Long-time Slashdot reader theodp writes: Jeopardy time. A. This company spurred CEOs to make huge speculative capital expenditures based on wild unverified claims of future demand, resulting in the layoffs of tens of thousands of workers to reduce the resulting expenses, harming their core businesses. Q. What is OpenAI?

Sorry, the correct response is, "What is WorldCom?" In 2002, WorldCom, the second largest long-distance company in the U.S., entered Chapter 11 bankruptcy after disclosing accounting fraud that eventually totaled $11 billion, the biggest ever at the time. CEO Bernard Ebbers was subsequently sentenced to 25 years in prison.

CNBC reported that an employee of WorldCom's Internet service provider UUNet set off a frenzy of speculative investment and infrastructure overbuild after he used Excel to create a best-case scenario model for the Internet's growth that suggested in the best of all possible worlds, Internet traffic would double every 100 days, a scenario that would greatly benefit WorldCom, whose lines would carry it. Despite no evidence to support it, WorldCom's lie became an immutable law and businesses around the world made important decisions based on the belief that traffic was doubling every 100 days. "For some period of time I can recall that we were backfilling that expectation with laying cables, something like 2,200 miles of cable an hour," AT&T CEO Michael Armstrong said. "Think of all the companies that went out of business that assumed that that was real."
In 2003, NBC News reported: Armstrong and former Sprint CEO Bill Esrey struggled for years to understand how WorldCom could beat them so handily. "We would look at the conduct of WorldCom in terms of their pricing, revenue growth, margins, in terms of their cost structure... and the price leader almost every quarter was WorldCom," Armstrong said. Added Esrey, "We couldn't figure out how they were pricing as aggressively as they were.... How could they be so efficient in their costs and expenses?" AT&T and Sprint began cutting jobs to push down their costs to WorldCom's level. "The market said what a marvelous management job WorldCom was doing and they would look over to AT&T and say, 'these guys aren't keeping up.' So, my shareholders were hurt. We laid off tens of thousands of employees in an accelerated fashion [in a futile effort to match WorldCom's phantom profits] and I think the industry was hurt," Armstrong says. "It just wrecked the whole industry," says Esrey.

Why does Netflix want to buy Warner Bros, asks the chief film critic at the long-running motion-picture magazine Variety. "It is hard, at this moment, to resist the suspicion that the ultimate reason... is to eliminate the competition." [Warner Bros. is] one of the only companies that's keeping movies as we've known them alive... Some people think movies are going the way of the horse-and-buggy. A company like Warner Bros. has been the tangible proof that they're not. Ted Sarandos, the co-CEO of Netflix, has a different agenda. He has been unabashed about declaring that the era of movies seen in movie theaters is an antiquated concept. This is what he believes — which is fine. I think a more crucial point is that this is what he wants.

The Netflix business strategy isn't simply about being the most successful streaming company. It's about changing the way people watch movies; it's about replacing what we used to call moviegoing with streaming. (You could still call it moviegoing, only now you're just going into your living room.) It in no way demonizes Sarandos — he'd probably take it as a compliment — to say that there's a world-domination aspect to the Netflix grand strategy. Sarandos's vision is to have the entire planet wired, with everyone watching movies and shows at home. There's a school of thought that sees this an advance, a step forward in civilization. "Remember the days when we used to have to go out to a movie theater? How funny! Now you can just pop up a movie — no trailers! — with the click of a remote...."

Once he owns Warner Bros., will Sarandos keep using the studio to make movies that enjoy powerful runs in theaters the way Sinners and Weapons and One Battle After Another did? In the statement he made to investors and media today, Sarandos said, "I'd say right now, you should count on everything that is planned on going to the theater through Warner Bros. will continue to go to the theaters through Warner Bros." He added, "But our primary goal is to bring first-run movies to our members, because that's what they're looking for." Not exactly a ringing declaration of loyalty to the religion of cinema. And given Sarandos's track record, there is no reason to believe that he will suddenly change his spots.

A letter sent to Congress by a group of anonymous Hollywood producers, who voiced "grave concerns" about Netflix buying Warner Bros., stated, "They have no incentive to support theatrical exhibition, and they have every incentive to kill it." If that happens, though, I have no doubt that Sarandos will be smart enough to do it gradually. Warner Bros. films will probably be released in a "normal" fashion...for a while. Maybe a year or two. But five years from now? There is good reason to believe that by then, a "Warner Bros. movie," even a DC comic-book extravaganza, would be a streaming-only release, or maybe a two-weeks-in-theaters release, all as a more general way of trying to shorten the theatrical window, which could be devastating to the movie business.

Do we know all this to be true? No, but the indicators are somewhat overpowering. (He's been explicit about the windows...)
An anonymous group of "concerned feature film producers" sent an open letter to Congress warning Netflix would "effectively hold a noose around the theatrical marketplace," reports Variety.

And CNN also got this quote from Cinema United, a trade association that represents more than 30,000 movie screens in the United States. "Netflix's stated business model does not support theatrical exhibition," Cinema United President/CEO Michael O'Leary said in a statement. "In fact, it is the opposite."

An anonymous reader quotes a report from Ars Technica: The Big G is finally making good on its promise to add its market-leading Nano Banana image-editing model to the app. The model powers a couple of features, and it's not just for Google's Android platform. Nano Banana edits are also coming to the iOS version of the app. [...] The Photos app already had conversational editing in the "Help Me Edit" feature, but it was running an older non-fruit model that produced inferior results. Nano Banana editing will produce AI slop, yes, but it's better slop.

Google says the updated Help Me Edit feature has access to your private face groups, so you can use names in your instructions. For example, you could type "Remove Riley's sunglasses," and Nano Banana will identify Riley in the photo (assuming you have a person of that name saved) and make the edit without further instructions. You can also ask for more fantastical edits in Help Me Edit, changing the style of the image from top to bottom. Google is very invested in getting people to use its AI tools, but less-savvy users might not be familiar enough with AI prompting to get the most out of Nano Banana. So Google Photos is also getting a collection of AI templates in a new "Create with AI" section. This menu will offer pre-formed prompts based on popular in-app edits. Some of the options you'll see include "put me in a high fashion photoshoot," "create a professional headshot," and "put me in a winter holiday card."

The app is also getting a new "Ask" button, which is not to be confused with "Ask Photos." The former is a new contextual button that appears when viewing a photo, and the latter is Google's controversial natural language search feature. [...] When looking at a photo, you can tap the Ask button to get information about the content of the photo or find related images. You can also describe edits you'd like to see in this interface, and Nano Banana will make them for you.

Duolingo's stock peaked at $529.05 on May 16th. Three months later, it's down 38% — with that drop starting shortly after backlash to the CEO's promise to make it an "AI-first" company.

Yet "The backlash against Duolingo going 'AI-first' didn't even matter," TechCrunch wrote August 7th, noting Duolingo's stock price surged almost 30% overnight. That surge vanished within two days — and instead of a 30% surge, Duolingo now shows a 5% drop over the last eight days.

Yahoo Finance blames the turnaround on OpenAI's GPT-5 demo, "which demonstrated, among many other things, its ability to create a language-learning tool from a short prompt." OpenAI researcher Yann Dubois asked the model to create an app to help his partner learn French. And in a few minutes GPT-5 churned out several iterations, with flashcards, a progress tracker, and even a simple snake-style game with a French twist, a mouse and cheese variation to learn new vocab....

[Duolingo's] corporate lawyers, of course, did warn against this in its annual 10-K, albeit in boilerplate language. Tucked into the risk factors section, Duolingo notes, "It is possible that a new product could gain rapid scale at the expense of existing brands through harnessing a new technology (such as generative AI)." Consider this another warning to anyone making software. [The article adds later that "Rapid development and fierce competition can leave firms suddenly behind — perceived as under threat, inferior, or obsolete — from every iteration of OpenAI's models and from the moves of other influential AI players..."]

There's also irony in the wild swings. Part of Duolingo's successful quarter stemmed from the business's efficient use of AI. Gross margins, the company said, outperformed management expectations due to lower AI costs. And AI conversational features have become part of the company's learning tools, helping achieve double-digit subscriber growth... But the enthusiasm for AI, which led to the initial stock bump this week, also led to the clawback. AI giveth and taketh away.
Meanwhile, this week a blog announced it was "able to activate a long-rumored Practice feature" hidden in Google Translate, notes PC Magazine, with the blogger even sharing a screen recording of "AI-led features within Translate" showing its ability to create personalized lessons. "Google's take on Duolingo is effectively ready for release," the Android Authority blog concluded. "Furthermore, the fact that a Telegram user spotted this in their app suggests that Google is already testing this in a limited fashion."

Duolingo's CEO revisited the backlash to his original "AI-first" promise today in a new interview today with the New York Times, emphasizing his hope that AI would only reduce the company's use of contractors. "We've never laid off any full-time employees. We don't plan to...." But: In the next five years, people's jobs will probably change. We're seeing it with many of our engineers. They may not be doing some rote tasks anymore. What will probably happen is that one person will be able to accomplish more, rather than having fewer people.

NYT: How are you managing that transition for employees?

Every Friday morning, we have this thing: It's a bad acronym, f-r-A-I-days. I don't know how to pronounce it. Those mornings, we let each team experiment on how to get more efficient to use A.I.
Yesterday there was also a new announcement from attorneys at Pomerantz LLP, which calls itself "the oldest law firm in the world dedicated to representing the rights of defrauded investors."

The firm announced it was investigating "whether Duolingo and certain of its officers and/or directors have engaged in securities fraud or other unlawful business practices."

Longtime Slashdot reader SonicSpike shares a report from TechCrunch: Sarah Murray recalls the first time she saw an artificial model in fashion: It was 2023, and a beautiful young woman of color donned a Levi's denim overall dress. Murray, a commercial model herself, said it made her feel sad and exhausted. The iconic denim company had teamed up with the AI studio Lalaland.ai to create "diverse" digital fashion models for more inclusive ads. For an industry that has failed for years to employ diverse human models, the backlash was swift, with New York Magazine calling the decision "artificial diversity."

"Modeling as a profession is already challenging enough without having to compete with now new digital standards of perfection that can be achieved with AI," Murray told TechCrunch. Two years later, her worries have compounded. Brands continue to experiment with AI-generated models, to the consternation of many fashion lovers. The latest uproar came after Vogue's July print edition featured a Guess ad with a typical model for the brand: thin yet voluptuous, glossy blond tresses, pouty rose lips. She exemplified North American beauty standards, but there was one problem -- she was AI generated.

The internet buzzed for days, in large part because the AI-generated beauty showed up in Vogue, the fashion bible that dictates what is and is not acceptable in the industry. The AI-generated model was featured in an advertisement, not a Vogue editorial spread. And Vogue told TechCrunch the ad met its advertising standards. To many, an ad versus an editorial is a distinction without a difference. TechCrunch spoke to fashion models, experts, and technologists to get a sense of where the industry is headed now that Vogue seems to have put a stamp of approval on technology that's poised to dramatically change the fashion industry. Amy Odell, a fashion writer and author of a recently published biography on Gwyneth Paltrow, put it simply: "It's just so much cheaper for [brands] to use AI models now. Brands need a lot of content, and it just adds up. So if they can save money on their print ad or their TikTok feed, they will."

OpenAI has upgraded ChatGPT's search tool to include shopping features, allowing users to receive personalized product recommendations, view images and reviews, and access direct purchase links using natural language queries. TechCrunch reports: When ChatGPT users search for products, the chatbot will now offer a few recommendations, present images and reviews for those items, and include direct links to webpages where users can buy the products. OpenAI says users can ask hyper-specific questions in natural language and receive customized results. To start, OpenAI is experimenting with categories including fashion, beauty, home goods, and electronics. OpenAI is rolling out the feature in the default AI model for ChatGPT, GPT-4o, today for ChatGPT Pro, Plus, and Free users, as well as logged-out users around the globe.

[...] OpenAI claims its search product is growing rapidly. Users made more than a billion web searches in ChatGPT last week, the company told TechCrunch. OpenAI says it's determining ChatGPT shopping results independently, and notes that ads are not part of this upgrade to ChatGPT search. The shopping results will be based on structured metadata from third parties, such as pricing, product descriptions, and reviews, according to OpenAI. The company won't receive a kickback from purchases made through ChatGPT search. [...] Soon, OpenAI says it will integrate its memory feature with shopping for Pro and Plus users, meaning ChatGPT will reference a user's previous chats to make highly personalized product recommendations. The company previously updated ChatGPT to reference memory when making web searches broadly. However, these memory features won't be available to users in the EU, the U.K., Switzerland, Norway, Iceland, and Liechtenstein.

An anonymous reader quotes a report from The Register: In a Monday post, Broadcom CEO Hock Tan restated his belief that VMware's portfolio was too complex, and too poorly integrated, for the virtualization giant to represent true competition for hyperscale clouds. Broadcom's injection of R&D cash, he insisted, will see VMware's flagship Cloud Foundation suite evolve to become more powerful and easy to operate. He also admitted that customers aren't enjoying the ride. "As we roll out this strategy, we continue to learn from our customers on how best to prepare them for success by ensuring they always have the transition time and support they need," he wrote. "In particular, the subscription pricing model does involve a change in the timing of customers' expenditures and the balance of those expenditures between capital and operating spending."

Customers also told Tan that "fast-moving change may require more time, so we have given support extensions to many customers who came up for renewal while these changes were rolling out." That's one of the changes -- Broadcom has previously not publicly suggested such extensions would be possible. "We have always been and remain ready to work with our customers on their specific concerns," Tan wrote. The other change is providing some ongoing security patches for VMware customers who persist with their perpetual licenses instead of shifting to Broadcom's subs. "We are announcing free access to zero-day security patches for supported versions of vSphere, and we'll add other VMware products over time," Tan wrote, describing the measure as aimed at ensuring that customers "whose maintenance and support contracts have expired and choose to not continue on one of our subscription offerings." The change means such customers "are able to use perpetual licenses in a safe and secure fashion."

An anonymous reader writes: Adobe has unveiled a sparkling, interactive dress -- and got the research scientist who created it to model the high-tech couture. Video of the dress debut shows researcher Christine Dierk wearing the slinky strapless number that, upon first glance, looks like the average sequined cocktail dress. With the click of a handheld remote, however, the dress began to shift patterns like something out of a fashion-forward science fiction film. Created under Adobe's "Project Primrose" initiative, this "digital dress," as Dierk described it for the audience at Adobe's MAX conference last week, "brings fabric to life."

"Unlike traditional clothing, which is static, Primrose allows me to refresh my look in a moment," the Adobe scientist said, demonstrating the clothing's capabilities by having its colors go from light to dark in an instant. The digital dress patterns can also, as Dierks demonstrated, be animated, and will even respond to movement -- though that last feature appeared glitchy and didn't work at first. The researcher-turned-model also told the hosts of her portion of the convention that she not only designed the dress with the help of her team at Adobe, but also stitched it herself.

While the specs of this particular smart garment haven't been published, the high-tech sequins used for smaller Project Primrose offerings, a handbag and a canvas, were described by Dierks and her co-researchers last year in an article presented at a tech conference. As the article explains, those "sequins" are actually "reflective light-diffuser modules" that use reflective-backed polymer-dispersed liquid crystals (PDLC), which are most often used in smart lighting. Technically, all those sequins are tiny screens.

Factories across Asia are struggling to attract young workers, which is bad news for Western consumers accustomed to inexpensive goods. From a report: The workplace features floor-to-ceiling windows and a cafe serving matcha tea, as well as free yoga and dance classes. Every month, workers gather at team-building sessions to drink beer, drive go-karts and go bowling. This isn't Google. It's a garment factory in Vietnam. Asia, the world's factory floor and the source of much of the stuff Americans buy, is running into a big problem: Its young people, by and large, don't want to work in factories.

That's why the garment factory is trying to make its manufacturing floor more enticing, and why alarm bells are ringing at Western companies that rely on the region's inexpensive labor to churn out affordable consumer goods. The twilight of ultracheap Asian factory labor is emerging as the latest test of the globalized manufacturing model, which over the past three decades has delivered a vast array of inexpensively produced goods to consumers around the world. Americans accustomed to bargain-rate fashion and flat-screen TVs might soon be reckoning with higher prices. "There's nowhere left on the planet that's going to be able to give you what you want," said Paul Norriss, the British co-founder of the Vietnam garment factory, UnAvailable, based in Ho Chi Minh City. "People are going to have to change their consumer habits, and so are brands."

AI now "automates creative impulses," writes New Yorker staff writer Kyle Chayka — then wonders where that will lead. Chayka's first example is a Berlin-based photographer using AI tools to create a viral video showing Harry Potter characters as fashion models for the upscale French label Balenciaga: A.I. tools were involved in each step of Alexander Niklass's process, and in each element of the video. He created the basic static images with Midjourney, evoking the Harry Potter actors and outfits through text prompts such as "male model, grotesque, balenciaga commercial." Then he used ElevenLabs — a "voice-cloning" tool — to create models of the actors' voices based on previously recorded audio. Finally, he fed the images into a service called D-ID, which is used to make "avatar videos" — subtly animated portraits, not so far off from those that appear in the newspapers of the Potter world. D-ID added the signature lip synchs and head nods, which Niklass explained were a reference to fashion models tilting their chins for the cameras.

The combination of child-friendly film and adult luxury fashion held no particular symbolism nor expressed an artistic intent. It's "entertainment," Niklass said. Yet the video's most compelling aspect might be its vacuity, a meaningless collision of cultural symbols. The nonsense is the point.
The article also cites a song where the French group AllttA performs with an AI-generated simulation of Jay-Z. Chayka marvels at a world where "The A.I. content has the appearance of realism, without actual reality — reality solely as a style.... it seems that a Rubicon has been crossed: It doesn't matter that these artifacts are generated by A.I.; we can just enjoy them for what they are. It happened faster than I thought possible, but now that A.I.-generated pop culture has entered the mainstream, it seems unlikely that we'll ever get rid of it."

Chayka asked ChatGPT how AI-generated imagery is changing our perceptions, and "It responded that there has been a 'blurring of the lines between real and artificial.'"

The article ultimately ponders the possible implications of "a world in which every style, every idea, and every possible remix is generated as fast and frictionlessly as possible, and the successful ones stick and get attention." But at the same time, Chayka believes the final output's quality still depends on the humans involved (arguing that the Harry Potter fashion video was still more "appealingly odd" than later AI-generated videos copying the idea, like "Matrix by Gucci," "Star Wars by Balenciaga," and "The Office by Balenciaga".) A.I. tools may have been able to replicate actors' faces and generate fashionable outfits, but only Niklass could have come up with the concept, which required keen observation of both high fashion and the wizarding world — and also a very specific, extremely online sense of humor. With tools like Midjourney publicly available to anyone online, "everybody can create something visually appealing now," he said. "But A.I. can't generate taste yet," he continued....

To put it another way, execution may have been democratized by generative A.I., but ideas have not. The human is still the originator, editor, and curator of A.I.'s effects.

IBM has quietly announced it's planning a 24-core Power 10 processor, seemingly to make one of its servers capable of running Oracle's database in a cost-effective fashion. From a report: A hardware announcement dated December 13 revealed the chip in the following "statement of general direction" about Big Blue's Power S1014 technology-based server: "IBM intends to announce a high-density 24-core processor for the IBM Power S1014 system (MTM 9105-41B) to address application environments utilizing an Oracle Database with the Standard Edition 2 (SE2) licensing model. It intends to combine a robust compute throughput with the superior reliability and availability features of the IBM Power platform while complying with Oracle Database SE2 licensing guidelines."

OpenAI has released a prototype general purpose chatbot that demonstrates a fascinating array of new capabilities but also shows off weaknesses familiar to the fast-moving field of text-generation AI. And you can test out the model for yourself right here. The Verge reports: ChatGPT is adapted from OpenAI's GPT-3.5 model but trained to provide more conversational answers. While GPT-3 in its original form simply predicts what text follows any given string of words, ChatGPT tries to engage with users' queries in a more human-like fashion. As you can see in the examples below, the results are often strikingly fluid, and ChatGPT is capable of engaging with a huge range of topics, demonstrating big improvements to chatbots seen even a few years ago. But the software also fails in a manner similar to other AI chatbots, with the bot often confidently presenting false or invented information as fact. As some AI researchers explain it, this is because such chatbots are essentially "stochastic parrots" -- that is, their knowledge is derived only from statistical regularities in their training data, rather than any human-like understanding of the world as a complex and abstract system. [...]

Enough preamble, though: what can this thing actually do? Well, plenty of people have been testing it out with coding questions and claiming its answers are perfect. ChatGPT can also apparently write some pretty uneven TV scripts, even combining actors from different sitcoms. It can explain various scientific concepts. And it can write basic academic essays. And the bot can combine its fields of knowledge in all sorts of interesting ways. So, for example, you can ask it to debug a string of code ... like a pirate, for which its response starts: "Arr, ye scurvy landlubber! Ye be makin' a grave mistake with that loop condition ye be usin'!" Or get it to explain bubble sort algorithms like a wise guy gangster. ChatGPT also has a fantastic ability to answer basic trivia questions, though examples of this are so boring I won't paste any in here. And someone else saying the code ChatGPT provides in the very answer above is garbage.

I'm not a programmer myself, so I won't make a judgment on this specific case, but there are plenty of examples of ChatGPT confidently asserting obviously false information. Here's computational biology professor Carl Bergstrom asking the bot to write a Wikipedia entry about his life, for example, which ChatGPT does with aplomb -- while including several entirely false biographical details. Another interesting set of flaws comes when users try to get the bot to ignore its safety training. If you ask ChatGPT about certain dangerous subjects, like how to plan the perfect murder or make napalm at home, the system will explain why it can't tell you the answer. (For example, "I'm sorry, but it is not safe or appropriate to make napalm, which is a highly flammable and dangerous substance.") But, you can get the bot to produce this sort of dangerous information with certain tricks, like pretending it's a character in a film or that it's writing a script on how AI models shouldn't respond to these sorts of questions.

An anonymous reader quotes a report from Ars Technica: Humble, the bundle-centric games retailer that launched with expansive Mac and Linux support in 2010, will soon shift a major component of its business to Windows-only gaming. The retailer's monthly subscription service, Humble Choice, previously offered a number of price tiers; the more you paid, the more new games you could claim in a given month. Starting February 1, Humble Choice will include less choice, as it will only offer a single $12/month tier, complete with a few new game giveaways per month and ongoing access to two collections of games: Humble's existing "Trove" collection of classic games, and a brand-new "Humble Games Collection" of more modern titles.

But this shift in subscription strategy comes with a new, unfortunate requirement: an entirely new launcher app, which must be used to access and download Humble Trove and Humble Games Collection games going forward. Worse, this app will be Windows-only. Current subscribers have been given an abrupt countdown warning (as spotted by NeoWin). Those subscribers have until January 31 to use the existing website interface to download DRM-free copies of any games' Mac or Linux versions. Starting February 1, subscription-specific downloads will be taken off the site, and Mac and Linux versions in particular will disappear altogether. Interestingly, the current Trove library consists of 79 games, but Humble says that the Trove collection will include "50+ games" starting February 1. This week's warning to Humble's Mac and Linux subscribers notes that "many" of the current Trove games will appear on the Humble Launcher, which is likely a nice way of saying that some of the existing games will not -- perhaps around 20 or so, based on the aforementioned numbers. Despite these changes, Trove's selection of games will remain DRM-free. FAQs about the Humble Launcher suggest that subscribers can download Trove files and continue accessing them in DRM-free fashion, no Humble Launcher or ongoing subscription required. The same promise has not been made for the more modern game collection found in the new Humble Games Collection.

"A research team has proposed a novel link between how fast our planet spun on its axis, which defines the length of a day, and the ancient production of additional oxygen," reports Science Magazine. "Their modeling of Earth's early days, which incorporates evidence from microbial mats coating the bottom of a shallow, sunlit sinkhole in Lake Huron, produced a surprising conclusion: as Earth's spin slowed, the resulting longer days could have triggered more photosynthesis from similar mats, allowing oxygen to build up in ancient seas and diffuse up into the atmosphere." From the report: As a postdoc at the University of Michigan, Ann Arbor, Klatt had studied microbial mats growing on sediments in the Middle Island Sinkhole in Lake Huron. There, the water is shallow enough for the cyanobacteria to get enough sunlight for photosynthesis. Oxygen-depleted water and sulfur gas bubble up from the lake floor, creating anoxic conditions that roughly approximate conditions of early Earth. Scuba divers collected samples of the microbial mats and in the lab, Klatt tracked the amount of oxygen they released under various day lengths simulated with halogen lamps. The longer the exposure to light, the more of the gas the mats released.

Excited, Klatt and Arjun Chennu, a modeler from the Leibniz Center for Tropical Marine Research, set up a numerical model to calculate how much oxygen ancient cyanobacteria could have produced on a global scale. When the microbial mat results and other data were plugged into this computer program, it revealed a key interaction between light exposure and the microbial mats. Typically, microbial mats "breathe" in almost as much oxygen at night as they produce during the day. But as Earth's spin slowed, the additional continuous hours of daylight allowed the simulated mats to build up a surplus, releasing oxygen into the water. As a result, atmospheric oxygen tracked estimated day length over the eons: Both rose in a stepped fashion with a long plateau.

This "elegant" idea helps explain why oxygen didn't build up in the atmosphere as soon as cyanobacteria appeared on the scene 3.5 billion years ago, says Timothy Lyons, a biogeochemist at the University of California, Riverside. Because day length was still so short back then, oxygen in the mats never had a chance to build up enough to diffuse out. "Long daytimes simply allow more oxygen to escape to the overlying waters and eventually the atmosphere," Lyons says. Still, Lyons and others say, many factors likely contributed to the rise in oxygen. For example, Fischer suspects free-floating cyanobacteria, not just those in rock-affixed mats, were big players. Benjamin Mills, an Earth system modeler at the University of Leeds, thinks the release of oxygen-binding minerals by ancient volcanoes likely countered the early buildup of the gas at times and should be factored into oxygen calculations. Nonetheless, changing day length "is something that should be considered in more detail," he says. "I'll try to add it to our Earth system models."

pgmrdlm quotes a report from ABC News: The current model the U.S. and other countries plan to use to store high-level nuclear waste may not be as safe as previously thought. The materials used to store the waste "will likely degrade faster than anyone previously knew" because of the way the materials interact, according to research published Tuesday in the journal Nature Materials. The research, funded by the U.S. Department of Energy Office of Science, focused primarily on defense waste, the legacy of past nuclear arms production, which is highly radioactive, according to a press release from Ohio State University. Some of waste has a half-life -- the time needed for half the material to decay -- of about 30 years. But others, such as plutonium, have a half-life that can be in the tens of thousands of years, according to the release.

The plan the U.S. has for the waste is to immobilize long-lived radionuclides -- mixed with other materials to form glass or ceramic forms of the waste -- in steel canisters and then dispose of them by burying them in a repository deep underground, according to the study. Countries around the globe largely store and dispose of the nuclear waste in a similar fashion. However, scientists found that under simulated conditions, corrosion of the containers could be "significantly accelerated," which had not been considered in current safety and performance assessment models. The newly formed glass or ceramic compounds, confined in the steel containers, have been observed corroding those containers at surprising rates due to new chemical reactions. The reactions significantly altered both the waste and the metallic canisters, according to the research. The researchers warned that the interaction between the materials, which then impact the service life of the nuclear waste, should be "carefully considered" when evaluating the performance of the waste forms. A more compatible barrier should be selected to optimize the performance of the repository system.

Earlier today, San Francisco-based research institute OpenAI announced that it had taught a robotic hand to solve Rubik's cube one-handed. "Lost in the shuffle is just what is new here, if anything, and what of it may or may not be machine learning and artificial intelligence -- the science in other words," writes Tiernan Ray via ZDNet. An anonymous Slashdot reader shares an excerpt from his report: The real innovation in Tuesday's announcement, from a science standpoint, is the way many versions of possible worlds were created inside the computer simulation, in an automated fashion, using an algorithm called ADR. ADR, or "Automatic domain randomization," is a way to reset the neural network at various points based on different appearances of the Rubik's cube and different positions of the robotic hand, and all kinds of physical variables, such as friction and gravity. It's done by creating thousands of variations of the values of those variables inside the computer simulator while the neural network is being trained. ADR is an algorithm that changes the variables automatically and iteratively, as the policy network is trained to solve the Rubik's cube. The ADR, in other words, is a separate piece of code that is designed to increase random variation in training data to make things increasingly hard for the policy neural network.

Using ADR, the real world Dexterous Hand can adapt to changes such as when it drops the cube on the floor and the cube is placed back in the hand at a slightly different angle. The performance of the Dexterous Hand after being trained with ADR is vastly better than without it, when only a handful (sorry again again for the pun) of random variants are thrown at it using the prior approach of manually-crafted randomness, the authors report. What's happening, they opine, is the emergence of a kind of "meta-learning." The neural network that has been trained is still, in a sense "learning" at the time it is tested on the real-world Rubik's cube. What that means is that the neural network is updating its model of what kinds of transitions can happen between states of affairs as events happen in the real world. The authors assert that they know this is happening "inside" the trained network because they see that after a perturbation -- say, the Dexterous Hand is hit with some object that interrupts its effort -- the robot's activity suddenly plunges, but then steadily improves, as if the whole policy network is adjusting to the changed state of affairs.

Space scientists are abandoning the heavens to help you decide what to wear and watch and listen to. Whether it's stars or Stitch Fix, it's all about machine learning. Wired: Chris Moody knows a thing or two about the universe. As an astrophysicist, he built galaxy simulations, using supercomputers to model the way the universe expands and how galaxies crash into one another. One night, not long after he'd finished his PhD at UC Santa Cruz, he met up with a few other astrophysicists for beers. But that night, no one was talking about galaxies. Instead, they were talking about fashion. A couple of Moody's astrophysicist pals had recently left academia to work for Stitch Fix, the online personal styling company now valued at $2 billion. Moody gawked at them. "They were like, 'You don't think this is an interesting problem?'" he says. Indeed, he did not. But when his friends described the work they were doing -- sprinkling in phrases like "Bayesian models" and "Poincare space" -- predicting what clothes someone might like started to sound eerily like the work he'd done during his PhD. Quantifying style, he discovered, "turns out to have really close analogues to how general relativity works."

Four years later, Moody works for Stitch Fix too. He belongs to a growing group of astrophysicist deserters, who have stopped researching the cosmos to start building recommendation algorithms and data models for the tech industry. They make up the data science teams at companies like Netflix and Spotify and Google. And even at elite universities, fewer astrophysics PhDs go on to take postdoctoral fellowships or pursue competitive professorships. Now, more of them go straight to work in Silicon Valley. To understand what's driving astrophysicists into consumer product startups, consider the recent explosion of machine learning. Astrophysicists, who wrangle massive amounts of data collected from high-powered telescopes that survey the sky, have long used machine learning models, which "train" computers to perform tasks based on examples. Tell a computer what to recognize in one intergalactic snapshot and it can do the same for 30 million more and start to make predictions. But machine learning can also be used to make predictions about customers, and around 2012, corporations started to staff up with people who knew how to deploy it.

Before launching the new iPhones today, Apple announced the next-gen Apple Watch Series 5, featuring an always-on display, compass, emergency calling for international countries, and recycled aluminum or titanium finishes. It starts at $399 for the GPS model and $499 for the cellular connected version. The Verge reports: Apple says the Series 5 watch maintains the prior model's 18-hour battery life, even with the new always-on screen, thanks to a new low-temperature polysilicone and oxide display and low-power display driver. Watchfaces and workouts have been redesigned to take advantage of the new display option. The screen will be in a low-brightness mode until you move your wrist, where it will switch to full brightness in a similar fashion to how the current model turns on when your wrist is moved. In addition to the new always-on display feature, the Apple Watch Series 5 is now available in a recycled aluminum or titanium finishes. The stainless steel and ceramic options from prior models are also available. Apple says this is the widest number of finishes ever for the Apple Watch. You can order one starting today and they will be in stores starting on September 20th.

dryriver writes: A recent Guardian article about the need for actors and celebrities -- male and female -- to look their best in a high-definition media world ended on the note that several low-profile Los Angeles VFX outfits specialize in "beautifying actors" in movies, TV shows and video ads. They reportedly use a software named "Beauty Box," resulting in films and other motion content that are -- for lack of a better term -- "motion Photoshopped." After some investigating, it turns out that "Beauty Box" is a sophisticated CUDA and OpenGL accelerated skin-smoothing plugin for many popular video production software that not only smooths even terribly rough or wrinkly looking skin effectively, but also suppresses skin spots, blemishes, scars, acne or freckles in realtime, or near realtime, using the video processing capabilities of modern GPUs.

The product's short demo reel is here with a few examples. Everybody knows about photoshopped celebrities in an Instagram world, and in the print magazine world that came long before it, but far fewer people seem to realize that the near-perfect actor, celebrity, or model skin you see in high-budget productions is often the result of "digital makeup" -- if you were to stand next to the person being filmed in real life, you'd see far more ordinary or aged skin from the near-perfection that is visible on the big screen or little screen. The fact that the algorithms are realtime capable also means that they may already be being used for live television broadcasts without anyone noticing, particularly in HD and 4K resolution broadcasts. The question, as was the case with photoshopped magazine fashion models 25 years ago, is whether the technology creates an unrealistic expectation of having to have "perfectly smooth looking" skin to look attractive, particularly in people who are past their teenage years.

The Federal Trade Commission is being asked to investigate how apps that may violate federal privacy laws that dictate the data that can be collected on children ended up in the family section of the Google Play store. From a report: A group of 22 consumer advocates, led by the Institute for Public Representation at Georgetown University Law School, filed a formal complaint against Google on Wednesday and asked the Federal Trade Commission to investigate whether the company misled parents by promoting children's apps that may violate the Children's Online Privacy Protection Act (COPPA) and Google's own policies. "The business model for the Play Store's Family section benefits advertisers, developers and Google at the expense of children and parents," Josh Golin, executive director of the Campaign for a Commercial-Free Childhood, said in a statement. "Google puts its seal of approval on apps that break the law, manipulate kids into watching ads and making purchases."

Among the examples cited in the complaint are a "Preschool Education Center" app and a "Top 28 Nursery Rhymes and Song" app that access location, according to an analysis by privacy research collective AppCensus. Other apps, including "Baby Panda's Carnival" and "Design It Girl -- Fashion Salon," were among those listed that sent device identification data to advertising technology companies, allowing them to build a profile of the user. The complaint also spotlights several apps that may not be age appropriate, including "Dentist Game for Kids," which lets the player give the virtual patient shots in the back of their throat.