You are only seeing posts authors requested be public.
Register and Login to participate in discussions with colleagues.
Ars Technica
Amazon ready to use its own AI chips, reduce its dependence on Nvidia
Amazon is poised to roll out its newest artificial intelligence chips as the Big Tech group seeks returns on its multibillion-dollar semiconductor investments and to reduce its reliance on market leader Nvidia.
Executives at Amazon’s cloud computing division are spending big on custom chips in the hopes of boosting the efficiency inside its dozens of data centers, ultimately bringing down its own costs as well as those of Amazon Web Services’ customers.
The effort is spearheaded by Annapurna Labs, an Austin-based chip start-up that Amazon acquired in early 2015 for $350 million. Annapurna’s latest work is expected to be showcased next month when Amazon announces widespread availability of ‘Trainium 2,’ part of a line of AI chips aimed at training the largest models.
Ars Live: Our first encounter with manipulative AI
In the short-term, the most dangerous thing about AI language models may be their ability to emotionally manipulate humans if not carefully conditioned. The world saw its first taste of that danger in February 2023 with the launch of Bing Chat, now called Microsoft Copilot.
During its early testing period, the temperamental chatbot gave the world a preview of an "unhinged" version of OpenAI's GPT-4 prior to its official release. Sydney's sometimes uncensored and "emotional" nature (including use of emojis) arguably gave the world its first large-scale encounter with a truly manipulative AI system. The launch set off alarm bells in the AI alignment community and served as fuel for prominent warning letters about AI dangers.
On November 19 at 4 pm Eastern (1 pm Pacific), Ars Technica Senior AI Reporter Benj Edwards will host a livestream conversation on YouTube with independent AI researcher Simon Willison that will explore the impact and fallout of the 2023 fiasco. We're calling it "Bing Chat: Our First Encounter with Manipulative AI."
Mazda gives the 2025 CX-50 a big efficiency boost to 38 mpg with hybrid power
Mazda won itself a lot of fans over the years with the sharp styling and attention to detail of its engaging-to-drive vehicles. But in recent years its cars have been left behind by rivals when it comes to fuel efficiency. As a small automaker facing daunting development costs, it decided to partner with Toyota on advanced technologies, including a battery electric vehicle due by 2027. A more immediate benefit has been access to Toyota's hybrid powertrains, one of which can now be found inside the $33,970 2025 Mazda CX-50 Hybrid, which we've driven ahead of it going on sale this month.
Engineering stuffOpen the hood, and instead of one of Mazda's Skyactiv engines, you find one of Toyota's 2.5 L naturally aspirated Atkinson cycle engines, which drives the front wheels together with two electric motors (one is really the integrated starter motor) via a continuously variable transmission. At the rear, a third electric motor powers the rear wheels when necessary. The front traction motor provides 118 hp (88 kW), the rear motor can add an extra 54 hp (40 kW), and total output is rated at 219 hp (163 kW) and 163 ft-lb (220 Nm).
It's a handsomely styled vehicle. Credit: Jonathan Gitlin Starting at under $34,000, the base CX-50 is very well-equipped. Credit: Jonathan GitlinIt wasn't exactly a case of shoehorning the new powertrain into the CX-50, but Mazda has had to modify the front frame rails to fit the new engine, transmission, and hybrid motors, and there's a new rear subframe with the rear eAxle, as well as the 0.9 kWh hybrid traction battery, which lives under the rear seat. (This improves the front-rear weight distribution to 55:45, Mazda says.)
Smaller than an Escalade IQ, bigger than a Lyriq: The 2026 Cadillac Vistiq
Cadillac has another new electric vehicle on the way. It's the 2026 Vistiq, a three-row SUV that fills the gap in the American automaker's lineup. It will be the fifth Cadillac to use what until now has been called the Ultium battery platform.
Although parent company General Motors has caught a case of EV regret this year, it has also ironed out the problem it encountered when assembling battery cells into packs and expects to sell at least 200,000 EVs this year. Cadillac's Ultium journey began with the midsize Lyriq, which has since been joined by the compact Optiq, the hand-built, highly exclusive Celestiq, and imminently, the Escalade IQ, which rivals the Hummer EV in weight as a result of its whopping 450 mile range.
The Vistiq conforms to the design language started by the Lyriq. Credit: Cadillac If you're picturing an electric XT6 you've got the right idea. Credit: CadillacThe $77,395 Vistiq drops in between the Lyriq and Escalade IQ, filling the same niche—but electric—as the XT6 does in the gas-powered lineup. The Vistiq actually has a virtually identical wheelbase to the gas Escalade at 121.8 inches (3,094 mm), but it's a few inches shorter at 205 inches (5,207 mm) long. Yes, this is a large SUV, but it's a three-row Cadillac—what else do you expect?
There are some things the Crew-8 astronauts aren’t ready to talk about
The astronauts who came home from the International Space Station last month experienced some drama on the high frontier, and some of it accompanied them back to Earth.
In orbit, the astronauts aborted two spacewalks, both under unusual circumstances. Then, on October 25, one of the astronauts was hospitalized due to what NASA called an unspecified "medical issue" after splashdown aboard a SpaceX Crew Dragon capsule that concluded the 235-day mission. After an overnight stay in a hospital in Florida, NASA said the astronaut was released "in good health" and returned to their home base in Houston to resume normal post-flight activities.
The space agency did not identify the astronaut or any details about their condition, citing medical privacy concerns. The three NASA astronauts on the Dragon spacecraft included commander Matthew Dominick, pilot Michael Barratt, and mission specialist Jeanette Epps. Russian cosmonaut Alexander Grebenkin accompanied the three NASA crew members. Russia's space agency confirmed he was not hospitalized after returning to Earth.
FTX sues Binance for $1.76B in battle of crypto exchanges founded by convicts
The bankruptcy estate of collapsed cryptocurrency exchange FTX has sued the company's former rival Binance in an attempt to recover $1.76 billion or more. The lawsuit seeks "at least $1.76 billion that was fraudulently transferred to Binance and its executives at the FTX creditors' expense, as well as compensatory and punitive damages to be determined at trial."
The complaint filed yesterday in US Bankruptcy Court in Delaware names Binance and co-founder and former CEO Changpeng Zhao among the defendants. FTX founder Sam Bankman-Fried sold 20 percent of his crypto exchange to Binance in November 2019, but Binance exited that investment in 2021, the lawsuit said.
"As Zhao would later remark, he decided to exit his position in FTX because of personal grievances he had against Bankman-Fried," the lawsuit said. "In July 2021, the parties negotiated a deal whereby FTX bought back Binance's and its executives' entire stakes in both FTX Trading and [parent company] WRS. Pursuant to that deal, FTX's Alameda Research division directly funded the share repurchase with a combination of FTT (FTX's exchange token), BNB (Binance's exchange token), and BUSD (Binance's dollar-pegged stablecoin). In the aggregate, those tokens had a fair market value of at least $1.76 billion."
Is “AI welfare” the new frontier in ethics?
A few months ago, Anthropic quietly hired its first dedicated "AI welfare" researcher, Kyle Fish, to explore whether future AI models might deserve moral consideration and protection, reports AI newsletter Transformer. While sentience in AI models is an extremely controversial and contentious topic, the hire could signal a shift toward AI companies examining ethical questions about the consciousness and rights of AI systems.
Fish joined Anthropic's alignment science team in September to develop guidelines for how Anthropic and other companies should approach the issue. The news follows a major report co-authored by Fish before he landed his Anthropic role. Titled "Taking AI Welfare Seriously," the paper warns that AI models could soon develop consciousness or agency—traits that some might consider requirements for moral consideration. But the authors do not say that AI consciousness is a guaranteed future development.
"To be clear, our argument in this report is not that AI systems definitely are—or will be—conscious, robustly agentic, or otherwise morally significant," the paper reads. "Instead, our argument is that there is substantial uncertainty about these possibilities, and so we need to improve our understanding of AI welfare and our ability to make wise decisions about this issue. Otherwise there is a significant risk that we will mishandle decisions about AI welfare, mistakenly harming AI systems that matter morally and/or mistakenly caring for AI systems that do not."
Man gets 10 years for stealing $20M in nest eggs from 400 US home buyers
A Nigerian man living in the United Kingdom has been sentenced to 10 years for his role in a phishing scam that snatched more than $20 million from over 400 would-be home buyers in the US, including some savers who lost their entire nest eggs.
Late last week, the US Department of Justice confirmed that 33-year-old Babatunde Francis Ayeni pled guilty to conspiracy to commit wire fraud through "a sophisticated business email compromise scheme targeting real estate transactions" in the US.
To seize large down payments on homes, Ayeni and co-conspirators sent phishing emails to US title companies, real estate agents, and real estate attorneys. When unsuspecting employees clicked malicious attachments and links, a prompt appeared asking for login information that was then shared with the hackers.
Air quality problems spur $200 million in funds to cut pollution at ports
Raquel Garcia has been fighting for years to clean up the air in her neighborhood southwest of downtown Detroit.
Living a little over a mile from the Ambassador Bridge, which thousands of freight trucks cross every day en route to the Port of Detroit, Garcia said she and her neighbors are frequently cleaning soot off their homes.
“You can literally write your name in it,” she said. “My house is completely covered.”
Russia: Fine, I guess we should have a Grasshopper rocket project, too
Like a lot of competitors in the global launch industry, Russia for a long time dismissed the prospects of a reusable first stage for a rocket.
As late as 2016, an official with the Russian agency that develops strategy for the country's main space corporation, Roscosmos, concluded, "The economic feasibility of reusable launch systems is not obvious." In the dismissal of the landing prospects of SpaceX's Falcon 9 rocket, Russian officials were not alone. Throughout the 2010s, competitors including space agencies in Europe and Japan, and US-based United Launch Alliance, all decided to develop expendable rockets.
However, by 2017, when SpaceX re-flew a Falcon 9 rocket for the first time, the writing was on the wall. "This is a very important step, we sincerely congratulate our colleague on this achievement," then-Roscosmos CEO Igor Komarov said at the time. He even spoke of developing reusable components, such as rocket engines capable of multiple firings.
How a stubborn computer scientist accidentally launched the deep learning boom
During my first semester as a computer science graduate student at Princeton, I took COS 402: Artificial Intelligence. Toward the end of the semester, there was a lecture about neural networks. This was in the fall of 2008, and I got the distinct impression—both from that lecture and the textbook—that neural networks had become a backwater.
Neural networks had delivered some impressive results in the late 1980s and early 1990s. But then progress stalled. By 2008, many researchers had moved on to mathematically elegant approaches such as support vector machines.
I didn’t know it at the time, but a team at Princeton—in the same computer science building where I was attending lectures—was working on a project that would upend the conventional wisdom and demonstrate the power of neural networks. That team, led by Prof. Fei-Fei Li, wasn’t working on a better version of neural networks. They were hardly thinking about neural networks at all.
Marvel drops Captain America: Brave New World trailer
Marvel Studios dropped a full-length trailer for Captain America: Brave New World at the first ever Brazil D23 fan event this weekend. This is star Anthony Mackie's first cinematic appearance as the new Captain America after the Phase Four 2021 TV miniseries, The Falcon and the Winter Soldier. The event also featured a special look at Marvel's forthcoming Thunderbolts* film, followed by a new trailer.
As previously reported, it's the fifth film in the MCU's Phase Five, directed by Julius Onah (The Cloverfield Paradox) and building on events not just in F&WS but also the 2008 film The Incredible Hulk. Per the official premise:
After meeting with newly elected US President Thaddeus Ross, played by Harrison Ford in his Marvel Cinematic Universe debut, Sam finds himself in the middle of an international incident. He must discover the reason behind a nefarious global plot before the true mastermind has the entire world seeing red.
In addition to Mackie and Ford, the cast includes Liv Tyler as the president's daughter, Betty Ross, and Tim Blake Nelson as Samuel Sterns, both reprising their roles in 2008's The Incredible Hulk. (Ford replaces the late William Hurt, who played Ross in that earlier film.) Carl Lumbley plays Isaiah Bradley, reprising his F&WS role as a Korean War veteran who had been secretly imprisoned and given the Super Soldier Serum against his will, enduring 30 years of experimentation. (He told Sam he couldn't imagine how any black man could take up Captain America's shield because of what it represented to people like him, and one could hardly blame him.)
Review: Catching up with the witchy brew of Agatha All Along
The MCU's foray into streaming television has produced mixed results, but one of my favorites was the weirdly inventive, oh-so-meta WandaVision. I'm happy to report that the spinoff sequel, Agatha All Along, taps into that same offbeat creativity, giving us a welcome reminder of just how good the MCU can be when it's firing on all storytelling cylinders.
(Spoilers below, including for WandaVision and Multiverse of Madness. We'll give you another heads up when major spoilers for Agatha All Along are imminent.)
The true identity of nosy next-door neighbor Agnes—played to perfection by Kathryn Hahn—was the big reveal of 2021's WandaVision, even inspiring a jingle that went viral. Agnes turned out to be a powerful witch named Agatha Harkness, who had studied magic for centuries and was just dying to learn the source of Wanda's incredible power. Wanda's natural abilities were magnified by the Mind Stone, but Agatha realized that Wanda was a wielder of "chaos magic." She was, in fact, the Scarlet Witch. In the finale, Wanda trapped Agatha in her nosy neighbor persona while releasing the rest of the town of Westview from her grief-driven Hex.
Research monkeys still having a ball days after busting out of lab, police say
If you need any inspiration for cutting loose and relaxing this weekend, look no further than a free-wheeling troop of monkeys that broke out of their South Carolina research facility Wednesday and, as of noon Friday, were still "playfully exploring" with their newfound freedom.
In an update Friday, the police department of Yemassee, SC said that the 43 young, female rhesus macaque monkeys are still staying around the perimeter of the Alpha Genesis Primate Research Facility. "The primates are exhibiting calm and playful behavior, which is a positive indication," the department noted.
The fun-loving furballs got free after a caretaker "failed to secure doors" at the facility.
Claude AI to process secret government data through new Palantir deal
Anthropic has announced a partnership with Palantir and Amazon Web Services to bring its Claude AI models to unspecified US intelligence and defense agencies. Claude, a family of AI language models similar to those that power ChatGPT, will work within Palantir's platform using AWS hosting to process and analyze data. But some critics have called out the deal as contradictory to Anthropic's widely-publicized "AI safety" aims.
On X, former Google co-head of AI ethics Timnit Gebru wrote of Anthropic's new deal with Palantir, "Look at how they care so much about 'existential risks to humanity.'"
The partnership makes Claude available within Palantir's Impact Level 6 environment (IL6), a defense-accredited system that handles data critical to national security up to the "secret" classification level. This move follows a broader trend of AI companies seeking defense contracts, with Meta offering its Llama models to defense partners and OpenAI pursuing closer ties with the Defense Department.
New SMB-friendly subscription tier may be too late to stop VMware migrations
Broadcom has a new subscription tier for VMware virtualization software that may appease some disgruntled VMware customers, especially small to medium-sized businesses. The new VMware vSphere Enterprise Plus subscription tier creates a more digestible bundle that's more appropriate for smaller customers. But it may be too late to convince some SMBs not to abandon VMware.
Soon after Broadcom bought VMware, it stopped the sale of VMware perpetual licenses and started requiring subscriptions. Broadcom also bundled VMware's products into a smaller number of SKUs, resulting in higher costs and frustration for customers that felt like they were being forced to pay for products that they didn't want. All that, combined with Broadcom ditching some smaller VMware channel partners (and reportedly taking the biggest clients direct), have raised doubts that Broadcom's VMware would be a good fit for smaller customers.
“The challenge with much of the VMware by Broadcom changes to date and before the announcement [of the vSphere Enterprise Plus subscription tier] is that it also forced many organizations to a much higher offering and much more components to a stack that they were previously uninterested in deploying," Rick Vanover, Veeam's product strategy VP, told Ars.
Matter 1.4 has some solid ideas for the future home—now let’s see the support
Matter, the smart home standard that promises an interoperable future for home automation, even if it's scattered and a bit buggy right now, is out with a new version, 1.4. It promises more device types, improvements for working across ecosystems, and tools for managing battery backups, solar panels, and heat pumps.
"Enhanced Multi-Admin" is the headline feature for anybody invested in Matter's original promise, one where you can buy a device and it doesn't matter if your other gear is meant for Amazon (Alexa), Google, Apple, or whatever, it should just connect and work. With 1.4, a home administrator should be able to let a device onto their network just once, and then have that device picked up by whatever controller they're using. There have technically been ways for a device to be set up on, say, Alexa and Apple Home, but the process has been buggy, involves generating "secondary codes," and is kind of an unpaid junior sysadmin job.
What's now available is "Fabric Sync," which sounds like something that happens in a static-ridden dryer. But "Fabrics" is how the Connectivity Standards Alliance (CSA) describes smart home systems, like Alexa or Google Home. In theory, with every tech company doing their best, you'd set up a smart light bulb with your iPhone, add it to your Apple Home, but still have it be able to be added to a Google Home system, Android phones included. Even better, ecosystems that don't offer controls for entire categories, like Apple and smart displays (because it doesn't make any), should still be able to pick up and control them.
Verizon, AT&T tell courts: FCC can’t punish us for selling user location data
Verizon, AT&T, and T-Mobile are continuing their fight against fines for selling user location data, with two of the big three carriers submitting new court briefs arguing that the Federal Communications Commission can't punish them.
A Verizon brief filed on November 4 and an AT&T brief on November 1 contest the legal basis for the FCC fines issued in April 2024. T-Mobile also sued the FCC, but briefs haven't been filed yet in that case.
"Verizon's petition for review stems from the multiple and significant errors that the FCC, in purporting to enforce statutory consumer data privacy provisions, made in overstepping its authority," Verizon wrote. "The FCC's Forfeiture Order violated both the Communications Act and the Constitution, while failing to benefit the consumers it purported to protect."
Discord terrorist known as “Rabid” gets 30 years for preying on kids
A Michigan man who ran chat rooms and Discord servers targeting children playing online games and coercing them into self-harm, sexually explicit acts, suicide, and other violence was sentenced to 30 years in prison Thursday.
According to the US Department of Justice, Richard Densmore was a member of an online terrorist network called 764, which the FBI considers a "tier one" terrorist threat. He pled guilty to sexual exploitation of a child as "part of a broader indictment that charged him with other child exploitation offenses." In the DOJ's press release, FBI Director Christopher Wray committed to bring to justice any abusive groups known to be preying on vulnerable kids online.
“This defendant orchestrated a community to target children through online gaming sites and used extortion and blackmail to force his minor victims to record themselves committing acts of self-harm and violence,” Wray said. “If you prey on children online, you can’t hide behind a keyboard. The FBI will use all our resources and authorities to arrest you and hold you accountable.”
Space policy is about to get pretty wild, y’all
The global space community awoke to a new reality on Wednesday morning.
The founder of this century's most innovative space company, Elon Musk, successfully used his fortune, time, and energy to help elect Donald Trump to president of the United States. Already, Musk was the dominant Western player in space. SpaceX launches national security satellites and NASA astronauts and operates a megaconstellation. He controls the machines that provide essential space services to NASA and the US military. And now, thanks to his gamble on backing Trump, Musk has strong-armed himself into Trump's inner circle.
Although he may not have a cabinet-appointed position, Musk will have a broad portfolio in the new administration for as long as his relations with Trump remain positive. This gives Musk extraordinary power over a number of areas, including spaceflight. Already this week, he has been soliciting ideas and input from colleagues. The New York Times reported that Musk has advised Trump to hire key employees from SpaceX into his administration, including at the Department of Defense. This reflects the huge conflict of interest that Musk will face when it comes to space policy. His actions could significantly benefit SpaceX, of which he is the majority owner and has the final say in major decisions.