Enterprise computers are often the ugly ducklings of the PC world, viewed as dull, slow and less feature-rich than their consumer counterparts. But vendors at last weeks Consumer Electronics Show launched a selection of business machines (alongside their consumer offerings) hoping to capitalize on the rise of generative AI (genAI).Despite the focus on AI PCs or Copilot+ PCs, analysts said the vendors might be a bit ahead of the market.I liken it to the dot.com era, said Tom Butler, executive director of portfolio and product management for Lenovos worldwide commercial notebook business. When the dot.com era kicked off, companies immediately said, I need a dot.com instance. So, its very much like that right now. Companies, as we move into this AI PC era, [say] I need an AI PC.Heres a look at some of the noteworthy business PCs announced at CES 2025 and analysis of whether vendors are hitting the mark for enterprise customers.AsusIn addition to its Zenbook and Republic of Gamers (ROG) offerings, Asus unveiled the enterprise-focused ExpertBook B5, ExpertBook B3, ExpertCenter P400 AiO, and ExpertCenter P500. Although theyre notCopilot+ PCs(their neural processing unit (NPU) isnt powerful enough), they qualify asAI PCs; both B5 and B3 laptops include Intel vPro for manageability and have passed the MIL-STD 810H durability tests. Neither is super light, tipping the scales at about 3 pounds.The B5 supports up to 64GB of RAM and up to a 2TB SSD with RAID support, has an all-metal design, 16-in. screen, and security features including a fingerprint reader, facial recognition, and a smart card reader.The B3 has either a 14-in. or 16-in. display, supports up to 64GB of RAM and up to 1TB storage in dual SSDs. And it offers a variety of ports USB-A, USB-C, HDMI, and even an Ethernet port. (The B5 lacks Ethernet, unless you have a USB dongle.)On the desktop side, the ExpertCenter AiO (all-in-one) comes in two models, one with a 27-in. display, the other with a 24-in. screen. The P500 is a mini tower supporting up to 64GB of RAM and up to 4TB storage on one SSD and one hard drive.DellWhileDells rebranding plans, announced at CES, might be a bit perplexing, the company did introduce several new Dell Pro models designed for professional-grade productivity. They come in several flavors: Base, Plus, and Premium, and all qualify as Copilot+ PCs, based on their specs.At the Base level, there are the Dell Pro 14 and Dell Pro 16, designed to deliver essential performance for everyday productivity, Dell said. They feature Intel Core Ultra 5 processors, 16GB of RAM and 256GB of storage, and screen resolution of 19201200 pixels.The company also announced Dell Pro desktops, powered by either Intel or AMD processors, available in micro, slim, and tower form factors. They are, Dell said, the companys first commercial desktops with NPUs.One step up are the Dell Pro 13/14/16 Plus, with up to 32GB of RAM and 1TB of storage. They come in laptop or 2-in-1 form factors and use the same system BIOS to make ordering and management easier for IT departments. Dell claims up to 18.2 hours of battery life for the Pro 14 Plus, and 12.6 hours for the Pro 16 Plus.At the top of the heap are the Dell Pro 13/14 Premium models. Theyre the slimmest and lightest member sof the Pro portfolio, starting at 2.36 pounds. The Dell Pro 13 Premium offers up to 20.8 hours of battery life; the Pro 14 Premium provides up to 21.2 hours. Each can be ordered with an Intel Core Ultra 7 processor, up to 32GB RAM and up to 1TB of storage.HPHPs EliteBook line expanded with the companys announcement of a trio of Copilot+ PCs. The HP EliteBook Ultra G1i Next Gen AI PC is, HP said, designed for executives, with its 14-in. UWVA OLED screen on the Intel-powered model. (The Qualcomm model offers WLED screen technology.) They can have up to 32GB of RAM and up to 2TB of storage for the Intel version, or 1TB of storage for the Qualcomm version. HP touted studio quality microphones and a 9-megapixel camera for high quality video calls.The HP EliteBook X Flip G1i Next Gen AI PC has multiple use modes, including laptop, tablet, and tent configurations, with up to 32GB of RAM and up to 2TB of storage. Like the Ultra G1i, it has a 14-in. screen, although touch comes standard (its an option on the G1i) and its a WLED display, not OLED.The HP EliteBook X G1i Next Gen AI PC is powered by either Intel or AMD chips. The Intel model can hold up to 32GB of RAM; the AMD version offers up to 64GB. Both models can have up to 2TB of storage.The EliteBook X machines will be available in March, with HP saying only that the Ultra G1i is coming soon.LenovoLenovo launched an impressive array of devices, and the two models specifically aimed at businesses both contained surprises.The ThinkPad X9 14- and 15-in. Aura Editions are sleek, thin and light notebooks tested to meetMIL-SPEC 810H standards. Lenovo claims all-day battery life,and says the machines are designed to allow easy servicing of the battery and SSD. They offer up to 32GB of RAM and 1TB of storage.One thing, however, is missing: the X9 is the first ThinkPad to forego the trackstick. Lenovo hastened to note that its just for this model other ThinkPads will continue to have the trademark red nub in the middle of their keyboards.Two years ago, Lenovo showcased a laptop concept with a rollable screen; this year, that concept became a reality. The ThinkBook Plus Gen 6 Rollables 14-in screen expands upwards at the touch of a button, growing to 16.7 inches and providing 50% more screen space. As with the other new models, it offers up to 32GB of RAM and 1TB of storage.It is not a budget-friendly device, though; prices start at $3,500.AI a gimmick?As for whether these systems meet enterprise needs, analysts weighed in on what theyre seeing in the current PC market and they were somewhat dismissive of the AI hype.Current AI features are over-hyped and largely invisible or are seen as thats nice to have tobusiness unless the users are proactive in finding more detailed features, saidRanjit Atwal, senior director analyst at Gartner. All in all, businesses are unlikely to pay more than 5% above normal prices for an AI PC.I think the AI laptops are a gimmick for the most part, efforts by the OEMs to stay relevant and bottle lightning if they can, said Jeremy Roberts, senior research director at Info-Tech Research Group. I have yet to have any of my enterprise or mid-market clients profess to be excited or tell me theyre changing their refresh cycle or anything to incorporate AI features at the PC level.IDCsRyan Reith, group vice president, Worldwide Device Trackers, said he saw some interest in AI PCs early in 2024, but noted it has since waned.What we gathered throughout most of last year, especiallyaround the middle of the year, is that large enterprises and most developed markets around the world were allocating budget for these genAI PCs, Reith said. That enthusiasm declined in the second part of 2024 amid concerns that Microsoft and its partners had not delivered on expectations.Whats important to enterprisesAccording to Roberts, genAI features have not excited the enterprises he deals with. Things that excite enterprises are Autopilot compatibility for seamless deployment, TPM chips for encryption, specialized screens to limit viewing angles, decently powerful CPUs/GPUs depending on the use case, and repairability, he said.I dont think AI features even make the top 10, Roberts said. Most organizations wont be consuming AI features locally anyway theyll use cloud services like CoPilot or Gemini.Reith cited feedback from the head of commercial sales for a large OEM who said with tightened budgets and uncertainly about what can be achieved with genAI PCs, companies are shifting their spending to mid-range computers. They dont qualify as Copilot+ devices, but are still very good PCs.If you get 200 PCs at a mainstream level, as opposed to 50 at a premium level, theyre going more towards the 200, he said.Roberts agreed: Modern laptops are generally more than capable of handling the typical knowledge workers day-to-day. A Dell Latitude 5000 or 7000 (Dell Pro now?) series or a run-of-the-mill ThinkPad from Lenovo wont struggle with Slack, Teams, or PowerPoint.As for what corporate users want, size and weight are often key, Reith said and companies are now listening to their employees rather than choosing what they think they want.The thin and light PCs are the trend, Atwal said, adding, businesses essentially want to future proof their PCs with AI capabilities so [want to] have them include an NPU.Lenovos Butler said the configuration sweet spot has shifted; its now 32GB of memory and a minimum 512GB of storage. And screen sizes have edged up.Most business laptops are in the 14- to 16-in. range, said Roberts. Anything smaller is quite cramped. Anything larger can be cumbersome to lug around not to mention more expensive.Whither AI PCs, then?Reith and Roberts differ on the fate of the AI PC/Copilot+ PC.AI PCs are a solution looking for a problem, Roberts argued. Most end user computing managers wont be swayed by this branding and additional feature set. AI will continue to be delivered primarily via the cloud. I expect the CoPilot+ PC will go the way of the Ultrabook: branding attached to computers that are only marginally relevant to the people who buy and use them.Reith, however, believes its more an issue of timing. Theres a necessity to have on-device AI, he said. Its just getting pushed forward a little bit to when that inflection point really starts to kick in.Im trying to use my words cautiously, because we do not believe that this is dead in the water. It was a good chance that then passed. Its just that the timing was really bad, which no one could have predicted. But nobodys really backing off of the developments. Supply side is now just shifting some of their business plans around products.Whatever we thought was going to be the genAI PC volume in 2025, itll be slightly less than that, in our opinion, Reith said. But a lot of that will just get pushed forward to a ramp up thats maybe more towards the end of this year, and certainly into 2026.

The Apple Watch Series 10 has been on my wrist for three months now and it's clear that it's the best smartwatch Apple ever made. But that doesn't mean there aren't things that a future Series 11 could do better.The Apple Watch Series 10 looks great, if a little familiar.At this point, so early in the review, you're probably ready to scream "Apple Watch Ultra 2" at your screen. Don't worry, I haven't forgotten it, but the Apple Watch Ultra models are different and I'd argue the Series 10 is the better watch.For most Apple Watch buyers, the Ultras are just too big, too bulky, and too expensive. They come with features they don't need like cellular connectivity and a huge, rugged design that's built to deal with the great outdoors in ways the standard models aren't. Continue Reading on AppleInsider | Discuss on our Forums

Construction input prices decreased by 0.2% for the month of December, according to a new Associated Builders and Contractors (ABC) analysis of U.S. Bureau of Labor Statistics Producer Price Index data. Nonresidential construction input prices decreased 0.2% for the month, and overall construction input prices are 0.9% higher than a year ago.Decembers nonresidential construction input prices are now also 0.6% higher. Prices for steel mill products (-3.0%) and iron and steel (-2.5%) dropped while the price of softwood lumber (1.1%) and hot rolled steel bars, plates, and structural shapes (2.5%) increased.Chart credit: Associated Builders and ContractorsConstruction materials prices declined slightly in December and are virtually unchanged over the past two years, ABC Chief Economist Anirban Basu said. Of course, there is significant variability across input categories. Much of the recent moderation can be tracked to lower energy prices; diesel prices, for instance, are down rou...

Arrowhead Game Studios has a new patch for Helldivers 2, addressing crashes and jittering that occurs when ADSing with scopes. The crashes resulted from Illuminate ships despawning or when a player would quit the game and should thus be gone.Other miscellaneous issues that occurred include the targeting for STA-X3 W.A.S.P Launcher not working if you used one belonging to another player. Those who faced disconnects in Illuminate missions can also breathe easier as theyve been reduced.Helldivers 2 is available for PS5 and PC check out our review here. Its had a successful first year with over 15 million players (despite some bumps with the whole mandatory PSN linking and balance). With its first anniversary in February, one has to wonder if the developer has anything special planned. Stay tuned for updates.In the meantime, Arrowhead has a new project with creative director Johan Pilestedt stating hes working on the high concept. Sony also recently greenlit a movie based on the franchise.PATCH 01.002.005OverviewCrash fixesADS jitter fixMiscellaneous fixesFixesCrash Fixes and Soft-locks:Fixed a crash caused by Illuminate ships despawningFixed some potential crashes when the player quits the gameMiscellaneous fixesFixed an issue with jitter when aiming through scopesFixed an issue with the STA-X3 W.A.S.P Launchers targeting not properly working when using one owned by another playerReduced disconnections in Illuminate missions

Music History and Contemporary Art Destroyed in the Deadly Los Angeles WildfiresAn archive of scores by composer Arnold Schoenberg and the collections of countless contemporary artists have been lost in the blaze The Eaton Fire has devastated the community of Altadena. Public domain via Wikimedia CommonsThe wildfires still burning in Los Angeles have killed at least24 people, razed thousands of homes andhistoric landmarks and destroyed countless treasured itemsincluding creative works stored in the area.One such loss is a collection of music written by 20th-century Austrian-American composerArnold Schoenberg, who developed the 12-tone technique, an innovative style of musical composition. An archive of roughly 100,000 scores and parts written by Schoenberg had been kept in a building behind the house of his son, Larry Schoenberg, in Los Angeles Pacific Palisades neighborhood, which has been devastated by thefires.Its brutal, Larry, 83, tells theNew York Times Javier C. Hernndez. We lost everything. Composer Arnold Schoenberg moved to southern California in the 1930s. Florence Homolka via Wikimedia CommonsLarry runs the companyBelmont Music Publishers, which rents and sells Schoenberg scores. The fires didnt destroy any original Schoenberg manuscripts. But without the archive of copies, upcoming performances of Schoenberg pieces may be difficult. As Leon Botstein, the music director of the American Symphony Orchestra, tells the Times, the collection was an indispensable resource. Belmont provided the orchestra with scores for its performance of Schoenbergs Gurrelieder last year.Born in Vienna in 1874, Schoenberg rose up in the European music scene during the early 20th century. As theWashington Posts Michael Dirda wrote in 2023, the composers increasingly esoteric music regularly elicited catcalls and uproar from early audiences. After fleeing the Nazis in 1933, Schoenberg landed in California, where he began teaching at several universities. Per reports from January 12, the composers longtimehome in the Brentwood neighborhood has been spared from fire damage.Larrys home and the archivewhich included photos, letters, books and other composers arrangements of Schoenbergs workswas completely destroyed.Theres a finality here which is astonishing, Larry tells the Times. Theres no hope left that youre going to find or retrieve anything. And thats a different kind of grief.Los Angeles contemporary artistspainters, potters, sculptors and moreare also suffering the loss of creative works. As multidisciplinary artistKathryn Andrews tellsARTnews Karen K. Ho, There are certain things that cant be replaced.Like Larry Schoenberg, Andrews lives in Pacific Palisades. She evacuated her home before it burned down. This is the second time shes lost a house to wildfire: In 2020, theBobcat Fire destroyed her home in Juniper Hills, north of Los Angeles. This time, Andrews lost her own really lovely collection of art, which included works byRashid Johnson,Jim Shaw,Fredrik Nilsen and others. Shed built the collection over two decades.After the last fire, I had the equivalent of a writers blocka creative block, Andrews tells ARTnews. A lot of animals died in that one. Its not just the loss of stuff, you know, its the loss of nature, its the loss of a community, its the loss of dreams. It has a very intense impact.The fires have razed countless local galleries, including AltadenasAlto Beta, which contained a ten-painting show by artistMary Anna Pomonis when it burned down.PainterRachelle Sawatsky tellsCulturedTara Walters, who lives in the Malibu Village neighborhood, tellsHyperallergics Valentina Di Liscia, Matt Stromberg, Maya Pontone and Rhea Nayyar, Everything is gone from my house. My car. My paintings inside. All my heirlooms. My wedding dress. Everything.Kelly Akashi, an Altadena-based sculptor of glass and bronze, lost her home and studio, report theTimes Robin Pogrebin, Julia Halperin and Zachary Small. The building contained recent and archival work, including pieces shed selected to display in her first exhibition at theLisson Gallery later this month. Per the Times, She had considered naming one of her recent works Monument to Loss. Now it is actually lost.The fires also consumed the livelihood of interdisciplinary artist Ross Simonini. As he tells Cultured, the wildfire destroyed his familys home, his studio and nearly all the work I have ever made.That includes childhood drawings I made with my mom, the drawing that helped me believe I could be an artist and several new bodies of work, Simonini adds. I dont think I will ever stop grieving that loss, but the loving response from the art community has already started transforming that grief into something else: a feeling of deep human connection that Ive looked for all my life. It only took losing everything to get it.Get the latest stories in your inbox every weekday.Filed Under: Archives, Art History, Artists, Arts, California, Composers, Contemporary Art, Fire, Music, Musical History, Painters, Painting, Sculpture, Wildfire

NC America announced today Blade & Soul NEO, the modern reinterpretation of Blade & Souls signature fantasy universe, is coming soon to PC in North America and Europe. Featuring a beautiful visually enhanced experience, the original, well-crafted storyline, Infinite Windwalk (the glide ability), and a handful of enticing gameplay features, Blade & Soul Neo fulfills players desire to re-live that pure MMO action fantasy experience they once knew.Offering an exclusive pre-registration package, available today including a cosmetically enhanced character outfit, in-game currency, and more NC America will actively roll out new information, community activities and more on its revamped website between today and launch. You can pre-register for the title now. Blade & Soul NEO is a modern reinterpretation of Blade & Souls iconic Wuxia-themed, fantasy world, offering unparalleled action-based combat and character customization. Players can reconnect with the original spirit of the free-to-play, open-world MMORPG in a richer, visually enhanced experience set in a world of Asian mythology where fast-paced combat collides with the visual beauty.Gamers will re-live the original storyline as a hero determined to restore honor to their clan thats been wiped out by enemy forces. Players can pick one of four playable races based on animal lore, including the Dragon, Qilin, Phoenix and Turtle, and amongst seven classes, including Assassin, Blade Dancer, Blade Master, Destroyer, Force Master, Kung Fu Master, and Summoner.The features remastered graphics and animations: NCSofts graphic overhaul of the original MMORPG highlights enhanced textures and lighting, improved animations, and more environmental details and atmosphere, taking full advantage of the Unreal Engine 4 toolkit.Blade & Soul Neos action-based combat enables players to truly feel like a martial artist, with ability combos, counters, and flexible party roles. Likened to some of the combat systems traditionally found in fighting games, Blade & Soul Neo offers a real skill-based combat system.You can experience enhanced combat customization with more ways to expand ones own distinct fighting style. Through gameplay, players earn Skill Books with unique benefits to build custom combat techniques.Obtainable from the start, Infinite Windwalk, once limited to a stamina bar, is now infinite! Infinite Windwalk (which includes both walking and sprinting) will whisk players into the air for a whole new form of long-form transportation.Players can create their own uniquely designed character with this highly customizable system, including stylizing character hair, facial structures, eye color, height, and body, and then export and share their character templates with other players.Blade & Soul NEO players can engage in numerous different PVP activities ranging from Field Combat to Faction Combat to organized 6v6 battlegrounds.Daily insights on business use cases with VB DailyIf you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory shifts to practical deployments, so you can share insights for maximum ROI.Read our Privacy PolicyThanks for subscribing. Check out more VB newsletters here.An error occured.

App growth firm SplitMetrics appoints new CROCEO says new senior appointments are "instrumental" to company's "next level of growth"Image credit: SplitMetrics News by Vikki Blake Contributor Published on Jan. 14, 2025 SplitMetrics has appointed former Meta and Google sales director Eoin Hallahan as its new chief revenue officer.The company said Hallahan "brings 16 years of sales expertise, with over a decade in senior leadership roles across adtech, martech, and high-growth B2B startups," including roles as enterprise sales director for EMEA gaming at Meta, managing director EMEA & APAC at NextRoll, and Google's sales department. He worked most recently as sales leader at Supermetrics.The firm has also appointed former data.ai executive Dan de Vries as its product director.We are excited to have Eoin on our team as our new CRO," said CEO and co-founder, Max Kamenkov. "His extensive experience in adtech and martech coupled with his proven track record of scaling high-growth businesses makes him a perfect fit for us."Eoin will be an instrumental member of our senior leadership that takes SplitMetrics to the next level of growth and helps us achieve our long-term goals."Kamenkov added: "We are also excited to have Dan de Vries on our team as our new Product Director, who is already making a big impact within our team. Dans proven track record in product management will be invaluable to SplitMetrics as we continue to innovate with our products and services for app marketers and developers."

Justin Carter, Contributing EditorJanuary 14, 20251 Min Readvia GOGAt a GlanceGOG launched a program to preserve games last year, which promises ongoing support for older and delisted PC titles.Digital game storefront GOG is furthering its recent game preservation efforts by joining with the European Federation of Game Archives, Museums, and Preservation Projects (EFGAMP).The European group collects and preserves digital games, and represents "over 20 heritage institutions and other organizations dedicated to preserving video games as cultural heritage." GOG is the first Polish organization in its ranks, which includes the French-based MO5, Berlin's Computerspielemuseum, and the Video Game Museum in Rome.GOG's preservation effortsGOG launched its titular preservation program in November, which pledged to keep classic games (read: 10 years or older) up to date with current and upcoming Windows hardware. In December, the initiative expanded to include titles delisted by their developers, such as Blizzard's first two Warcraft games. Players who purchased them on GOG before their delisting can still access them, and the company pledged all delisted games joining the program will "remain compatible with modern systems, preserving their experience for years to come.""Classic games and the mission to safeguard them for future generations have always been at the core of our work," said GOG's managing director Maciej Gobiewski. "As a European company,we feel a responsibility to lead in preserving gaming heritage. Joining EFGAMP reinforces this commitment."EFGAMP COO Andreas Lange noted the storefront's "unique perspective" in game distribution "complements the work of our existing members by bringing further digital expertize to our collective efforts. GoG is a fantastic addition to EFGAMP, whose members are primarily rooted in the cultural heritage sector."Now that it's part of the EFGAMP, Gobiewski said GOG will "expand institutional collaboration with museums and governmental and non-governmental organizations worldwide. [...] We are also discussing exciting new game preservation projects, which we look forward to sharing soon."Read more about:Game preservationAbout the AuthorJustin CarterContributing Editor, GameDeveloper.comA Kansas City, MO native, Justin Carter has written for numerous sites including IGN, Polygon, and SyFy Wire. In addition to Game Developer, his writing can be found at io9 over on Gizmodo. Don't ask him about how much gum he's had, because the answer will be more than he's willing to admit.See more from Justin CarterDaily news, dev blogs, and stories from Game Developer straight to your inboxStay UpdatedYou May Also Like

Parallels has added support for x86 emulation in Parallels Desktop 20.2, product manager Mikhail Ushakov wrote in a blog post last week. The early technology preview will let you emulate Intel-based hardware on an M1-or-greater Mac, a first for Parallels since Apples Arm transition in 2020 but dont expect stellar performance.Parallels says users will be able to:Run existing x86_64 Windows 10, Windows 11*, Windows Server 2019/2022, and some Linux distributives with UEFI BIOS via Parallels Emulator.Create new Windows 10 21H2 and Windows Server 2022 virtual machines. Image: ParallelsHowever, performance will be really slow, with up to seven-minute boot times, Ushakov says. Other limitations include no external USB device support, Windows 11 24H2 isnt supported, and you can only emulate 64-bit operating systems, though Ushakov says you can run 32-bit apps.He writes that the option to start one of these VMs is hidden for now to avoid false expectations from those who dont need x86 emulation.Parallels has a longer list of caveats in a separate article. Parallels now puts Apple Intelligence Writing Tools in the Windows context menu. Image: ParallelsVersion 20.2 brings some other changes, including support for automatic time and time zone syncing in macOS virtual machines on Apple silicon. It also adds Apples AI-powered Writing Tools to the Windows right-click menu in Word, Powerpoint, and the classic version of Outlook. Before, you had to use a keyboard shortcut or the macOS menu bars Edit menu.

Author(s): Towards AI Editorial Team Originally published on Towards AI. What happened this week in AI by LouieThe Biden administrations Framework for Artificial Intelligence Diffusion has been unveiled, setting out sweeping rules for managing how advanced AI technologies are exported and utilized globally. The framework creates a three-tier system for AI chip exports and model weights, aiming to balance U.S. technological leadership with international cooperation. The rules triggered immediate pushback from Nvidia and other tech companies while receiving cautious support from some national security experts. The framework was introduced in the final days of the Biden administration, leading to speculation about its longevity and the incoming Trump administrations potential to alter or repeal it. The 120-day comment period might also allow for significant changes.At its core, the framework establishes different levels of access to U.S. AI technology. The top tier of 18 close allies maintains essentially unrestricted access. The middle tier, covering most countries, faces caps on computing power imports unless they meet specific security requirements. This group is effectively capped at around 50,000 advanced AI chips through 2027, absent exceptions and agreements. The bottom tier of restricted countries remains largely blocked from access to chips that meet specific criteria.The rules also control exports of AI model weights above certain thresholds (initially a rather arbitrary threshold of 10 computational operations or more during training). However, open-weight models remain uncontrolled. The framework requires security standards for hosting powerful AI systems weights internationally.Nvidia responded forcefully against the framework, calling it unprecedented and misguided while arguing it would derail innovation and economic growth worldwide. The company suggests the rules will push countries toward Chinese alternatives rather than achieve their intended security goals. Other critics warn about diplomatic fallout, noting that many allies find themselves in tier two. However, the administration counters that China lacks the capacity to backfill restricted chip exports in the near term. They argue this creates leverage to encourage the adoption of U.S. security standards in exchange for computing access.Why should you care?This framework could significantly reshape global AI development, but whether it will do so effectively remains to be seen. The broader question is whether this framework will achieve its stated goals or disrupt innovation, backfire against the US current tech leadership, and exacerbate fragmentation in the global AI landscape. Will tier-two countries align with U.S. standards, or will they seek alternatives in Chinese or open-source technologies? Can the U.S. bureaucracy implement such a complex system without causing delays and inefficiencies? Will these rules actually enhance national security, or are they merely symbolic gestures that fail to address the real risks of AI? In the near term, could this create a new bottleneck to GPUs and LLM token access even in the US, as the global data center capacity planned for tier 2 countries gets delayed?Given our belief in the capability, utility, and continued rapid pace of growth of these models, the answers to these questions will shape not just the trajectory of U.S. AI leadership but also the global dynamics of technology and power in the years to come.Hottest News1. Mistral AI Announced Codestral 25.01, an Updated Version of the Codestral ModelCodestral 25.01 features a more efficient architecture and an improved tokenizer than the original, generating and completing code about 2 times faster. The model is now leading the coding benchmarks in its weight class and SOTA for FIM use cases across the board.2. Microsoft Makes Powerful Phi-4 Model Fully Open-Source on Hugging FaceMicrosoft has released its Phi-4 model as a fully open-source project with downloadable weights on Hugging Face. Although Phi-4 was actually revealed by Microsoft last month, its usage was initially restricted to Microsofts new Azure AI Foundry development platform. Now, Phi-4 is available outside that proprietary service and comes with a permissive MIT License, allowing it to be used for commercial applications. The model has received 60,000 downloads on Hugging Face so far.3. Nvidia CEO Says His AI Chips Are Improving Faster Than Moores LawNvidia CEO Jensen Huang claims that the companys AI chips surpass Moores Law, achieving performance improvements 30x faster than previous models. This evolution could lower AI inference costs as newer chips enhance computing capability.4. Mark Zuckerberg Gave Metas Llama Team the OK To Train on Copyrighted Works, Filing ClaimsThe plaintiffs in a copyright lawsuit allege that Meta CEO Mark Zuckerberg approved using pirated content from LibGen to train Llama models despite internal concerns. They accuse Meta of stripping copyright data to conceal infringement and illegally torrenting LibGen, aiming to bypass legal methods.5. NVIDIA Announces Nemotron Model Families To Advance Agentic AIAs mentioned briefly last week, at CES 2025, NVIDIA CEO Jensen Huang launched new Nemotron models, including the Llama Nemotron large language models (LLMs) and Cosmos Nemotron vision language models (VLMs), to improve agentic AI and boost enterprise productivity. The Llama Nemotron models, built on Llama foundation models, allow developers to create AI agents for applications like customer support, fraud detection, and supply chain optimization.6. Cohere Introduces North: A Secure AI Workspace To Get More DoneCohere launched the early access program for North, an all-in-one secure AI workspace platform for improving the quality and speed of their work. North helps employees across teams and industries offload routine tasks and focus their time on where they can add the most value to their company. North combines LLMs, search, and agents into an intuitive platform that integrates AI into your daily work.Five 5-minute reads/videos to keep you learning1. Building Large Action Models: Insights from MicrosoftMicrosofts framework for building LAMs represents a significant advancement in AI, enabling a shift from passive language understanding to active real-world engagement. This article explores the frameworks comprehensive approach, which encompasses data collection, model training, agent integration, and rigorous evaluation, and provides a robust foundation for building LAMs.2. Visualize and Understand GPU Memory in PyTorchThe article explains GPU memory usage in PyTorch, offering tools to visualize memory and optimize usage. It details estimating memory for model parameters, optimizer states, activations, and gradients, considering factors like batch size.3. Best Practices for Building and Deploying Scalable APIs in 2025Building, deploying, and managing an API is a crucial step for scaling products, connecting systems, and ensuring everything communicates seamlessly. This article explores the world of APIs what they are, why you might need one, and what deployment options are available.4. GPT-4o Python Charting Insanity: Prompting For Instant Data VisualsGPT-4o greatly simplifies the process of creating working Python data visualization code by displaying the visualization within the chat window. This post shares a step-by-step tutorial on generating hands-on, no-code visuals from datasets.5. My 6 Secret Tips for Getting an ML Job in 2025There are many different ways to get a job in ML. This blog post shares six tips on landing one. It emphasizes questions you should ask yourself, such as how to demonstrate your skill and how to show that you can actually add value to a project.Repositories & ToolsFACTS Leaderboard is a benchmark that measures LLMs factual accuracy, addressing hallucinations.Trending Open LLM Models is a directory of 5000+ open-source models.Top Papers of The Week1. rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep ThinkingrStar-Math demonstrates that small language models can surpass OpenAI o1 in math reasoning without distillation. Implementing deep thinking via Monte Carlo Tree Search and self-evolution, rStar-Math enhances models to solve advanced problems, raising Qwen2.5-Math-7B to 90% accuracy. It solves 53.3% of USA Math Olympiad problems, placing in the top 20% of students.2. REINFORCE++: A Simple and Efficient Approach for Aligning Large Language ModelsThis paper presents REINFORCE++, an enhanced variant of the REINFORCE algorithm that incorporates key optimization techniques from PPO while eliminating the need for a critic network. This approach offers simplicity, enhanced training stability, and reduced computational overhead.3. VideoRAG: Retrieval-Augmented Generation over Video CorpusVideoRAG dynamically retrieves relevant videos based on query relevance and utilizes visual and textual information in output generation. It leverages large video language models to process video content for retrieval and integrates these videos with queries.4. Dive into Time-Series Anomaly Detection: A Decade ReviewIn recent years, advances in machine learning have led to diverse time-series anomaly detection methods vital for fields such as cybersecurity and healthcare. This survey categorizes these methods under a process-centric taxonomy. It performs a meta-analysis to identify trends in research, highlighting the shift from statistical measures to machine learning algorithms for detecting anomalies.5. TabPFN: A Transformer That Solves Small Tabular Classification Problems in a SecondThis paper presents TabPFN, a trained Transformer that can perform supervised classification for small tabular datasets in less than a second. It does not need hyperparameter tuning and is competitive with state-of-the-art classification methods.Quick Links1. LlamaIndex introduces Agentic Document Workflows, an architecture that combines document processing, retrieval, structured outputs, and agentic orchestration to automate end-to-end knowledge work.2. New OpenAI job listings reveal the companys robotics plans. In a post on X, Caitlin Kalinowski said that OpenAIs robotics team will focus on general-purpose, adaptive, and versatile robots that can operate with human-like intelligence in dynamic, real-world settings.3. Researchers open-source Sky-T1, a reasoning AI model that can be trained for less than $450. Sky-T1 appears to be the first truly open-source reasoning model because it can be replicated from scratch; the team released the dataset they used to train it and the necessary training code.Whos Hiring in AI2025 Summer Internship, Tech Research Scientist PhD @Spotify (London, UK)AI / Large Language Model Architect @Accenture (Multiple US Locations)Python Developer @InvestorFlow (Remote)Data Visualization Analyst @RYZ Labs (Remote within Argentina & Uruguay)AI-Driven Quality Engineering Specialist (PCD-PWD) @CloudWalk (Remote)Software Engineer III Backend @Deputy (Melbourne, FL, USA)Junior Backend Engineer Search @BlaBlaCar (Paris or Remote from France)Interested in sharing a job opportunity here? Contact [emailprotected].Think a friend would enjoy this too? Share the newsletter and let them join the conversation.Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming asponsor. Published via Towards AI