa, control two-thirds of the rough diamond supply. DeBeers this week said its rough sales dropped 23 per cent in the first quarter.

It is not enough. While rough stone inventory has stabilised of late, polished diamond stocks remain high. At more than $20bn at the end of 2023, these were near five-year highs, up a third since the end of 2022, according to Bank of America. Worse, as LGDs have taken market share, their prices have declined too, to about 15 per cent or less of their natural counterparts. Diamond miners spent years maintaining that romantic buyers would prefer the allure of rare, natural stones. It increasingly appears they were wrong.

Synthetic diamonds are nothing new, having appeared about 70 years ago mostly for industrial purposes. But in the past decade LGDs have taken off. In 2015, LGD supply barely featured as a rival to natural stones. By last year it was more than 10 per cent of the global diamond jewellery market, according to specialist Paul Zimnisky. This has created a competitive frenzy among producers. LGDs' lower costs have enabled them to slash prices. In October, WD Lab Grown Diamonds, America's second-largest maker of synthetics, filed for bankruptcy. It has since had to shift its business away from retail towards industrial customers.

President Joe Biden signed a foreign aid package that includes a bill that would ban TikTok if China-based parent company ByteDance fails to divest the app within a year. The Verge: The divest-or-ban bill is now law, starting the clock for ByteDance to make its move. The company has an initial nine months to sort out a deal, though the president could extend that another three months if he sees progress. While just recently the legislation seemed like it would stall out in the Senate after being passed as a standalone bill in the House, political maneuvering helped usher it through to Biden's desk. The House packaged the TikTok bill -- which upped the timeline for divestment from the six months allowed in the earlier version -- with foreign aid to US allies, which effectively forced the Senate to consider the measures together. The longer divestment period also seemed to get some lawmakers who were on the fence on board.

An anonymous reader shares a report: Qualcomm is cheating on the Snapdragon X Plus/Elite benchmarks given to OEMs and the press. SemiAccurate doesn't use these words lightly but there is no denying what multiple sources are telling us. [...] Then there were the actual 'briefings' for the X Pro SoC. To call them pathetic is giving them more than their due. The deck was 11 slides, three of which were empty/fluff, five 'benchmark' slides with woefully inadequate disclosure, and two infographic summary slides. The last was the slide below with the 'deep technical' stats [screenshots in the linked article], much of which we told you about last week. And more.

The rest of the 'disclosure' for Snapdragon X Pro was a list of features that all fall under the guise of exactly what you would expect. The rest was filled with deep 'details' like the GPU capabilities of 3.8TFLOPS. That's it. No specs, no capabilities, no nothing. It was truly pathetic. But wait there is more, or less really, with statements like it having AV1 encode and decode. Trivialities like frame rates and resolutions were seemingly not needed for such technical briefs. See what we mean by pathetic? Those 10 cores are arranged how again? That 42MB of cache is what level? Shall I go on about the bare minimum basics or do you get the point now? SemiAccurate was planning to ask Qualcomm about their cheating on benchmarks at the promised briefing but, well, they lied to us and cut us out of the pathetic bits they did brief on. We honestly would have liked to know why they were cheating but we kind of think they will do their usual response to bad news and pretend it never happened like last time. If they actually do explain things we will of course update this article as we always do.

Nvidia, in a blog post: To help customers make more efficient use of their AI computing resources, NVIDIA today announced it has entered into a definitive agreement to acquire Run:ai, a Kubernetes-based workload management and orchestration software provider. Customer AI deployments are becoming increasingly complex, with workloads distributed across cloud, edge and on-premises data center infrastructure.

Managing and orchestrating generative AI, recommender systems, search engines and other workloads requires sophisticated scheduling to optimize performance at the system level and on the underlying infrastructure. Run:ai enables enterprise customers to manage and optimize their compute infrastructure, whether on premises, in the cloud or in hybrid environments. The deal is valued at about $700 million.

Longtime Slashdot reader sfraggle writes: Biplane shoot-'em up, Sopwith, is celebrating 40 years today since its first release back in 1984. The game is one of the oldest PC games still in active development today, originating as an MS-DOS game for the original IBM PC. The 40th anniversary site has a detailed history of how the game was written as a tech demo for the now-defunct Imaginet networking system. There is also a video interview with its original authors. "The game involves piloting a Sopwith biplane, attempting to bomb enemy buildings while avoiding fire from enemy planes and various other obstacles," reads the Wiki page. "Sopwith uses four-color CGA graphics and music and sound effects use the PC speaker. A sequel with the same name, but often referred to as Sopwith 2, was released in 1985."

You can play Sopwith in your browser here.

Okian Warrior writes: For $10,000, you can now get a flamethrower mounted on a robotic dog. Just load the webpage and scroll down. I saw this on the news today. *Definitely* we need to have a conversation about where AI is going. The robot, called the Thermonator, is constructed by Ohio flame throwing manufacturer Throwflame and features one of the company's ARC flamethrowers mounted on its back. The 26-pound robotic quadruped "can shoot fire in a 30-foot stream and comes with a built-in fuel tank powered by gasoline," notes Gizmodo. "The company says the robot also has an hour-long battery, a laser sight, and lidar mapping, and it can be remotely controlled via the company's app."

The company says its product is designed for "wildfire control and prevention," "agriculture management," "ecological conservation," "entertainment and SFX," and "snow and ice removal." It can be yours for the low price of $9,420 with free shipping.

Construction has begun on a $12 billion high-speed rail project to connect Las Vegas and Los Angeles by the end of the decade. The project, backed by $3 billion in federal support, aims to reduce travel time to under two hours and significantly cut greenhouse gas emissions. Popular Science reports: Brightline expects its trains will depart every 40 minutes from a station outside of the Vegas strip and another one in the LA suburb of Rancho Cucamonga. When it's completed, the train will travel at 186 miles per hour, making it the fastest train in the U.S. and comparable to Japan's famous bullet trains. For context, Brightline's most recently completed train connecting parts of Florida is estimated to top out around 130 miles per hour. Both of those still fall far short of the speed achieved by the world fastest commuter train in Shanghai, which can reportedly reach a speed of 286 miles per hour. Still, the new train could complete the 218 mile trip between Sin City and a suburb of the City of Angels in just 2 hours and 10 minutes. That same trip would take about four hours by car, and that's without substantial traffic.

Once built, the trains will reportedly include onboard Wi-Fi, restrooms, and food and drinks available for purchase. Brightline hasn't provided an exact price for how much an individual train ticket will cost but has instead said they expect it to be roughly equivalent to the price of an airline flight. Brightline reportedly believes the train could attract 11 million one-way passengers annually once it's up and running. The U.S. Department of Transportation estimates the new train could cut back 400,000 tons of carbon dioxide per year and create 35,000 new jobs. Department of Transportation Secretary Pete Buttigieg described the moment as a "major milestone in building the future of American rail." The ceremony symbolically took place on Earth Day. "Partnering with state leaders and Brightline West, we're writing a new chapter in our country's transportation story that includes thousands of union jobs, new connections to better economic opportunity, less congestion on the roads, and less pollution in the air," Buttigieg said in a statement.

The Federal Trade Commission narrowly voted Tuesday to ban nearly all noncompetes, employment agreements that typically prevent workers from joining competing businesses or launching ones of their own. From a report: The FTC received more than 26,000 public comments in the months leading up to the vote. Chair Lina Khan referenced on Tuesday some of the stories she had heard from workers. "We heard from employees who, because of noncompetes, were stuck in abusive workplaces," she said. "One person noted when an employer merged with an organization whose religious principles conflicted with their own, a noncompete kept the worker locked in place and unable to freely switch to a job that didn't conflict with their religious practices." These accounts, she said, "pointed to the basic reality of how robbing people of their economic liberty also robs them of all sorts of other freedoms."

The FTC estimates about 30 million people, or one in five American workers, from minimum wage earners to CEOs, are bound by noncompetes. It says the policy change could lead to increased wages totaling nearly $300 billion per year by encouraging people to swap jobs freely. The ban, which will take effect later this year, carves out an exception for existing noncompetes that companies have given their senior executives, on the grounds that these agreements are more likely to have been negotiated. The FTC says employers should not enforce other existing noncompete agreements.

An anonymous reader quotes a report from the New York Times: Generative A.I. technologies can write poetry and computer programs or create images of teddy bears and videos of cartoon characters that look like something from a Hollywood movie. Now, new A.I. technology is generating blueprints for microscopic biological mechanisms that can edit your DNA, pointing to a future when scientists can battle illness and diseases with even greater precision and speed than they can today. Described in a research paper published on Monday by a Berkeley, Calif., startup called Profluent, the technology is based on the same methods that drive ChatGPT, the online chatbot that launched the A.I. boom after its release in 2022. The company is expected to present the paper next month at the annual meeting of the American Society of Gene and Cell Therapy. "Its OpenCRISPR-1 protein is built on a similar structure as the fabled CRISPR-Cas9 DNA snipper, but with hundreds of mutations that help reduce its off-target effects by 95%," reports Fierce Biotech, citing the company's preprint manuscript published on BioRxiv. "Profluent said it can be employed as a 'drop-in replacement' in any experiment calling for a Cas9-like molecule."

While Profluent will keep its LLM generators private, the startup says it will open-source the products of this initiative. "Attempting to edit human DNA with an AI-designed biological system was a scientific moonshot," Profluent co-founder and CEO Ali Madani, Ph.D., said in a statement. "Our success points to a future where AI precisely designs what is needed to create a range of bespoke cures for disease. To spur innovation and democratization in gene editing, with the goal of pulling this future forward, we are open-sourcing the products of this initiative."

Nicola Davis reports via The Guardian: If every day appears to go in a blur, try seeking out new and interesting experiences, researchers have suggested, after finding memorable images appear to dilate time. Researchers have previously found louder experiences seem to last longer, while focusing on the clock also makes time dilate, or drag. Now researchers have discovered the more memorable an image, the more likely a person is to think they have been looking at it for longer than they actually have. Such images were also easier for participants to recall the next day. Prof Martin Wiener, co-author of the study who is based at George Mason University in the U.S., said the findings could help develop improve artificial intelligence that interacts with humans, while they also offer opportunities to tweak our perceptions, given research has previously shown non-invasive brain stimulation can be used to lengthen a perceived interval.

The results from two groups, totaling about 100 people, revealed participants were more likely to think they had been looking at small, highly cluttered scenes -- such a crammed pantry -- for a shorter duration than was the case, whereas the reverse occurred when people viewed large scenes with little clutter, such as the interior of an aircraft hangar. The team also carried out experiments involving 69 participants that found images known from previous work to be more memorable were more likely to be judged as having been shown for longer than was the case. Crucially, the effect seemed to go both ways.

"We also found that the longer the perceived subjective duration of an image, the more likely you were to remember it the next day," said Wiener. When the team carried out an analysis using deep learning models of the visual system, they discovered more memorable images were processed faster. What's more, the processing speed for an image was correlated with how long participants thought they had been looking at it. "Images may be more memorable because they are processed faster and more efficiently in the visual system, and that drives the perception of time," said Wiener. The team suggest time dilation might serve a purpose, enabling us to gather information about the world around us. The findings have been published in the journal Nature Human Behavior.