An anonymous reader quotes a report from VICE: It's early on a Sunday morning in late 1994, and you're shuffling your way through Fitzrovia in Central London, bloodstream still rushing after a long night at Bagley's. The sun comes up as you come down. You navigate side streets that you know like the back of your hand. But your hand's stamped with a party logo. And your brain's kaput. Coffee... yes, coffee. Good idea. Suddenly, you find yourself outside a teal blue cafe. Walking in is like entering an alien world; rows of club kids, tech heads, and game developers sit in front of desktops, lost in the primitive version of some new reality. Tentacular cables hang from the ceiling. Ambient techno reverberates from wall to wall. Cigarette smoke fills the air.

Welcome to Cyberia, the world's first internet cafe. Which, if you're too young to remember, are basically cafes with computers in them. It all began when Eva Pascoe, a Polish computing student living in London, crossed paths with Tim Berners Lee and other early internet mavericks at the dawn of the 90s. "I was very interested in cyberfeminism and wanted to figure out how women could reclaim tech," she recalls. The internet was still in its infancy. Diabolically slow dial-up modems only emerged around 1992; the World Wide Web was a pipe dream until 1993 and hardly anyone had the internet at home. But there wasn't just a lack of javascript; Eva remembers there being no good java, either. "There were no coffee shops in London," she says, which today seems ludicrous. "Just greasy spoons and everyone drank tea. I wanted a European-style cafe."

Linking up with like-minded pioneers David Rowe and husband and wife Keith and Gene Teare, Eva found a spot on the corner of Whitfield Street and launched Cyberia there in 1994. With Hackers-style aesthetics and futuristic furniture, it was based around a U-shaped layout that meant visitors could see each other's screens. "I wanted women to feel safe, because a lot of the stuff on the net was dodgy," she explains. Many of Eva's mates chipped in to help out -- architects, interior designers, graphic artists, publishers, and ravers among them.

And then there was the Amish community in Pennsylvania. Eva had to fly out there to negotiate for the "Cyberia.com" domain name they had bought. "It was a proper barn with horse carts and a wall of modems as they were running a bulletin board and an early ecommerce company. Apparently, there was always one family nominated to be the tech support," she remembers. Back in London, Cyberia quickly became a hotspot. "Virtually the second we opened, we had three lines deep around the block," she says. It's hard to imagine, but nowhere else in the world was doing what they were doing. It was the world's first cybercafe. "If you wanted to collect your emails, we were the only place in town," Eva says. Cyberia opened around 20 cafes worldwide, including branches in Bangkok, Paris, and Rotterdam. "For a fleeting moment it became like a sexier version of Richard Branson's Virgin empire: there was Cyberia Records, Cyberia Channel (a pioneering streaming service), Cyberia Payments, the Cyberia magazine, a Cyberia show on UK TV -- even a Cyberia wedding," writes VICE's Kyle MacNeill. He attended Cyberia's 30th birthday party in September and spoke with some of the cafe's original innovators, "shooting the shit about the good times and the not-so-good coffee."

Slashdot readers damn_registrars and GFS666 share the news of the passing of Thomas E. Kurtz, co-inventor of the BASIC programming language back in the 1960s. He was 96. Hackaday reports: The origins of BASIC lie in the Dartmouth Timesharing System, like similar timesharing operating systems of the day, designed to allow the resources of a single computer to be shared across many terminals. In this case the computer was at Dartmouth College, and BASIC was designed to be a language with which software could be written by average students who perhaps didn't have a computing background. In the decade that followed it proved ideal for the new microcomputers, and few were the home computers of the era which didn't boot into some form of BASIC interpreter. Kurtz continued his work as a distinguished academic and educator until his retirement in 1993, but throughout he remained as the guiding hand of the language.

Lawrence Livermore National Laboratory's "El Capitan" supercomputer is now ranked as the world's most powerful, exceeding a High-Performance Linpack (HPL) score of 1.742 exaflops on the latest Top500 list. Engadget reports: El Capitan is only the third "exascale" computer, meaning it can perform more than a quintillion calculations in a second. The other two, called Frontier and Aurora, claim the second and third place slots on the TOP500 now. Unsurprisingly, all of these massive machines live within government research facilities: El Capitan is housed at Lawrence Livermore National Laboratory; Frontier is at Oak Ridge National Laboratory; Argonne National Laboratory claims Aurora. [Cray Computing] had a hand in all three systems.

El Capitan has more than 11 million combined CPU and GPU cores based on AMD 4th-gen EPYC processors. These 24-core processors are rated at 1.8GHz each and have AMD Instinct M1300A APUs. It's also relatively efficient, as such systems go, squeezing out an estimated 58.89 Gigaflops per watt. If you're wondering what El Capitan is built for, the answer is addressing nuclear stockpile safety, but it can also be used for nuclear counterterrorism.

A new six-episode Dune series premiers tonight on HBO and Max — a prequel to the Denis Villeneuve-directed Dune movies set 10,000 years before the birth f Paul Atreides. The Hollywood Reporter writes that it "draws on source material from the 2012 novel Sisterhood of Dune by Brian Herbert and Kevin J. Anderson, and Frank Herbert's 1965 novel Dune, the origin of the Dune universe." Cord-cutters can stream Dune: Prophecy online without cable on Max, with subscriptions starting at $9.99 per month through both Prime Video and the Max website directly. Amazon offers a seven-day free trial to the Max channel. Those who want to watch Dune: Prophecy online without a traditional cable service can also get Max as an add-on to existing streaming services, including Hulu and DirecTV Stream.
The San Francisco Chronicle describes the series as "">all palace intrigues, agonizing deaths and magical mind games." Taking a further cue from the network's top-rated Game of Thrones, this show indulges more sex and nudity than the Dune movies allow. It could be argued that elements like this introduce a liveliness often missing from the portentous big-screen behemoths, marking an improvement. Another fun touch here: Many characters are constantly baked.

Set a millennium before Frank Herbert's novels and the films' events, and a century after humans overthrew their "thinking machine" overlords, the psychoactive "Spice" from the desert planet Arrakis is already the most valued substance in the universe. It's not only vital for spaceship navigation and to expand the mental powers of sorceressy sisterhoods like the Bene Gesserit, it's the club drug of choice for younger members of the galaxy-ruling Great Houses. As ever with "Dune" business, control of the Spice trade fuels much of the conflict and character motivations.

Of which there are just enough to keep things interesting without becoming confusing... While the show can't match the outsize visual scope of Denis Villeneuve's films, it does pleasingly approximate those vast alien landscapes, Brutalist edifices and high-ceilinged chambers on a TV budget. For those who find Villeneuve's formal gigantism oppressive, the series' more human scale might be another welcome change of pace... There may not be an original thought in this "Dune" product's Spice-soaked head, but it is one professionally put-together piece of this sort of entertainment.
"Tasked with making more material with less money and time, Prophecy cannot hope to equal Villeneuve's aesthetic accomplishments," writes Variety. "But at its best, the show does justice to the intricate politics and ethical debates that form a cornerstone of Frank Herbert's fictional universe... The primary Dune plot finds many echoes throughout Prophecy..."

On the other hand, Vulture argues the six-episode series is "stuck in prequel quicksand," even calling it "an act of cowardice and abdication of creativity" (while also noting moments where it "feels like it's stretching itself to be something other than what we expect..."

An anonymous reader quotes a report from The Guardian: This year's Web Summit, in Lisbon, was all about artificial intelligence -- and a robot sorting laundry. Digit, a humanoid built by the US firm Agility Robotics, demonstrated how far AI has come in a few years by responding to voice commands -- filtered through Google's Gemini AI model -- to sift through a pile of colored T-shirts and place them in a basket. It wasn't a seamless demonstration but the enthusiastic response, nearly two years on from the launch of ChatGPT, reflected the excitement about all things AI that pervaded Europe's biggest annual tech conference.

[...] Digit is being used in warehouses by GXO, a US logistics company, to lift boxes and place them on conveyor belts. According to the chief executive of Agility Robotics, Peggy Johnson, a new role could be created managing teams of Digits doing physical work. "Employees who were previously doing this physical work, appreciate the fact that they can hand that off to Digit," she said. "Then it allows them to do a number of other things, one of which is to be a robot manager." "Talk of a bust in the AI boom could not be heard over the shouts of encouragement for Digit as it pondered different shades of garment," reports The Guardian. "Nonetheless, the voices of caution were there, discussing familiar themes such as safety, jobs and the climate, as AI comes to influence a huge range of industries."

Back in 1976, Mattel Electronics Auto Race became the very first handheld game to use only solid-state electronics, according to Wikipedia. (Its only mechanical elements were its on/off switch and hand-operated controls...) Nearly half a century goes by — until the ancient and broken gizmo reaches long-time Slashdot reader Shayde, who "dove into disassembling the unit and figuring out the problem."

Ironically, at one point his voltimeter stopped working, because...its batteries were dead. But a tri-wing screwdriver reveals the game's beautiful 1976 circuitboard — before the video fast forwards through "an almost comical attempt by me, a systems software engineer, to sauter the connections back onto this 48-year-old connector." (Instead he ends up replacing the machine's 9-volt battery connector...) On his Patreon page, he writes that filming the video "took a stupidly long time to put together." But their Slashdot submission acknowledges that in the end, "Taking it apart and debugging it was fun. (Slight spoiler: I figured out what was wrong, was an easy fix), and the game plays great now!"

Any Slashdot readers have memories of playing Mattel Electronics Auto Race? My one experience felt like that time that a gaming magazine had nine children (ages 9 to 12) try to play old 1970s-era videogames like Pong. ("Wow. The score is tied. It's so exhilarating..." "My line is so beating the heck out of your stupid line...")

Though it breaks down faster than CO2, methane is a greenhouse gas over 80 times as potent as carbon dioxide, reports the Washington Post. It suddenly started increasing in the atmosphere in 2007 — and then in 2020, its growth rate doubled.

While scientists have suspected it was natural gas, some researchers have a new theory... "The changes that we saw in the last couple of years — and even since 2007 — are microbial," said Sylvia Michel, lead author of the paper published last month in the Proceedings of the National Academy of Sciences. "Wetlands, if they are getting warmer and wetter, maybe they're producing more methane than they used to...."

Michel and her co-authors analyzed samples of methane, or CH4, from 22 sites around the globe at a Colorado laboratory. Then they measured the "heaviness" of that methane — specifically, how many of the molecules had a heavier isotope of carbon in them, known as C13. Different sources of methane give off different carbon signatures. Methane produced by microbes — mostly single-celled organisms known as archaea, which live in cow stomachs, wetlands and agricultural fields — tends to be "lighter," or have fewer C13 atoms. Methane from fossil fuels, on the other hand, is heavier, with more C13 atoms. As the amount of methane has risen in the atmosphere over the past 15 years, it's also gotten lighter and lighter. The scientists used a model to analyze those changes and found that only large increases in microbial emissions could explain both the rising methane and its changing weight....

Researchers say it doesn't mean that the world can just keep burning natural gas. If wetlands are releasing methane faster than ever, they argue, there should be an even greater push to curb methane from the sources humans can control, such as cows, agriculture and fossil fuels.
The article includes this quote from Stanford University professor Rob Jackson (who works on the Global Methane Budget, a project tracking the world's methane sources). "You can turn a wrench in an oil and gas field to quench methane emissions," Jackson said. "There's no wrench for the Congo or the Amazon." Another recent study found that two-thirds of current methane emissions are caused by humans — from fossil fuels, rice cultivation, reservoirs and other sources. "Methane forms biologically in warm, wet, low-oxygen environments," Jackson said. "The wetlands of a rice paddy and the gut of the cow are all similar." But evidence is also emerging that natural wetlands may be responding to warming temperatures by pumping out more methane. Satellite data from recent years has shown global methane hot spots in the tropical wetlands of the Amazon and the Congo. "Wetlands will emit more methane as temperatures warm," Jackson said. "This may be the start of a reinforcing feedback, that higher temperatures release more methane from natural ecosystems...."

Over 100 countries have pledged to reduce their methane emissions by 30 percent by 2030, compared with 2020 levels — but so far, that pledge has yet to see results. Instead, satellite measurements show concentrations are rising at a rate that is in line with the worst-case climate scenarios.

From an opinion piece on GamesIndustry.biz about the recently launched PS5 Pro that went on sale this week: What I'd argue is actually more interesting about PS5 Pro in a wider perspective isn't what Sony has done to the chips in the system -- it's what they've chosen not to include, and what it tells us about the decision-making process that's likely occurring for the company's future hardware. PS5 Pro doesn't have a disc drive. Anyone who wants to play disc-based games on the system will need to buy one of the add-on drives Sony started selling when the PS5 Slim model was released, adding further to the cost of the already very expensive device.

To add insult to injury, Sony doesn't seem to have made any effort whatsoever to ensure that those drives are actually well-stocked for the launch of the Pro. I can only speak directly to the situation in Japan, where they've been out of stock at most major retailers for months and even second-hand units are being sold at three to four times SRP by scalpers. But asking around suggests that the situation isn't much better in other regions. That's a very rough welcome to PS5 Pro ownership for anyone upgrading who has a collection of games on disc.

It's possible, of course, that Sony excluded the drive simply because its cost would push the Pro's price tag even higher. However, the incongruity of Sony's "Pro" console lacking the basic ability to play the games Sony sells at retailers all around the world is striking, and it's difficult to see the decision to accept that incongruity -- and the inconvenience it would inevitably cause for customers -- as anything other than strategic.

Digital sales make up a bigger and bigger portion of the industry's revenues every year, but physical game sales are still a very big deal -- and physical games are products that fall outside the control of publishers and platform holders in a way that they have found increasingly irritating in recent years. People who buy physical games can sell them second-hand or lend them to their friends, retailers with physical games in stock can discount them or include them in bundles as they see fit.

Anton Shilov reports via Tom's Hardware: FFmpeg is an open-source video decoding project developed by volunteers who contribute to its codebase, fix bugs, and add new features. The project is led by a small group of core developers and maintainers who oversee its direction and ensure that contributions meet certain standards. They coordinate the project's development and release cycles, merging contributions from other developers. This group of developers tried to implement a handwritten AVX512 assembly code path, something that has rarely been done before, at least not in the video industry.

The developers have created an optimized code path using the AVX-512 instruction set to accelerate specific functions within the FFmpeg multimedia processing library. By leveraging AVX-512, they were able to achieve significant performance improvements -- from three to 94 times faster -- compared to standard implementations. AVX-512 enables processing large chunks of data in parallel using 512-bit registers, which can handle up to 16 single-precision FLOPS or 8 double-precision FLOPS in one operation. This optimization is ideal for compute-heavy tasks in general, but in the case of video and image processing in particular.

The benchmarking results show that the new handwritten AVX-512 code path performs considerably faster than other implementations, including baseline C code and lower SIMD instruction sets like AVX2 and SSSE3. In some cases, the revamped AVX-512 codepath achieves a speedup of nearly 94 times over the baseline, highlighting the efficiency of hand-optimized assembly code for AVX-512.

While Amazon, Google, and other companies build new data centers — sometimes for their AI projects — parts of America "are facing higher electric bills," reports the Washington Post: The facilities' extraordinary demand for electricity to power and cool computers inside can drive up the price local utilities pay for energy and require significant improvements to electric grid transmission systems. As a result, costs have already begun going up for customers — or are about to in the near future, according to utility planning documents and energy industry analysts. Some regulators are concerned that the tech companies aren't paying their fair share, while leaving customers from homeowners to small businesses on the hook. In Oregon, electric utilities are warning regulators that consumers need protections from rising rates caused by data centers. From Virginia to Ohio and South Carolina, companies are battling over the extent of their responsibility for increases, attempting to fend off anger from customers. In the Mid-Atlantic, the regional power grid's energy costs shot up dramatically, and data centers are cited as among root causes of rate increases of up to 20 percent expected in 2025...

The tech firms and several of the power companies serving them strongly deny they are burdening others. They say higher utility bills are paying for overdue improvements to the power grid that benefit all customers. In some cases, they said in response to criticism from consumer and business advocates that they are committed to covering additional costs. But regulators — and even some utilities — are growing skeptical.

A jarring example of fallout on consumers is playing out on the Mid-Atlantic regional power grid, called PJM Interconnection, which serves 13 states and D.C. The recent auction to secure power for the grid during periods of extreme weather and high demand resulted in an 800 percent jump in the price that the grid's member utilities had to pay. The impact will be felt by millions by the spring, according to public records. Power bills will increase as much as 20 percent for customers of a dozen utilities in Maryland, Ohio, Pennsylvania, New Jersey and West Virginia, regulatory filings show. That includes households in the Baltimore area, where annual bills will increase an average of $192, said Maryland People's Counsel David Lapp, a state appointee who monitors utilities. The next auction, in 2025, could be more painful, Lapp said, leaving customers potentially "looking at increases of as much as $40 to $50 a month...."

Advocates cite another source of cost-shifting onto consumers: discounted rates that power companies and local government officials use to entice tech companies to build data centers... Google worked out a deal with Dominion Energy, blessed by regulators, to pay 6 cents per kilowatt hour for its power. That is less than half of what residential customers pay, as well as substantially less than is paid by businesses...
The article points out that in Pennsylvania, "Amazon's novel plan to fuel a data center from a reactor at the nearby Susquehanna nuclear plant is now in jeopardy, after regulators blocked it Friday. They cited potential impact on consumers as among their concerns. The plan threatens to leave other ratepayers stuck with a bill of $50 million to $140 million, according to testimony from [power utility] AEP and utility conglomerate Exelon."

And meanwhile, one Virginia retiree complained about a proposed $54 million transmission line and substation for an Amazon data center. "They are already making money hand over fist, and now they want us to pay for this?

The feds have indicted Aleksei Andriunin, a 26-year-old Russian national and founder of Gotbit, on charges of wire fraud and conspiracy to commit market manipulation. Crypto News reports: According to the U.S. Attorney's Office, the indictment alleges that Andriunin and his firm participated in a long-running scheme to artificially boost trading volumes for various cryptocurrency companies, including some based in the United States, to make them appear more popular and increase their trading value. Andriunin allegedly led these activities between 2018 and 2024 as Gotbit's CEO. He could face up to 20 years in prison, additional fines, and asset forfeiture if convicted, according to the U.S. Attorney's Office. Prosecutors say the scheme involved "wash trading," where the firm used its software to make fake trades that inflated a cryptocurrency's trading volume. This practice, called market manipulation, can mislead investors by giving the impression that demand for a particular cryptocurrency is higher than it actually is. Wash trades are illegal in traditional finance and are considered fraudulent because they deceive investors and manipulate market behavior.

Court documents also identify Gotbit's two directors, Fedor Kedrov and Qawi Jalili, as co-conspirators. The indictment claims Gotbit documented these activities in detailed records, tracking differences between genuine and artificial trading volumes. The firm allegedly pitched these services to prospective clients, explaining how Gotbit's tactics would bypass detection on public blockchains, where transactions are recorded transparently. The U.S. Department of Justice has announced that it seized over $25 million worth of cryptocurrency assets connected to these schemes and made four arrests across multiple firms. If you've been following the crypto industry, you're probably familiar with "pump-and-dump" schemes that have popped up throughout the years. Although it's a form of market manipulation, it's not quite the same as "wash trading."

In a pump-and-dump scheme, the perpetrator artificially inflates the price of a security (often a low-priced or thinly traded stock) by spreading misleading or exaggerated information to attract other buyers, who then drive up the price. Once the price has risen due to increased demand, the manipulators "dump" their shares at the inflated price, selling to the new buyers and pocketing the profits. The price typically crashes after the dump, leaving unsuspecting investors with overvalued shares and significant losses.

Wash trading, on the other hand, involves simultaneously buying and selling of the same asset to create the illusion of higher trading volume and activity. The purpose is to mislead other investors about the asset's liquidity and demand, often giving the impression that it is more popular or actively traded than it actually is. Wash trades usually occur without real changes in ownership or price movement, as the buyer and seller may even be the same person or entity. This tactic can manipulate prices indirectly by creating a perception of interest, but it does not involve a direct inflation followed by a sell-off, like a pump-and-dump scheme.

When it comes to introducing liability for software products, "the EU and U.S. are taking very different approaches," according to Lawfare's cybersecurity newsletter. "While the U.S. kicks the can down the road, the EU is rolling a hand grenade down it to see what happens." Under the status quo, the software industry is extensively protected from liability for defects or issues, and this results in systemic underinvestment in product security. Authorities believe that by making software companies liable for damages when they peddle crapware, those companies will be motivated to improve product security... [T]he EU has chosen to set very stringent standards for product liability, apply them to people rather than companies, and let lawyers sort it all out.

Earlier this month, the EU Council issued a directive updating the EU's product liability law to treat software in the same way as any other product. Under this law, consumers can claim compensation for damages caused by defective products without having to prove the vendor was negligent or irresponsible. In addition to personal injury or property damages, for software products, damages may be awarded for the loss or destruction of data. Rather than define a minimum software development standard, the directive sets what we regard as the highest possible bar. Software makers can avoid liability if they prove a defect was not discoverable given the "objective state of scientific and technical knowledge" at the time the product was put on the market.

Although the directive is severe on software makers, its scope is narrow. It applies only to people (not companies), and damages for professional use are explicitly excluded. There is still scope for collective claims such as class actions, however. The directive isn't law itself but sets the legislative direction for EU member states, and they have two years to implement its provisions. The directive commits the European Commission to publicly collating court judgements based on the directive, so it will be easy to see how cases are proceeding.

Major software vendors used by the world's most important enterprises and governments are publishing comically vulnerable code without fear of any blowback whatsoever. So yes, the status quo needs change. Whether it needs a hand grenade lobbed at it is an open question. We'll have our answer soon.

Libreboot is a distribution of coreboot "aimed at replacing the proprietary BIOS firmware contained by most computers."

So then what exactly is GNU Boot? Its home page explains... In November 2022, Libreboot began to include non-libre code. We have made repeated efforts to continue collaboration with those developers to help their version of Libreboot remain libre, but that was not successful. Now we've stepped forward to stand up for freedom, ours and that of the wider community, by maintaining our own version — a genuinely libre Libreboot, that after some hurdles gave birth to this project: GNU Boot.
But today, Phoronix writes: While priding itself on being "100% free", last December [GNU Boot] had to drop some motherboard support and CPU code after discovering they were shipping some files that are non-free by their free software standards. Today they announced another mistake in having inadvertently been shipping additional non-free code.

GNU Boot discovered an issue with non-free code affecting not only them but also some of the Linux distributions that pride themselves on being fully free software / 100% open-source. This latest snafu they say is "more problematic" than their prior non-free code discover due to impacting the free software Linux distributions too. The issue at hand though comes down to test data contained within the archive and that containing non-free code in the form of microcode, BIOS bits, and Intel Management Engine firmware.
"We also contacted Replicant..." according to the announcement, "a free Android distro that also ships vboot source code." And in addition, "We had to re-release all the affected tarballs." (Which at this point is three release candidates...)

An anonymous reader shares a report: If you dread the thought of calling to change an airline ticket or negotiate your internet bill, a new artificial intelligence tool may provide a solution. DoNotPay, which offers an assortment of consumer-friendly services like tracking subscriptions, generating burner phone numbers, and searching for unclaimed property, now features a bot that will call customer service numbers for users, navigate through phone menus and sit through hold music, then politely but firmly advocate on users' behalf.

The company shared examples of its AI calling a cellphone provider for help porting a phone number and talking with an airline to cancel a flight within the 24-hour cancellation window. Joshua Browder, CEO and founder of DoNotPay, says getting updates on lost luggage and seeking compensation for flight delays are also common use cases. DoNotPay already offered tools to connect to customer service agents via chat windows, and to draft and send emails, faxes, and even snail mail to companies on behalf of users.

But while the service's artificial intelligence had enough smarts to wait on hold for users, then hand over a call when an agent was available, until recently AI models were not capable of carrying on a convincing voice conversation with a human operator in real time. Browder says that changed with Open AI's GPT-4o model, unveiled in May. "That has reduced the delay by about 70%, so instead of it taking three seconds to come up with a response, it now takes under a second, and that's finally fast enough to hold these phone conversations," he says. "So now we're doing thousands of these calls."

An anonymous reader shares a report: Every year for the past seven, Nathan Benaich, the founder and solo general partner at the early-stage AI investment firm Air Street Capital, has produced a magisterial "State of AI" report. Benaich and his collaborators marshal an impressive array of data to provide a great snapshot of the technology's evolving capabilities, the landscape of companies developing it, a survey of how AI is being deployed, and a critical examination of the challenges still facing the field.

One of the big takeaways from this year's report, which was published late last week, is that OpenAI's lead over other AI labs has largely eroded. Anthropic's Claude 3.5 Sonnet, Google's Gemini 1.5, X's Grok 2, and even Meta's open-source Llama 3.1 405 B model have equaled, or narrowly surpassed on some benchmarks, OpenAI's GPT-4o.ââBut, on the other hand, OpenAI still retains an edge for the moment on reasoning tasks with the release of its o1 "Strawberry" model -- which Air Street's report rightly characterized as a weird mix of incredibly strong logical abilities for some tasks, and surprisingly weak ones for others.

Another big takeaway, Benaich told me, is the extent to which the cost of using a trained AI model -- an activity known as "inference" -- is falling rapidly. There are several reasons for this. One is linked to that first big takeaway: With models less differentiated from one another on capabilities and performance, companies are forced to compete on price.ââAnother reason is that engineers for companies such as OpenAI and Anthropic -- and their hyperscaler partners Microsoft and AWS, respectively -- are discovering ways to optimize how the largest models run on big GPU clusters. The cost of outputs from OpenAI's GPT-4o today is 100-times less per token (which is about equivalent to 1.5 words) than it was for GPT-4 when that model debuted in March 2023. Google's Gemini 1.5 Pro now costs 76% less per output token than it did when that model was launched in February 2024.â

New research challenges assumptions about decision-making, revealing people tend to believe they have sufficient information regardless of actual data at hand. A study by Gehlbach, Robinson, and Fletcher, published earlier this month, found participants consistently overestimated their knowledge when given partial information on a hypothetical school merger scenario.

Nearly 90% favored merger when presented pro-merger facts, while only 25% did when given opposing data. However, opinions shifted when full information was provided, suggesting malleability of views despite initial overconfidence. Researchers caution this bias could be exploited in today's fractured media landscape, where partial or misleading information often circulates unchecked.

An anonymous reader quotes a report from TorrentFreak: Korean game publisher Nexon is using the U.S. legal system to address online copyright infringement. The company obtained a DMCA subpoena that requires Discord to hand over the personal details of suspected pirates. While Discord has shared information in the past, it doesn't plan to cooperate any longer, refusing to play the role of 'anti-piracy police'. [...] The messaging platform wrote that it is prepared to file a motion to quash the subpoena, if needed. It further urged Nexon to withdraw their demands, and cease sending any similar 'defective' subpoenas going forward. To support its stance, Discord made a list of twenty-two general objections and reservations. Among other things, the company wants to protect user privacy and their first amendment right to anonymous speech.

"Discord objects to the Requests as infringing its users' decisions to remain anonymous, an aspect of their freedom of speech protected by the First Amendment. The Requests improperly seek to unmask anonymous speakers and consequently compel disclosure of material protected by the First Amendment," it reads. This strongly-worded letter didn't have the desired result, however. Instead of backing off, Nexon doubled down, filing a motion to compel (PDF) at a Texas federal court late last week. The game company refutes Discord's objections and asks the court to enter an order requiring Discord to produce the requested user data. Nexon says that it needs this information to protect its copyrights. "Discord's failure to cooperate discovery has impeded Nexon's ability to discover relevant, non-privileged information that will support its potential claims against the users who have provided access to the infringing material," Nexon writes. While the court has yet to rule on the matter, Discord is expected to file a formal motion to quash the subpoena in response, as indicated in its earlier communications.

An anonymous reader quotes a report from TechCrunch: In late September, Shield AI cofounder Brandon Tseng swore that weapons in the U.S. would never be fully autonomous -- meaning an AI algorithm would make the final decision to kill someone. "Congress doesn't want that," the defense tech founder told TechCrunch. "No one wants that." But Tseng spoke too soon. Five days later, Anduril cofounder Palmer Luckey expressed an openness to autonomous weapons -- or at least a heavy skepticism of arguments against them. The U.S.'s adversaries "use phrases that sound really good in a sound bite: Well, can't you agree that a robot should never be able to decide who lives and dies?" Luckey said during a talk earlier this month at Pepperdine University. "And my point to them is, where's the moral high ground in a landmine that can't tell the difference between a school bus full of kids and a Russian tank?"

When asked for further comment, Shannon Prior, a spokesperson for Anduril said that Luckey didn't mean that robots should be programmed to kill people on their own, just that he was concerned about "bad people using bad AI." In the past, Silicon Valley has erred on the side of caution. Take it from Luckey's cofounder, Trae Stephens. "I think the technologies that we're building are making it possible for humans to make the right decisions about these things," he told Kara Swisher last year. "So that there is an accountable, responsible party in the loop for all decisions that could involve lethality, obviously." The Anduril spokesperson denied any dissonance between Luckey (pictured above) and Stephens' perspectives, and said that Stephens didn't mean that a human should always make the call, but just that someone is accountable.

Last month, Palantir co-founder and Anduril investor Joe Lonsdale also showed a willingness to consider fully autonomous weapons. At an event hosted by the think tank Hudson Institute, Lonsdale expressed frustration that this question is being framed as a yes-or-no at all. He instead presented a hypothetical where China has embraced AI weapons, but the U.S. has to "press the button every time it fires." He encouraged policymakers to embrace a more flexible approach to how much AI is in weapons. "You very quickly realize, well, my assumptions were wrong if I just put a stupid top-down rule, because I'm a staffer who's never played this game before," he said. "I could destroy us in the battle."

When TC asked Lonsdale for further comment, he emphasized that defense tech companies shouldn't be the ones setting the agenda on lethal AI. "The key context to what I was saying is that our companies don't make the policy, and don't want to make the policy: it's the job of elected officials to make the policy," he said. "But they do need to educate themselves on the nuance to do a good job." He also reiterated a willingness to consider more autonomy in weapons. "It's not a binary as you suggest -- 'fully autonomous or not' isn't the correct policy question. There's a sophisticated dial along a few different dimensions for what you might have a soldier do and what you have the weapons system do," he said. "Before policymakers put these rules in place and decide where the dials need to be set in what circumstance, they need to learn the game and learn what the bad guys might be doing, and what's necessary to win with American lives on the line." [...] "For many in Silicon Valley and D.C., the biggest fear is that China or Russia rolls out fully autonomous weapons first, forcing the U.S.'s hand," reports TechCrunch. "At the Hudson Institute event, Lonsdale said that the tech sector needs to take it upon itself to 'teach the Navy, teach the DoD, teach Congress' about the potential of AI to 'hopefully get us ahead of China.' Lonsdale's and Luckey's affiliated companies are working on getting Congress to listen to them. Anduril and Palantir have cumulatively spent over $4 million in lobbying this year, according to OpenSecrets."

TorrentFreak's Andy Maxwell reports: In August, New Zealand's Justice Minister authorized Kim Dotcom's immediate arrest and extradition. Dotcom's response to his followers on X was simple: "I'm not leaving." Another post mid-September -- "we are very close to disaster" -- led to Dotcom disappearing for three weeks. On his return, Dotcom said X had suspended his account, based on an extremely serious allegation. After accusing Elon Musk of failing to help, yesterday Dotcom warned that a Trump loss would see Musk indicted and "fighting for his life." Dotcom has a plan to avoid extradition; chaos like this provides the fuel.

The details of Dotcom's "plan" to stay in New Zealand are yet to be revealed. Given Dotcom's history, exhausting the judiciary with every possible avenue of appeal is pretty much guaranteed, no matter how unlikely the prospects of success. At the same time, it's likely that Dotcom will use social media to preach to the existing choir. He will also try to appeal to those who loathe him, and those who merely hate him, by focusing on a common grievance. "People keep suggesting that I should leave this corrupt US colony like a fugitive on the run. Hell no," he told 1.7 million X followers recently. "Corrupt US colony" and the interchangeable "obedient" variant are clearly derogatory, catering to theories of joint complicity and sniveling weakness. This rhetoric has been visible on Dotcom's social media accounts for some time, but the main theme is Dotcom's belligerent, out-of-the-blue support for Russia's invasion of Ukraine. [...]

Some people believe that Dotcom genuinely supports Russia and, with his quotes regularly appearing on state-run news channels, arguing otherwise is a pretty tough ask. A different assessment starts with the things Dotcom values most -- his family, his wealth, and his freedom -- and applies that to a reputation of doing whatever it takes to protect and maintain those three, non-negotiable aspects of his life. Right now, his best chance is to tilt the chess board via a change at the White House, and then carefully exploit a change in policy. Dotcom's colleagues took a plea deal from the U.S. and New Zealand that Dotcom insists he would never accept; certainly not if Biden was in power. A Donald Trump win, on the other hand, would introduce an administration Dotcom could be seen to negotiate with, on previously unthinkable terms, without losing face. Previous reluctance to admit any wrongdoing could suddenly seem trivial after the prevention of World War 3.

[Since 2022, Dotcom supported narratives more closely aligned with those of the Kremlin, in particular the claim that United States policy is the root cause of the current conflict. The amplification of anti-Ukraine rumors in the United States, strategically links alleged U.S. policy failures to billions of dollars in military aid, all at taxpayers' expense. This toxic mix, Dotcom insists, heralds the collapse of the dollar, the dismantling of the "US Empire," and ultimately a global human catastrophe; World War 3, no holds barred.]

The decline of journalism has been attributed to many factors, from slow adaptation to the internet to the dominance of tech giants in advertising. But a veteran journalist offers a new perspective: the death of the hyperlink could be changing the fundamental nature of the internet, with significant implications for the news industry. Matt Pearce: There is a real bias against hyperlinking that has developed on platforms and apps over the last five years in particular. It's something that's kind of operating hand-in-hand with the rise of algorithmic recommendations. You see this on Elon Musk's version of Twitter, where posts with hyperlinks are degraded. Facebook itself has decided to detach itself from displaying a lot of links. That's why you get so much AI scum on Facebook these days. Instagram itself has always been kind of hostile to linking. TikTok as well...

If you degrade hyperlinks, and you degrade this idea of the internet as something that refers you to other things, you instead have this stationary internet where a generative AI agent will hoover up and summarize all the information that's out there, and place it right in front of you so that you never have to leave the portal... That was a real epiphany to me, because the argument against one form of this legislation was, "My God, you'll destroy this fundamental way of how the internet works." I'm like, dude, these companies are already destroying the fundamental way of how the internet works.

[...] If you look at what technology has done to journalism over the last 10 years, it was journalists who figured out how to make Twitter work for them. It was journalists who figured out how to be really good on Instagram and Tik Tok. I know there's this argument about content creators and versus journalists, but I'm like, we're all in the same ecosystem. If you're performing the functions of a journalist, you're a journalist. Some people are really good on different platforms. But it's hard to imagine a scenario where Google is going to be the party that creates a more humane, intelligent, responsive form of journalism.