By Billy Perrigo and Tharin PillayJanuary 27, 2025 1:55 PM ESTA new Chinese AI model, created by the Hangzhou-based startup DeepSeek, has stunned the American AI industry by outperforming some of OpenAIs leading models, displacing ChatGPT at the top of the iOS app store, and usurping Meta as the leading purveyor of so-called open source AI tools. All of which has raised a critical question: despite American sanctions on Beijings ability to access advanced semiconductors, is China catching up with the U.S. in the global AI race? At a supposed cost of just $6 million to train, DeepSeeks new R1 model, released last week, was able to match the performance on several math and reasoning metrics by OpenAIs o1 model the outcome of tens of billions of dollars in investment by OpenAI and its patron Microsoft. The Chinese model is also cheaper for users. Access to its most powerful versions costs some 95% less than OpenAI and its competitors. The upshot: the U.S. tech industry is suddenly faced with a potentially cheaper and more powerful challenger, unnerving investors, who sold off American tech stocks on Monday morning. Yet not everyone is convinced. Some American AI researchers have cast doubt on DeepSeeks claims about how much it spent, and how many advanced chips it deployed to create its model.Few, however, dispute DeepSeeks stunning capabilities. Deepseek R1 is AI's Sputnik moment, wrote prominent American venture capitalist Marc Andreessen on X, referring to the moment in the Cold War when the Soviet Union managed to put a satellite in orbit ahead of the United States.So, what is DeepSeek and what could it mean for U.S. tech supremacy? What is DeepSeek?DeepSeek was founded less than two years ago by the Chinese hedge fund High Flyer as a research lab dedicated to pursuing Artificial General Intelligence, or AGI. A spate of open source releases in late 2024 put the startup on the map, including the large language model v3, which outperformed all of Meta's open-source LLMs and rivaled OpenAI's closed-source GPT4-o.At the time, Liang Wenfeng, the CEO, reportedly said that he had hired young computer science researchers with a pitch to solve the hardest questions in the world"critically, without aiming for profits. Early signs were promising: his products were so efficient that DeepSeeks 2024 releases sparked a price war within the Chinese AI industry, forcing competitors to slash prices. This year, that price war looks set to reach across the Pacific Ocean.Yet DeepSeeks AI looks different from its U.S. competitors in one important way.What DeepSeeks success could mean for American tech giantsAt a moment when Google, Meta, Microsoft, Amazon and dozens of their competitors are preparing to spend further tens of billions of dollars on new AI infrastructure, DeepSeeks success has raised a troubling question: Could Chinese tech firms potentially match, or even surpass, their technical prowess while spending significantly less?Meta, which plans to spend $65 billion on AI infrastructure this year, has already set up four war rooms to analyze DeepSeeks models, seeking to find out how the Chinese firm had managed to train a model so cheaply and use the insights to improve its own open source Llama models, tech news site The Information reported over the weekend.In the financial markets, Nvidias stock price dipped more than 15% on Monday morning on fears that fewer AI chips may be necessary to train powerful AI than previously thought. Other American tech stocks were also trading lower.While [DeepSeek R1] is good news for users and the global economy, it is bad news for U.S. tech stocks, says Luca Paolini, chief strategist at Pictet Asset Management. It may result in a nominal downsizing of capital investment in AI and pressure on margins, at a time when valuation and growth expectations are very stretched.But American tech hasnt lostat least not yet.For now, OpenAIs o1 Pro model is still considered the most advanced in the world. The performance of DeepSeek R1, however, does suggest that China is much closer to the frontier of AI than previously thought, and that open-source models have just about caught up to their closed-source counterparts.Perhaps even more worrying for companies like OpenAI and Google, whose models are closed source, is how muchor rather, how littleDeepSeek is charging consumers to access its most advanced models. OpenAI charges $60 per million tokens, or segments of words, outputted by its most advanced model, o1. By contrast DeepSeek charges $2.19 for the same number of tokens from R1nearly 30 times less.It erodes the industrial base, it erodes the margin, it erodes the incentive for further capital investment into western [AI] scaling from private sources, says Edouard Harris, the chief technology officer of Gladstone AI, an AI firm that works closely with the U.S. government. but is Deepseek being transparent?DeepSeeks success was all the more explosive because it seemed to call into question the effectiveness of the U.S. governments strategy to constrain Chinas AI ecosystem by restricting the export of powerful chips, or GPUs, to Beijing. If DeepSeeks claims are accurate, it means China has the ability to create powerful AI models despite those restrictions, underlining the limits of the U.S. strategy. DeepSeek has claimed it is constrained by access to chips, not cash or talent, saying it trained its models v3 and R1 using just 2,000 second-tier Nvidia chips. Money has never been the problem for us, DeepSeeks CEO, Liang Wenfeng, said in 2024. Bans on shipments of advanced chips are the problem. (Current U.S. policy makes it illegal to export to China the most advanced types of AI chips, the likes of which populate U.S. datacenters used by OpenAI and Microsoft.)But are those claims true? My understanding is DeepSeek has 50,000 H100s, Scale AI CEO Alexandr Wang recently told CNBC in Davos, referring to the highest-powered Nvidia GPU chips currently on the market. They cant talk about [them], because it is against the export controls that the U.S. has put in place. (An H100 cluster of that size would cost in the region of billions of dollars.)In a sign of how seriously the CCP is taking the technology, Liang, Deepseeks CEO, met with Chinas premier Li Qiang in Beijing last Monday. In that meeting, Liang reportedly told Li that DeepSeek needs more chips. DeepSeek only has access to a few thousand GPUs, and yet theyre pulling this off, says Jeremie Harris, CEO of Gladstone AI. So this raises the obvious question: what happens when they get an allocation from the Chinese Communist Party to proceed at full speed? Even though China might have achieved a startling level of AI capability with fewer chips, experts say more computing power will always remain a strategic advantage. On that front, the U.S. remains far ahead. It's never a bad thing to have more of it, says Dean Ball, a research fellow at George Mason University. No matter how much you have of it, you will always use it.Where does this leave Americas tech rivalry with China?The short answer: from Washingtons perspective, in uncertain waters.In the closing days of the Biden Administration, outgoing National Security Adviser Jake Sullivan warned that the speed of AI advancement was the most consequential thing happening in the world right now. And just days into his new job, President Trump announced a new $500 billion venture, backed by OpenAI and others, to build the infrastructure vital for the creation of artificial general intelligence the next leap forward in AI, with systems advanced enough to make new scientific breakthroughs and reason in ways that have so far remained in the realm of science fiction.And although questions remain about the future of U.S. chip restrictions on China, Washingtons priorities were apparent in President Trumps AI executive order, also signed during his first week in office, which declared that it is the policy of the United States to sustain and enhance Americas global AI dominance in order to promote human flourishing, economic competitiveness, and national security. Maintaining this dominance will mean, at least in part, understanding exactly what Chinese tech firms are doingas well as protecting U.S. intellectual property, experts say.There's a good chance that DeepSeek and many of the other big Chinese companies are being supported by the [Chinese] government, in more than just a monetary way, says Edouard Harris of Gladstone AI, who also recommended that U.S. AI companies harden their security measures.Where does AI go from here?Since December, OpenAIs new o1 and o3 models have smashed records on advanced reasoning tests designed to be difficult for AI models to pass.DeepSeek R1 does something similar, and in the process exemplifies what many researchers say is a paradigm shift: instead of scaling the amount of computing power used to train the model, researchers scale the amount of time (and thus, computing power and electricity) the model uses to think about a response to a query before answering. It is this scaling of what researchers call test-time compute that distinguishes the new class of reasoning models, such as DeepSeek R1 and OpenAIs o1, from their less sophisticated predecessors. Many AI researchers believe theres plenty of headroom left before this paradigm hits its limit.Some AI researchers hailed DeepSeeks R1 as a breakthrough on the same level as DeepMinds AlphaZero, a 2017 model that became superhuman at the board games Chess and Go by purely playing against itself and improving, rather than observing any human games.Thats because R1 wasnt pretrained on human-labeled data in the same way as other leading LLMs.Instead, DeepSeeks researchers found a way to allow the model to bootstrap its own reasoning capabilities essentially from scratch.Rather than explicitly teaching the model on how to solve a problem, we simply provide it with the right incentives, and it autonomously develops advanced problem-solving strategies, they claim.The finding is significant because it suggests that powerful AI capabilities might emerge more rapidly and with less human effort than previously thought, with just the application of more computing power. DeepSeek R1 is like GPT-1 of this scaling paradigm, says Ball.Ultimately, Chinas recent AI progress, instead of usurping U.S. strength, might in fact be the beginning of a reorderinga step, in other words, toward a future where, instead of a hegemonic power, there are many competing centers of AI power.China will still have their own superintelligence(s) no more than a year later than the US, absent [for example] a war, wrote Miles Brundage, a former OpenAI policy staffer, on X. So unless you want (literal) war, you need to have a vision for navigating multipolar AI outcomes.

The big picture: Element Six has introduced a copper-plated diamond composite material designed to address increasing thermal management challenges in advanced semiconductor devices. By lowering temperatures, some chips could be fed more voltage and be pushed to run even faster. Synthetic diamond specialist Element Six, part of the De Beers Group (think diamond jewelry), said more than half of all electronic device failures today are heat-related. And considering that data centers are expected to account for 10 percent of the total US power demand by 2029, thermal management is becoming an ever-pressing matter.This is where E6's Cu-diamond composite comes into play. According to the company, its thermal conductivity is in the 800 W/mk range about twice as high as copper, but not quite as high as diamond on its own. Diamond, as you may know, is the leading thermally conductive material with conductivity values that are about five times that of copper.The Cu-diamond material can be manufactured into complex shapes with specific features, like holes or curves, to fit a variety of applications (think heatspreaders and other inserts designed to fit between a chip and its cooling solution). E6 noted that thermal conductivity and coefficient of thermal expansion can be fine-tuned by varying the copper-diamond composition to meet a client's specific needs.According to the company, it's also cost effective. It's hard to imagine anything with diamonds in it synthetic or natural being affordable, but we'll reserve judgement until pricing is revealed.Related reading: CPU makers are experimenting with alternative substrates to double clock speedsDon't expect to see this tech show up in consumer-grade hardware, at least not anytime soon. E6 expects the first applications to utilize the material to include high-end chips for AI processing, RF power amplifiers for wireless and defense communications, and powerful semiconductor lasers.Element Six will showcase its new material at the SPIE Photonics West conference for optics and photonics. The show starts on January 28 and runs through the 30th in San Francisco.

Cutting corners: Microsoft introduced Windows Server Update Services in 2005 as an improvement over previous software update mechanisms, designed for Windows patch downloads and servicing. Two decades later, one of the most popular tools in the Windows enterprise ecosystem is losing features ahead of its planned, definitive demise. Microsoft already told its enterprise customers that Windows Server Update Services (WSUS) will eventually go the way of the dodo. While the service will continue to function for now, it will no longer receive development updates as the company shifts its focus to streamlined, cloud-based Windows management platforms.Recently, Microsoft reiterated that WSUS will soon lose some of its existing functionality. According to a new deprecation announcement, starting April 18, 2025, WSUS will stop synchronizing updates for Windows drivers. In on-premises environments such as those used by traditional end customers, small office/home office (SOHO) setups, and small businesses drivers will instead need to be downloaded directly from Microsoft's official Update Catalog.New dark wallpaper in Windows Server 2025However, Microsoft warns that these downloads cannot be directly imported into WSUS. Alternative methods will be required to keep drivers up-to-date outside the platform. The company strongly encourages its customers to transition to cloud-based solutions such as Microsoft Intune and Windows Autopatch. For resourceful administrators, there remains the option to integrate custom driver packages into their Windows images prior to deployment.In its original announcement regarding WSUS deprecation, Microsoft noted that only a small portion of WSUS users (34%) were still using driver synchronization. Among these, most had already implemented alternative solutions, while only 8% expressed concern about the potential impacts of deprecation. // Related StoriesMicrosoft clarifies that "deprecation" refers to the stage in a product's lifecycle when a feature is no longer actively developed. Deprecated features may be completely removed in future releases, though they should continue to work until they are no more.Judging from on Microsoft's latest announcement on the Windows Message Center, the deprecation of WSUS driver synchronization means the feature will no longer be available starting this April. This decision is part of an ongoing trend where Microsoft removes long-standing features from its software and cloud services.The extensive list of deprecated features in Windows 11 is just the tip of the iceberg. As the WSUS case clearly shows, no single customer is really safe from Microsoft's chopping block.

Dennis Quaid transforms into a notorious serial killer in the teaser trailer for Happy Face, an upcoming crime drama at Paramount+.Happy Face is inspired by the true-life story of Melissa G. Moore, played by Annaleigh Ashford. As a teenager, Moore discovered her father (Quaid) was the Happy Face serial killer. With her father behind bars, Moore changed her name to disassociate herself from her familys name that lived in infamy. Decades later, Melissa is forced to confront the incarcerated killer after an innocent man faces the death penalty for a crime her father committed.Recommended VideosEvery violent act is a rock dropped in water, Ashford says in the trailer. If you stay silent, the ripples of trauma just keep pulling everyone they touch underneath.Please enable Javascript to view this contentMoores story served as the basis of theHappy Facepodcast and her autobiography, Shattered SilenceHappy Face | Teaser Trailer | Paramount+Happy Facesensemble includes series regulars James Wolk, Tamera Tomakili, Khiyla Aynne, and Benjamin Mackey.Jennifer Cacicio (Your Honor) is the showrunner and executive producer alongside Robert and Michelle King, Liz Glotzer, Melissa G. Moore, Conal Byrne, Will Pearson, Michael Showalter, and Jordana Mollick. The Kings are the husband-and-wife creators behind several shows, includingThe Good Wife,The Good Fight, Evil, andElsbeth. Showalter directed the first episode.CBS Studios producesHappy Face in association with King Size Productions, iHeartPodcasts, and Semi-Formal Productions.Happy Facepremieres on Thursday, March 20, 2025, with two episodes on Paramount+. New episodes will be released on Thursdays, with the season finale on May 1. Season 1 will consist of eight episodes.Editors Recommendations

Table of ContentsTable of ContentsWhat is DeepSeek?What can DeepSeek do?Who can use DeepSeek?Why is DeepSeek suddenly such a big deal?A year-old startup out of China is taking the AI industry by storm after releasing a chatbot which rivals the performance of ChatGPT while using a fraction of the power, cooling, and training expense of what OpenAI, Google, and Anthropics systems demand. Heres everything you need to know about Deepseeks V3 and R1 models and why the company could fundamentally upend Americas AI ambitions.DeepSeek (technically, Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd.) is a Chinese AI startup that was originally founded as an AI lab for its parent company, High-Flyer, in April, 2023. That May, DeepSeek was spun off into its own company (with High-Flyer remaining on as an investor) and also released its DeepSeek-V2 model. V2 offered performance on par with other leading Chinese AI firms, such as ByteDance, Tencent, and Baidu, but at a much lower operating cost.Recommended VideosThe company followed up with the release of V3 in December 2024. V3 is a 671 billion-parameter model that reportedly took less than 2 months to train. Whats more, according to a recent analysis from Jeffries, DeepSeeks training cost of only US$5.6m (assuming $2/H800 hour rental cost). That is less than 10% of the cost of Metas Llama. Thats a tiny fraction of the hundreds of millions to billions of dollars that US firms like Google, Microsoft, xAI, and OpenAI have spent training their models.Related Introducing DeepSeek-V3!Biggest leap forward yet: 60 tokens/second (3x faster than V2!) Enhanced capabilities API compatibility intact Fully open-source models & papers 1/n pic.twitter.com/p1dV9gJ2Sd DeepSeek (@deepseek_ai) December 26, 2024Benchmark tests put V3s performance on par with GPT-4o and Claude 3.5 Sonnet. A December 2024 Op-Ed in The Hill categorized DeepSeeks success as Americas Sputnik Moment.DeepSeek released its R1-Lite-Preview model in November 2024, claiming that the new model could outperform OpenAIs o1 family of reasoning models (and do so at a fraction of the price). The company estimates that the R1 model is between 20 and 50 times less expensive to run, depending on the task, than OpenAIs o1. DeepSeek subsequently released DeepSeek-R1 and DeepSeek-R1-Zero in January 2025. The R1 model, unlike its o1 rival, is open source, which means that any developer can use it.As such V3 and R1 have exploded in popularity since their release, with DeepSeeks V3-powered AI Assistant displacing ChatGPT at the top of the app stores. Venture capitalist Marc Andreesen, in a recent social media post, called DeepSeeks chatbotone of the most amazing and impressive breakthroughs Ive ever seen and a profound gift to the world.As an open-source large language model, DeepSeeks chatbots can do essentially everything that ChatGPT, Gemini, and Claude can. That includes text, audio, image, and video generation. Whats more, DeepSeeks newly released family of multimodal models, dubbed Janus Pro, reportedly outperforms DALL-E 3 as well as PixArt-alpha, Emu3-Gen, and Stable Diffusion XL, on a pair of industry benchmarks. DeepSeek-R1, rivaling o1, is specifically designed to perform complex reasoning tasks, while generating step-by-step solutions to problems and establishing logical chains of thought, where it explains its reasoning process step-by-step when solving a problem.oh boy #deepseek Alexios Mantzarlis (@mantzarlis.com) 2025-01-27T16:50:40.640ZWhat DeepSeeks products cant do is talk about Tienanmen Square. Or the Yellow Umbrella protests. Or President Xi Jinpings likeness to Winnie the Pooh. Basically, if its a subject considered verboten by the Chinese Communist Party, DeepSeeks chatbots will not address it or engage in any meaningful way.Andrew Tarantola / DeepSeek / Digital TrendsAs an open-source LLM, DeepSeeks model can be used by any developer for free. OpenAI charges $200 per month for the Pro subscription needed to access o1. DeepSeeks models are available on the web, through the companys API, and via mobile apps. You will need to sign up for a free account at the DeepSeek website in order to use it, however the company has temporarily paused new sign ups in response to large-scale malicious attacks on DeepSeeks services. Existing users can sign in and use the platform as normal, but theres no word yet on when new users will be able to try DeepSeek for themselves.Since the release of ChatGPT in November 2023, American AI companies have been laser-focused on building bigger, more powerful, more expansive, more power and resource-intensive large language models. Rather than seek to build more cost-effective and energy-efficient LLMs, companies like OpenAI, Microsoft, Anthropic, and Google instead saw fit to simply brute force the technologys advancement by, in the American tradition, simply throwing absurd amounts of money and resources at the problem. In 2024 alone, xAI CEO Elon Musk was expected to personally spend upwards of $10 billion on AI initiatives. OpenAI and its partners just announced a $500 billion Project Stargate initiative that would drastically accelerate the construction of green energy utilities and AI data centers across the US. Google plans to prioritize scaling the Gemini platform throughout 2025, according to CEO Sundar Pichai, and is expected to spend billions this year in pursuit of that goal. Meta announced in mid-January that it would spend as much as $65 billion this year on AI development.DeepSeek just showed the world that none of that is actually necessary that the AI Boom which has been helping spur the American economy in recent months and which has made GPU companies like Nvidia exponentially more wealthy than they were in October 2023, may be nothing more than a sham. It also calls into question just how much of a lead the US actually has in AI, despite repeatedly banning shipments of leading-edge GPUs to China over the past year.The bottom line is the US outperformance has been driven by tech and the lead that US companies have in AI, Keith Lerner, an analyst at Truist, told CNN. The DeepSeek model rollout is leading investors to question the lead that US companies have and how much is being spent and whether that spending will lead to profits (or overspending).In short, DeepSeek just beat the American AI industry at its own game, showing that the current mantra of growth at all costs is no longer valid. DeepSeek clearly doesnt have access to as much compute as U.S. hyperscalers and somehow managed to develop a model that appears highly competitive, Srini Pajjuri, semiconductor analyst at Raymond James, told CNBC.If a Chinese startup can build an AI model that works just as well as OpenAIs latest and greatest, and do so in under two months and for less than $6 million, then what use is Sam Altman anymore?Time will tell if the DeepSeek threat is real the race is on as to what technology works and how the big Western players will respond and evolve, Michael Block, market strategist at Third Seven Capital, told CNN. Markets had gotten too complacent on the beginning of the Trump 2.0 era and may have been looking for an excuse to pull back and they got a great one here.Editors Recommendations

Better to share technology broadly, in bite-size increments, than to smother it with regulation from the start, prompted by our worst fears.

"containment breached" Alien: Earth will bring the horror home The prequel series is set a couple of years before the events of the original 1979 film, Alien. Jennifer Ouellette Jan 27, 2025 2:56 pm | 11 Credit: YouTube/FX/Hulu Credit: YouTube/FX/Hulu Story textSizeSmallStandardLargeWidth *StandardWideLinksStandardOrange* Subscribers only Learn moreAlien: Earth, an all-new original series, premieres this summer on Hulu and with Hulu On Disney+. FX/Hulu dropped a surprise short teaser for its upcoming spinoff series, Alien: Earth, during the AFC Championship game last night. What makes it intriguing is the way it's shot entirely from a xenomorph's point of view as the creature races through a spaceship's corridor while a "containment breached!" warning repeats. The final shot said the spaceship is headed on a crash course toward Earth.The official premise is short and sweet: "When a mysterious space vessel crash-lands on Earth, a young woman (Sydney Chandler) and a ragtag group of tactical soldiers make a fateful discovery that puts them face-to-face with the planets greatest threat." We know very little yet about the specifics of the series, other than that it is set two years before the events of the first film, Alien (1979).It's promising that showrunner Noah Hawley has said that the style and mythology will be closer to that film, rather than Prometheus (2012) or Alien: Covenant, both of which were also prequels. In the prequels, Ridley [Scott] made the technology thousands of years more advanced than the technology of Alien, which is supposed to take place in those movies future," he said last January. "Theres something about that that doesnt really compute for me. I prefer the retro-futurism of the first two films, and so thats the choice Ive madetheres no holograms. The convenience of that beautiful Apple Store technology is not available to me.Chandler's character is named Wendy, and apparently she has "the body of an adult and the consciousness of a child." The eminently watchable Timothy Olyphant plays her synth mentor and trainer, Kirsh, and here's hoping he brings some space cowboy vibes to the role. The cast also includes Alex Lawther as the soldier named CJ; Samuel Blenkin as a CEO named Boy Kavalier; Essie Davis as Dame Silvia; Adarsh Gouray as Slightly; Kit Young as Tootles; and Sandra Yi Sencindiver as a senior member of the Weyland-Yutani Corporation. I think we can expect at least some cast members to end up as xenomorph fodder.Alien: Romulus was a welcome return to the franchise's horror roots, and Alien: Earth will bring the horror to our home planet. Theres something about seeing a Xenomorph in the wilds of Earth with your own eyes, Hawley told Deadline Hollywood in September. I cant tell you under what circumstances youll see that, but youll see it and youre going to lock your door that night.As for creature design, "What was really fun for me was to really engage with the creature, bring some of my own thoughts to the design while not touching the silhouette, because thats sacrosanct," he said. "But some of the elements as we know, whatever the host is informs what the final creature is. I just wanted to play around a little bit to make it as scary as it should be.Alien: Earth premieres on FX/Hulu this summer. Credit: FX/Hulu Jennifer OuelletteSenior WriterJennifer OuelletteSenior Writer Jennifer is a senior writer at Ars Technica with a particular focus on where science meets culture, covering everything from physics and related interdisciplinary topics to her favorite films and TV series. Jennifer lives in Baltimore with her spouse, physicist Sean M. Carroll, and their two cats, Ariel and Caliban. 11 Comments

What's the deal? Trump may leave TikToks algorithm under ByteDances control Everything insiders are saying about Trumps plan to save TikTok. Ashley Belanger Jan 27, 2025 2:39 pm | 2 Credit: Bloomberg / Contributor | Bloomberg Credit: Bloomberg / Contributor | Bloomberg Story textSizeSmallStandardLargeWidth *StandardWideLinksStandardOrange* Subscribers only Learn moreIt was apparently a busy weekend for key players involved in Donald Trump's efforts to make a deal to save TikTok.Perhaps the most appealing option for ByteDance could be if Trump blessed a merger between TikTok and Perplexity AIa San Francisco-based AI search company worth about $9 billion that appears to view a TikTok video content acquisition as a path to compete with major players like Google and OpenAI.On Sunday, Perplexity AI submitted a revised merger proposal to TikTok-owner ByteDance, reviewed by CNBC, which sources told AP News included feedback from the Trump administration.If the plan is approved, Perplexity AI and TikTok US would be merged into a new entity. And once TikTok reaches an initial public offering of at least $300 billion, the US government could own up to 50 percent of that new company, CNBC reported. In the proposal, Perplexity AI suggested that a "fair price" would be "well north of $50 billion," but the final price will likely depend on how many of TikTok's existing investors decide to cash out following the merger.ByteDance has maintained a strong resistance to selling off TikTok, especially a sale including its recommendation algorithm. Not only would this option allow ByteDance to maintain a minority stake in TikTok, but it also would leave TikTok's recommendation algorithm under ByteDance's control, CNBC reported. The deal would also "allow for most of ByteDances existing investors to retain their equity stakes," CNBC reported.But ByteDance may not like one potential part of the deal. An insider source told AP News that ByteDance would be required to allow "full US board control."According to AP News, US government ownership of a large stake in TikTok would include checks to ensure the app doesn't become state controlled. The government's potential stake would apparently not grant the US voting power or a seat on the merged company's board.A source familiar with Perplexity AI's proposal confirmed to Ars that the reporting from CNBC and AP News is accurate.Trump denied Oracles involvement in talksOver the weekend, there was also a lot of speculation about Oracle's involvement in negotiations. NPR reported that two sources with direct knowledge claimed that Trump was considering "tapping software company Oracle and a group of outside investors to effectively take control of the app's global operations."That would be a seemingly bigger grab for the US than forcing ByteDance to divest only TikTok's US operations."The goal is for Oracle to effectively monitor and provide oversight with what is going on with TikTok," one source told NPR. "ByteDance wouldn't completely go away, but it would minimize Chinese ownership."Oracle apparently met with the Trump administration on Friday and has another meeting scheduled this week to discuss Oracle buying a TikTok stake "in the tens of billions," NPR reported.But Trump has disputed that, this weekend saying that he "never" spoke to Oracle about buying TikTok, AP News reported."Numerous people are talking to me. Very substantial people," Trump said, confirming that he would only make a deal to save TikTok "if the United States benefits."All sources seemed to suggest that no deal was close to being finalized yet. Other potential Big Tech buyers include Microsoft or even possibly Elon Musk (can you imagine TikTok merged with X?). On Saturday, Trump suggested that he would likely announce his decision on TikTok's future in the next 30 days.Meanwhile, TikTok access has become spotty in the US. Google and Apple dropped TikTok from their app stores when the divest-or-ban law kicked in, partly because of the legal limbo threatening hundreds of billions in fines if Trump changes his mind about enforcement. That means ByteDance currently can't push updates to US users, and anyone who offloads TikTok or purchases a new device can't download the app in popular distribution channels."If we can save TikTok, I think it would be a good thing," Trump said.Could Trumps plan violate divest-or-ban law?The divest-or-ban law is formally called the Protecting Americans from Foreign Adversary Controlled Applications Act. For months, TikTok was told in court that the law required either a sale of TikTok US operations or a US ban, but now ByteDance seems to believe there's another option to keep TikTok in the US without forcing a sale.It remains unclear if lawmakers will approve Trump's plan if it doesn't force a sale of TikTok. US Representative Raja Krishnamoorthi (D-Ill.), who co-sponsored the law, issued a statement last week insisting that "ByteDance divesting remains the only real solution to protect our national security and guarantee Americans access to TikTok."Krishnamoorthi declined Ars' request to comment on whether leaked details of Trump's potential deal to save TikTok could potentially violate the divest-or-ban law. But debate will likely turn on how the law defines "qualified divestiture."Under the law, qualified divestiture could be either a "divestiture or similar transaction" that meets two conditions. First, the transaction is one that Trump "determines, through an interagency process, would result in the relevant foreign adversary controlled application no longer being controlled by a foreign adversary." Second, the deal blocks any foreign adversary-controlled entity or affiliate from interfering in TikTok US operations, "including any cooperation" with foreign adversaries "with respect to the operation of a content recommendation algorithm or an agreement with respect to data sharing."That last bit seems to suggest that lawmakers might clash with Trump over ByteDance controlling TikTok's algorithm, even if a company like Oracle or Perplexity serves as a gatekeeper to Americans' data safeguarding US national security interests.Experts told NPR that ByteDance could feasibly maintain a minority stake in TikTok US under the law, with Trump seeming to have "wide latitude to interpret" what is or is not a qualified divestiture. One congressional staffer told NPR that lawmakers might be won over if the Trump administration secured binding legal agreements "ensuring ByteDance cannot covertly manipulate the app."The US has tried to strike just such a national security agreement with ByteDance before, though, and it ended in lawmakers passing the divest-or-ban law. During the government's court battle with TikTok over the law, the government repeatedly argued that prior agreementalso known as "Project Texas," which ensured TikTok's US recommendation engine was stored in the Oracle cloud and deployed in the US by a TikTok US subsidiarywas not enough to block Chinese influence. Proposed in 2022, the agreement was abruptly ended in 2023 when the Committee on Foreign Investment in the United States (CFIUS) determined only divestiture would resolve US concerns.CFIUS did not respond to Ars' request for comment.The key problem at that point was ByteDance maintaining control of the algorithm, the government successfully argued in a case that ended in a Supreme Court victory."Even under TikToks proposed national security agreement, the source code for the recommendation engine would originate in China," the government warned.That seemingly leaves a vulnerability that any Trump deal allowing ByteDance to maintain control of the algorithm would likely have to reconcile."Under Chinese national-security laws, the Chinese government can require a China-based company to 'surrender all its data,'" the US argued. That ultimately turned TikTok into "an espionage tool" for the Chinese Communist Party.There's no telling yet if Trump's plan can set up a better version of Project Texas or convince China to sign off on a TikTok sale. Analysts have suggested that China may agree to a TikTok sale if Trump backs down on tariff threats.ByteDance did not respond to Ars' request for comment.Ashley BelangerSenior Policy ReporterAshley BelangerSenior Policy Reporter Ashley is a senior policy reporter for Ars Technica, dedicated to tracking social impacts of emerging policies and new technologies. She is a Chicago-based journalist with 20 years of experience. 2 Comments

The incoming president signed a suite of executive orders last weekYURI GRIPAS/POOL/EPA-EFE/ShutterstockIn the first week of his second term, US president Donald Trump has thrown the countrys scientific apparatus into chaos, while at the same time anointing a new tech oligarchy. With a blitz of executive orders, hes set the US on a path that will derail climate goals, biomedical research and pandemic readiness.On 20 January, the first day of his presidency, Trump declared a national energy emergency and proposed expanding fossil fuel production on federal land, including a controversial move to open Alaskas Arctic National

We followed the No. 1 ranked female eater, Miki Sudo, and her husband, Nick Wehry, as they trained for one of the last Major League Eating contests of 2024 the Norms hotcake eating championship in Las Vegas. Over the past three decades, competitive eating in America has evolved from a casual hobby into an increasingly ambitious sport. These athletes train their bodies like machines, crafting every meal and strategizing every bite to beat the competition. Major League Eating, the sport's governing body, works with sponsors to host about 70 contests annually, with up to five figures in prize money.Show more