In BriefPosted:2:51 PM PDT March 21, 2025Image Credits:Kyle Grillot/Bloomberg / Getty ImagesAnduril might build a weapons factory in the UKFactories are all the rage in defense tech: Anduril announced a billion-dollar megafactory in Ohio earlier this year, while Saronic said last month its planning its own factory to mass produce autonomous warships.Now, Anduril is considering building a factory in the U.K. as it expands beyond U.S. defense contracts.If we get enough orders, absolutely we are planning to open a facility in the UK, Rich Drake, Andurils general manager for U.K. and Europe, told Bloomberg.Anduril would use the factory for drones and R&D and might choose somewhere near Oxford and Cambridge, Sifted reported earlier.European defense budgets are breaking records as the U.S. winds down aid to Ukraine, so theres plenty of opportunities.Anduril, which is based in California and is reportedly in talks to raise at a $28 billion valuation, didnt respond to a request for comment.Topics

In a blog post last July, Meta CEO Mark Zuckerberg said that selling access to Metas openly available Llama AI models isnt [Metas] business model. Yet Meta does make at least some money from Llama through revenue-sharing agreements, according to a newly unredacted court filing.The filing, submitted by attorneys for the plaintiffs in the copyright lawsuit Kadrey v. Meta, in which Meta stands accused of training its Llama models on hundreds of terabytes of pirated ebooks, reveals that Meta shares a percentage of the revenue that companies hosting its Llama models generate from users of those models.The filing doesnt indicate which specific hosts pay Meta. But Meta lists a number of Llama host partners in various blog posts, including AWS, Nvidia, Databricks, Groq, Dell, Azure, Google Cloud, and Snowflake.Developers arent required to use a Llama model through a host partner. The models can be downloaded, fine-tuned, and run on a range of different hardware. But many hosts provide additional services and tooling that makes getting Llama models up and running simpler and easier.Zuckerberg mentioned the possibility of licensing access to Llama models during an earnings call last April, when he also floated monetizing Llama in other ways, like through business messaging services and ads in AI interactions. But he didnt outline specifics.[I]f youre someone like Microsoft or Amazon or Google and youre going to basically be reselling these services, thats something that we think we should get some portion of the revenue for, Zuckerberg said. So those are the deals that we intend to be making, and weve started doing that a little bit.More recently, Zuckerberg asserted that most of the value Meta derives from Llama comes in the form of improvements to the models from the AI research community. Meta uses Llama models to power a number of products across its platforms and properties, including Metas AI assistant,Meta AI.I think its good business for us to do this in an open way, Zuckerberg said during Metas Q3 2024 earnings call. [I]t makes our products better rather than if we were just on an island building a model that no one was kind of standardizing around in the industry.The fact that Meta may generate revenue in a rather direct way from Llama is significant because plaintiffs in Kadrey v. Meta claim that Meta not only used pirated works to develop Llama, but facilitated infringement by seeding, or uploading, these works. Plaintiffs allege that Meta used surreptitious torrenting methods to obtain ebooks for training, and in the process due to the way torrenting works shared the ebooks with other torrenters.Meta plans to significantly up its capital expenditures this year, largely thanks to its increasing investments in AI. In January, the company said it wouldspend $60 billion-$80 billion on CapEx in 2025 roughly double Metas CapEx in 2024 primarily on data centers and growing the companys AI development teams. Likely to offset a portion of the costs, Meta is reportedly considering launching a subscription service for Meta AI thatll add unspecified capabilities to the assistant. Updated 3/21 at 1:54 p.m.: A Meta spokesperson pointed TechCrunch to this earnings call transcript for additional context. Weve added a Zuckerberg quote from it specifically a quote about Metas intent to revenue share with large hosts of Llama models.

The Penn Station redevelopment saga, recently quiet but never dormant, has taken an abrupt plot twist, as proposals emerge and re-emerge against a background of renewed attention toand a high-level edict requiringclassical design. General Project Plan: Update or Scrap?On March 6, Assemblyman Tony Simone (with support from City Councilman Erik Bottcher, Borough President Mark Levine, and State Senator Brad Hoylman-Sigal) called for amending the General Project Plan (GPP) to replace the ten Vornado Realty commercial towers proposed in former governor Andrew Cuomos initial GPP with a mixed-use complex comprising one-third housing and two-thirds offices.Simones proposal would rely on a new GPP that would still override local zoning to facilitate construction but reject two controversial aspects: the option of seizing private property through eminent domain; and the demolition of Block 780, bordered by 30th and 31st Streets between Seventh and Eighth Avenues, to replace residences, businesses, and St. John the Baptist Catholic Church with a proposed southern expansion of the station. Simones GPP proposal calls for about 5,000 housing units (up from the existing GPPs 1,800) plus a public park on the Seventh Avenue site of the newly demolished Hotel Pennsylvania.The existing GPP, predicated on financing a new Penn Station with tax revenue from the towers, has been decoupled from the station renovations since Governor Kathy Hochuls announcement in June 2023, as the weak commercial real estate market made that revenue stream unlikely. For community activists, preservationists, and proponents of alternative plans, the GPP remains a punching bag, yet the Governor has not withdrawn it outright. Her stated openness to proposals by any architect, any design firm, any engineer has not taken the form of an RFP. Work on the redesign has been the province of a selected 55-member Station Working Advisory Group (SWAG) since September 2024.Washburns plan is one of several that would restore McKim, Mead, and Whites original Beaux Arts entrance arcade and colonnades, features that were destroyed when the original station was demolished in 1963. (Courtesy Grand Penn Community Alliance)The Pitch from the Grand Penn Community AllianceA few days after Simones announcement, another proposal seized the spotlight. At a March 11 press conference at The New York Historical Society, Grand Penn Community Alliance (GPCA) executive director Alexandros Washburn presented details, from financial charts to virtual-reality simulations, and announced that the GPCA would soon submit documents to the U.S. Department of Transportation.The GPCAs plan is not a new vision but instead the latest iteration of a verdant neoBeaux-Arts design that Washburn presented at Cooper Union in January 2023. Today it benefits from support from Thomas D. Klingenstein, the Claremont Institute chairman and financier of right-wing causes who donated $10 million to Republican campaigns during the 2024 election. The scheme by Washburn differs from the two leading proposalsthe Metropolitan Transportation Authority (MTA)s plan and the public-private partnership (P3) of ASTM-Halmar, HOK, and PAUby requiring that Madison Square Garden (MSG) find a new site. Where the Garden now stands, Grand Penn proposes a park roughly the size of Bryant Park. The ASTM-Halmar/HOK/PAU P3s plan can accommodate a future Garden move but does not require it. The MTAs design leaves MSG in place.A fourth scheme by Richard Camerons Beaux-Arts Atelier, supported by ReThinkNYC and known as the McKim Variations, offers three versions contingent on the Garden moving, both the Garden and 2PENN (Two Penn Plaza) moving, or both remaining in place, with the Garden redesigned to harmonize with his station that evokes the original structure, designed by McKim, Mead and White.A single train concourse is central to the the Grand Penn plan, which would allow for the implementation of through-running trains and greater programmatic flexibility. (Courtesy Grand Penn Community Alliance)Washburn was previously the chief urban designer for the City and public works advisor to Senator Daniel Patrick Moynihan. He recently told AN that part of the method of Grand Penn is to give big wins to all the stakeholders. Recalling his contention, writing in Metropolis in 2007 (while serving in the Bloomberg administration), that nature is the new civic ideal, he noted that Grand Penn provides ample green space for the West Side, which is undersupplied with parks. Further community benefits include sparing certain landmarks that we would love to carry forward into our future, the Church of St. John in particular, though the plan sacrifices part of Block 780 for a southern station expansion.For the railroads, Grand Penn offers doubled capacity with the creation of a 604,000-square-foot train concourse: The most important aspect of rail operations to be supported by the station is flexibility. We give flexibility through that enormous, open single train concourse to reprogram trains as through-running or as commuter or as intercity or as regional rail.As for MSG, GPCA proposes a new arena across Seventh Avenue. Washburn is unfazed by MSG Entertainment president James Dolans well-known opposition to relocating the Garden. It has to be a business deal that appeals to them, he said. Mr. Dolan is a very good businessman; hes actually even a visionary, when you look at the Sphere and other projects hes done. At the vacant Hotel Pennsylvania site and adjoining property extending to the intersection of 34th Street and Seventh Avenue, a new Garden would retain essential transit access. The savings from building a new station without the Garden above are substantial, considering gains in speed, safety, and simplified logistics. The cost of the new MSG, including land, is estimated at $3.5 billion and is included in the overall $7.5 billion estimate, a figure that matches the MTA plans figures. As for Vornado, they would gain both from the land purchase for the arena and the rising value of their other local properties.Weve developed a set of measured drawings, cost-estimated them, and they are now a reasonable alternative set, Washburn explained. We are submitting those to US DOT, and they perform to certain standards. For instance, we have 3.1 times the number of entries that the current station does. We have 1.99 times the amount of square footage on the platform. Critically, particularly for observers concerned with the ventilation, circulation, and safety problems raised in the June 2023 MSG-Penn Station Compatibility Report, GPCAs plan has 10 times the emergency ventilation.Washburn contends that the GPCAs plan reframes the Penn conundrum to reduce stakeholder conflicts and give the city an infrastructural asset that can last at least a century. Were the only plan that states the problem correctly, which is how to get the best train station, Washburn said. The answer, he thinks, requires a new arena and an opening to above.Some opponents argue that the Grand Penn plan prioritizes architectural aesthetics over much-needed improvements to transit infrastructure. (Courtesy Grand Penn Community Alliance)Reanimating Public DebateWhen approached for comment on the Simone and Grand Penn plans, the MTA press office referred AN to MTA CEO Janno Liebers remarks at a March 11 press conference, emphasizing what actually has been done to make Penn Station better (e.g., the upgraded 33rd Street concourse) rather than longer-range plans. We are, as a transportation agency, focused [on] what we can do now. Our focus is doing things that can deliver for customers now, without tearing up Penn Station and making it unlivable for another generation.Critiquing the MTAs current plan on multiple grounds, from its reliance on a flawed 2021 technical review to its cost estimates (recently adjusted to account for miscalculated HVAC expenses), Sam Turvey, chairperson of ReThinkNYC, said, Governor Hochul should sponsor the design competition for Penn Station which she promised in June of 2023 once the transit options are fairly evaluated and a track plan determined. He links the current logjam to interagency turf battles and calls for Amtrak management to make everybody in New Jersey Transit, the MTA, and Amtrak check their egos at the door and specifically for MTA management to understand that New York deserves better. Transportation engineer Robert Paaswell, director emeritus of CUNYs University Transportation Research Center, also prefers placing the transit horse before the architectural cart, questioning whether a GPP (commercial or mixed), Grand Penn, or any development plan can avoid worsening congestion if it proceeds without first bringing the MTA up to a state of good repair and operations. For a station in the epicenter of what should be a rebirth of American rail, including a future for high-speed rail, he said, unless you put what the future of rail is in there its hard to do any planning.Paaswell contrasted his work on the Port Authority redesign jury with the current Penn arrangement: SWAG isnt a working group. Its a bunch of people sitting in a room nodding at presentations that are made. The eight or ten people that we had on the Port Authority were a real working group, because we were given the plans, and we each had to go around the table and hammer them out, hour after hour, rather than just listening and then saying, Oh, this is good. Well vote on alternative A or alternative B. His recommendations for Penn include an independent competition; vetting of city, state, and Regional Plan Association numbers by population experts and economic forecasters; an openness to outside investors (whether in P3 form or not); and everybody sacrificing a little bit of their own authority to get a much higher buy-in.The Grand Penn proposal is one of several that has advocated for relocating Madison Square Garden. (Courtesy Grand Penn Community Alliance)Horse, Meet Cartand Watch for the ElephantPAUs Vishaan Chakrabarti supports Simones amendments for more housing and more sensitivity to historic fabric while favoring a basic reordering of priorities. The GPP puts the cart before the horse, he said, in the sense that, whether its office or residential, its still primarily talking about transit-oriented density without talking about how to fix the transit. Youve got to start with the infrastructure and then figure out what is the right form of development around it. Noting that operational questions about through-running service, as seen in Londons Elizabeth Line and other systems, are separate from architectural questions, he pointed out that our plan can work with the existing train shed, and it can work with a reconfigured track layout that provides through-running.As for Grand Penn, he is skeptical. Moving the Garden, an idea he has advocated in past proposals, strikes him as no longer achievable. We have not had a governor since Eliot Spitzer who was interested in moving the Garden, he said, recalling the history of efforts to find an alternate double-block site near transit, including his own past proposal for the two blocks south of Macys. It would be lovely to move Madison Square Garden, but the government doesnt want to pay to do it, and the Garden has no interest in doing it.Contrasting the visuals prepared by other teams with the ASTM-Halmar/HOK/PAU P3s extensive structural and mechanical drawings and actual financial backing, he views the MTA plan and his own teams as the only serious contenders. I dont understand why replacing an underground station with an arena on top of it with an underground station with a park on top of it is better, he said. He also offered some stylistic feedback: The neoclassical design seems like a naked appeal to the President. Like, come on, go full-on Albert Speer. If youre going to do it, do it.Backed by right-wing financier Thomas D. Klingenstein, the GPCA plan aligns with the Trump administrations promotion of classical architecture. (Courtesy Grand Penn Community Alliance)Here, Chakrabarti has identified the living-room elephant that some Penn observers, and even active participants, are reluctant to name. The various proposals reflect a widely shared belief that Penn Station is long overdue for an overhaul, yet any suggestion that the current political regime could be appropriate agents for its replacement is certain to be divisive. GPCA has support from the National Civic Art Society (NCAS), whose board includes Klingenstein. The NCAS is one of the few arts organizations to support the White Houses 2020 and 2025 executive orders mandating classical architecture in federal buildings, opposed vigorously by the AIA and recently likened to 20th-century dictators anti-modernist paranoia by Steven Holl in Dezeen. Chakrabarti pointed out that the newer order is much shorter and less stylistically prescriptive, as it requires federal buildings to respect regional, traditional, and classical architectural heritage. Still, even a transient convergence of interests between urban design advocates and the Trump administration is a salient case of politics creating strange bedfellows.Public Review or Back-Room Deal?On March 19, Governor Hochul spoke of revising Penn plans to avoid destroying a neighborhood, encouraging preservationists by specifically stating opposition to Amtraks intentions for a southern expansion that would demolish Block 780. To date, however, she has not announced a public design competition that would follow through on her 2023 comments, while calls for an independent review of transit fundamentals await official answers. Hochuls ongoing conversations with the President have given rise to the suggestion that improving Penn Station may be one area where their interests could convergeand, with the Department of Transportations deadline for ending congestion pricing now postponed by a month, wide-ranging speculation about the potential tradeoffs, far removed from Penn Station, that could be involved.Bill Millard is a regular contributor to AN.

From the top of the Ferris Wheel at the 1893 Chicago World Columbian Exposition, Simon Pokagon, an Indigenous rights activist born in 1830, addressed the burgeoning city, and saw an apocalyptic tide of change. How unlike the Chi-Kag-Ong of the red man! he wrote in the New York Times Sunday Magazine. The shoreline of the lake, with its fleet of canoes; the marsh and winding river, with flags and rushes fringed; the scattering wigwams and the red men were nowhere to be seen. But in their place rose roof-on-roof, with steeples tall, smoking towers and masts of ships as far as [the] eye could see. All had changed, except the sun and sky above. They had not, because the great spirit, in his wisdom, hung them beyond the white mans reach. Pokagon, a member of the Potawatomi tribe, and his critique of the dispossession and distortion that came along with this tide of canonical architecture has a special place in the permanent installation at the MacArthur Foundation headquarters in Chicago, which opened this fall. Curated by two Indigenous artists and scholars (John Low, of the Pokagon Band of Potawatomi, and Andrea Carlson, of the Ojibwe tribe), We Will Tell Our Story crafts an intervention that corrects the historical record and materializes Simon Pokagons critique of Chicago architecture from within, in building with its own history of distorted settler-colonial narratives.The exhibition was designed by Michigan-based Seven Generations A+E. (Courtesy MacArthur Foundation)Located in Holabird and Roches 1895 Marquette Building, an archetypal example of the Chicago Schools technological and formal innovations, the building has been celebrated, and landmarked, for its lobby, decorated with mosaics by Louis Tiffany and J.A. Holzer that depict Jacque Marquettes settler-colonial exploration of Great Lakes region in 17th century and his interactions with Indigenous people, who are represented inaccurately. The clothing and architecture depicted (teepees, feather headdresses) are associated with plains Indians further west. The European settlers are seen serenely passing a peace pipe in one mosaic, and in another Indigenous Americans calmly listen to Marquettes impassioned exhortations to the Christian faith.We Will Tell Our Story focuses on these misrepresentations explicitly. The consequences of this meeting were not Peace and Prosperity, but were conflict, dispossession, and diaspora, reads one panel. The first section of the museum visitors see as they exit the original lobby is labeled The dishonesty of the Marquette Building. While the original lobby is atmospheric, allegorical, and falsely idealized, the new exhibition, designed by Michigan-based Seven Generations A+E, which is owned by the Pokagon Band of Potawatomi, is didactic, scholarly, and text-heavy. Its primary architectural organizing device is a set of birch wood shelves and frames inspired by the domed wigwam, vernacular shelter for Great Lakes Indigenous people, which wraps around the perimeter of the exhibition, creating a sense of curvilinear embrace as it arches over visitors heads. The goal was to reference in an abstract and respectful way a traditional architectural or structural methodology, said Alex Hokkanen, Manager of Design Research + Sustainability at Seven Generations A+E.The lighting and materials similarly draw aesthetic counterpoints to the lobby. Delicate, bright track lighting, the light-hued wood, and brilliant block panels of Indigenous ribbonwork patterns contrasts with the more subdued colors and subtle lighting of the lobbys pointillist mosaics. Three curving exhibition walls that center the Potawatomi tribes experience are themed Recollections, Power, Time, and Land, all orbiting a structural column recast as the center-point of this circle.Birch wood shelves and frames inspired by the domed wigwam wraps around the perimeter of the exhibition. (Courtesy MacArthur Foundation)Throughout, theres an intense focus on not historicizing Indigenous people. The installation addresses the 20th-century urbanization of Native people in Chicago and elsewhere and emphasizes that no matter how old the cultural practices of Indigenous people may be, their persistence today makes them contemporary, not prehistoric artifacts. Were not gone, were not conquered, were also not discovered, said Low. Furthermore, the exhibit makes it clear that the wigwam and effigy mound can lay claim to the title of the first Chicago School of architecture.The contemporary dynamism of Indigenous life is expressed in the architecture of the exhibition as well. The cellular organization of the wigwam-like structure can be used to frame 2D works, as its done now with a selection of drawings and photographs of Indigenous people by Indigenous artists, and as a display case for 3D objects, so that new exhibitions can be rotated in and out. The present tense, as a general guiding principal, was very strong, said Hokkanen. The project began several years ago, when the MacArthur Foundation sought to refresh a previous exhibition in this space, which focused on the foundation and the Marquette building. In conversation with Indigenous community members, the foundation realized they had the opportunity to cede more leadership to Native voices, Native leaders, and Native artists, said Jamie Waters, a staff member at the MacArthur Foundation who was co-chair of the exhibition working group.Exhibition content emphasizes that no matter how old the cultural practices of Indigenous people may be, their persistence today makes them contemporary. (Courtesy MacArthur Foundation)The urge to demonstrably alter the lobby to point out its historical inaccuracies would be understandable, but between its landmark protections and the obvious progression in Indigenous depiction the two spaces demonstrate, the curators felt this fundamental contrast should be embraced. In and of itself its a history lesson of how white people thought about Indians at the time the Marquette building was built, said Low. We dont want to lose that lesson.Editors Note: Zach Mortices wife is a staff member at the MacArthur Foundation, though she was not involved in the organization, production, or promotion of We Will Tell Our Story.Zach Mortice isa Chicago-based design journalist and critic focused on architecture and landscape architectures relationship to public policy.

The White-Silsbee House is a stately, Federal period, brick mansion built in Salem, Massachusetts for wealthy merchant, Joseph White. Constructed in 1811 for Joseph White, Jr., the house uniquely faces Oliver Street and not the Salem Common, like that of his brothers home nextdoor. Joseph White (1780-1816) was bred to the sea, and grew up at the time of Salems commercial expansion into the markets of India, China, and Sumatra in the far reaches of the Indian Ocean. Captain White evidently had command of an East India vessel on at least one voyage. After Joseph Whites untimely death in 1816, at the age of 36, his widow Eliza continued to live here until 1831 when the property was sold to William Silsbee (1779-1833), a merchant, and his wife, Mary. The three-story, hipped-roof, dwelling featuring a fan light transom and Ionic portico. Most of the windows contain 6/6 sash with exterior storm windows, capped by splayed stone lintels with molded and beaded keystones. A full-length window is centered above the entrance porch; the third story windows are shortened.

Kerry Wan/ZDNETThe Amazon Echo Frames (3rd Gen) are a natural way for glasses wearers to control their smart home, communicate with friends and family, and listen to audio wherever and whenever.The latest model features a lighter build, longer battery life, improved speakers, and new controls that will take some getting used to.While Alexa in your ears is no ChatGPT voice assistant, it's capable enough to answer general questions and complete most smart home tasks.Amazon's Echo Frames (3rd Gen) are on sale for $180 at Amazon. That's 33% savings on these frame's regular price, and one of the best offers we've seen in a few months. The last time we saw a similar offer was at QVC back in January.The headline for this article could've gone many ways --I wore Amazon's Echo Frames at an airport, and the TSA didn't stop me orThese smart glasses let me take calls at CES hands-free --but I settled with "Alexa on my face" because that's the use case most people will likely have for the latest smart glasses from Amazon. It's not as weird as it sounds, I promise.Also: CES 2025: The 8 most advanced smart glasses we tried - and were impressed byWith these being the company's third generation of Echo Frames, Amazon's made mostly iterative updates -- the build is lighter, the battery lasts longer, and there's supposedly more bass -- while staying true to the glasses' original purpose: Giving you a direct (and natural) communication path with Amazon's popular voice assistant, Alexa. And right now, QVC is discounting the Echo Frames by $90, reducing the price of these glasses to $180 now.Of course, you can do other things with the glasses, too, as I'll detail in my two-week account of wearing the Echo Frames below. details View at QVC In the hierarchy of smart glasses, you can think of the Echo Frames as the entry-level pair, ideal for users who want something discrete but with just enough technology to scratch that consumer itch. I've worn a few too many pairs of smart glasses over the past year, and these from Amazon may be the most normal-looking of them all. That's a good thing, as I learned during my flight to CES in early January.Also: Amazon developing smart glasses to help drivers deliver packages faster - reportWanting to test both the Echo Frames and the Meta Ray-Ban during my trip, I went through the usual security screenings and found myself stopped twice because my camera-equipped Meta glasses looked suspicious. No one questioned the Echo Frames, which, even from up close, look like a regular pair of plastic-made prescriptions. (Or maybe no one expected me to have two pairs of smart glasses. I'm not sure.) The Meta Ray-Ban (left) has a built-in camera and is noticeably thicker and heavier than the Amazon Echo Frames (right). Kerry Wan/ZDNETTrue to Amazon's word, the Echo Frames feel very light on the face, with most of the weight distributed to the sides of the frames, where the speakers, buttons, and other minuscule components are stored. The rubberized tips on the ends of the glasses certainly help with the fit, though I've found it harder to fold the glasses after adjusting them.While the glasses are easily stored in the included carrying case, if they're not folded properly, issues arise when you need to charge them. Because, unlike the Meta Ray-Ban's USB-C carrying case, there's a new separate charging dock for the Echo Frames, and aligning the wireless charging coils with the ones on the glasses can be quite the doozy. You either have to force the glasses into the gap -- which feels like something you wouldn't want to do with glasses -- or collapse and slot them in just right. Nine times out of 10, I'm doing option one.In terms of functionality, the Echo Frames don't have multimodal AI cameras built into them like the Meta Ray-Bansor project visual overlays like XR glasses. Instead, they can stream Bluetooth audio, take calls and send text messages, read notifications, and communicate with Alexa-supported smart home devices like how your phone or smart hub normally would. Again, entry-level glasses -- and that's not a bad thing.Also: New Meta Ray-Ban AI features roll out, making the smart glasses even more temptingWhile my house is not assmart as my ZDNET colleague Maria Diaz's, I enjoyed using the Echo Frames to turn on and off various smart switches, which saved me from having to walk over to my phone and use an app when it's charging in the kitchen or on my nightstand. The scale of Alexa-compatible devices is large enough that I could make commands to secure my three-year-old August entry lock, too. The same "Alexa" wake word is used for commands. Kerry Wan/ZDNETBesides the smart features, I've also enjoyed listening to podcasts and making calls during my evening walks. Perhaps it's the ability to listen to my on-device audio while staying aware of my surroundings, or the fact that the speakers on the Echo Frames favor the mid and high frequencies more than the bassy, engulfing lows. It's probably a bit of both. The takeaway is that the glasses are great for vocal-only audio tracks and passable for anything multi-layered.Also: 10 reasons the Apple Vision Pro is secretly brilliantLastly, I wish the button navigations weren't so complicated, but since I'm mainly using voice commands to get things done, the need to press the front or back button to accept and decline things (or mute the microphone) didn't bother me all that much. If you're coming from an older pair of Echo Frames, you'll have to retrain your muscle memory as the touch strip and swipe gestures are no more.ZDNET's buying adviceAt a list price of $270 (but currently discounted to $180 at Amazon), the Echo Frames (3rd Gen) aren't necessarily priced to compete. The list price is just $30 less than Meta's more capableRay-Ban smart glasses.However, Amazon's glasses have two big things going for them: Alexa integration and design. Smart home enthusiasts will absolutely reap the benefits of having a controller that they can use at all times (with a battery life that can last them all day), and first-time smart glasses shoppers will find these much less intimidating than the ones with mini projectors and cameras built in. Plus, you'll have a smaller chance of being stopped by the TSA. When will this deal expire? Deals are subject to sell out or expire at any time, though ZDNET remains committed to finding, sharing, and updating the best product deals for you to score the best savings. Our team of experts regularly checks in on the deals we share to ensure they are still live and obtainable. We're sorry if you've missed out on this deal, but don't fret -- we're constantly finding new chances to score savings and sharing them with you atZDNET.com. Show more This article was originally published on January 26, 2024, and was updated on March 21, 2025.Featured reviews

The headquarters of semiconductor company Marvell Technology in Santa Clara, CaliforniaGetty ImagesMarvell Technology is a global semiconductor company specializing in data infrastructure solutions, generating approximately $5.5 billion in annual revenue. The company is a leader in AI-driven infrastructure, providing critical semiconductor solutions for industries such as data centers, automotive and networking. The company has grown substantially, and part of the fuel of that growth has been in better harnessing the insights that data provide. The leader who has helped lead Marvell to become a data driven culture is Nishit Sahay.Sahay has been with Marvell for seven years, and he has been the Chief Information Officer of for roughly two. He has spearheaded the development of Marvells digital foundation, implemented AI-driven efficiencies, managed significant acquisitions and positioned the company for sustained innovation.Marvell's Market Position in Data InfrastructureMarvell has evolved beyond its traditional semiconductor roots to become a key player in data infrastructure. Its components are deeply embedded in industries that rely on processing, moving and securing large volumes of data. This shift has aligned the company with the rise of artificial intelligence, an area in which it now plays a pivotal role."Marvell is one of the top semiconductor companies fueling AI infrastructure. Whether its data centers, automotive or carrier networks, everything relies on data, and were at the heart of it," said Sahay. By leaning into data infrastructure, Marvell has distinguished itself among industry giants and made itself integral to the modern data economy.Managing Data, AI and Business TransformationAs CIO, Sahay leads four major areas: traditional IT operations, the data office, enterprise AI initiatives and business process transformation. Each of these contributes to an overarching goal of creating a nimble, data-driven enterprise that is aligned to business outcomes.Nishit Sahay, CIO of Marvell Technology Apple Photos Clean Up"At Marvell, data is central to our culture. We dont rely on subjective conversations; every decision is backed by data analytics and insights," noted Sahay. His approach integrates advanced technology not just in IT, but across all business units, creating a culture of analytics and operational rigor.Building a Digital Foundation for Agility and SecurityMarvell has invested heavily in developing a scalable, resilient and secure technology infrastructure to support its expanding business. According to Sahay, this includes systems for HR, engineering and other key functions, all designed with automation and integration in mind.We wanted automation across all functions, from HR systems to engineering platforms, explained Sahay. "This ensures agility and efficiency while maintaining a strong security foundation." These digital underpinnings, while largely invisible to most users, are essential for enabling innovation and maintaining performance at scale.Leveraging Data for Business GrowthData has long been core to Marvells operations, and its role has only grown more critical in recent years. The companys data office supports functions ranging from analytics and data science to governance and engineering, all of which contribute to enterprise-wide visibility and smarter decision-making.We automated our data-driven culture, making analytics accessible to decision-makers, said Sahay. This helped us navigate crises, optimize inventory and balance supply-demand efficiently. The company is also implementing a data mesh architecture to unify structured and unstructured data, ensuring integrity and compliance as AI becomes more central to operations.From Acquisition to Integration: Building Strategic SynergyMarvells growth has been bolstered by a series of strategic acquisitions, including the $10 billion purchase of Inphi in 2021. The integration of acquired companies has been executed with uncommon speed and accuracy, hallmarks of Marvells operational discipline.Many companies struggle with acquisition integration, Sahay noted while pivoting with pride. At Marvell, we perfected the process, integrating Inphis supply chain, CRM and portals within two months with 99.6% data accuracy. The ability to integrate systems and processes so efficiently ensures that strategic acquisitions become accretive to the business much faster.Silicon Design in the Cloud: Rethinking an Industry NormIn a bold move, Marvell has begun shifting its silicon design processes to the cloud, something most semiconductor companies have resisted due to legacy workflows and infrastructure demands. Sahay and his team partnered with hyperscalers like AWS to re-architect the necessary platforms.Unlike other semiconductor firms that use cloud for burst capacity, we are fully moving silicon design workflows to the cloud, noted Sahay. "This allows us to optimize workloads using CPUs, GPUs and ARM-based systems more efficiently." This shift not only modernizes Marvells design capabilities but also positions it at the forefront of customizable compute and engineering agility.AI as a Core Business DriverArtificial intelligence has become a key enabler for innovation and efficiency across Marvells value streams. The companys AI strategy focuses on product development, enterprise-wide efficiency, change management and long-term investment in platforms and governance.AI isnt just about automation; its about reimagining how we work, said Sahay. "We encourage engineers to build AI solutions, ensuring innovation happens at every level." With over 40 applications in production, 60% developed outside the IT function, Marvell is democratizing access to AI and embedding it deeply across the organization.Looking to the Future: Quantum Computing and Emerging TechnologiesSahay is already setting his sights on the next wave of transformative technologies. Chief among these is quantum computing, which he sees as both a promising and disruptive force for the semiconductor industry.Quantum computing has limitless potential but also significant risks, especially in security, he said with a note of excitement. Were closely monitoring developments to ensure Marvell stays ahead of the curve. As the company builds on its strong digital and AI foundations, it is also preparing for a future where quantum may play a central role.With a robust digital foundation, a forward-thinking AI strategy and a commitment to future technologies, Sahay has helped Marvell remain well-positioned for continued success in the fast-evolving tech landscape.Peter High is President of Metis Strategy, a business and IT advisory firm. He has written three bestselling books, including his latest Getting to Nimble. He also moderates the Technovation podcast series and speaks at conferences around the world. Follow him on Twitter @PeterAHigh.

The power of partnership has been demonstrated by the AI Alliances milestones and global expansion ... More in its first year of operation.PixabayThe AI Alliance was founded in December 2023 by IBM and Meta, along with 50 other initial members. Over a years time, its membership has grown quickly, so that now it has more than 140 worldwide members and allows companies, non-profits and academic institutions of all sizes to collaborate on building a robust and open AI ecosystem. In the bigger picture, the AI Alliance has become an important force in the democratization of AI, which is why its worth reviewing what it has accomplished in its first year-plus of existence.Before we get into those specifics, its important to understand why the organization was founded in the first place. For most of AIs history, open source development was a fragmented effort that led to underperforming models. Prior to 2023, there were few nonprofit institutions capable of training AI models with even GPT-2 capabilities. At that time, large tech companies dominated proprietary AI, and open source AI was mainly confined to niche applications.Everything changed in 2023 when multiple new base models with permissive licenses were released. Then by mid-2023, Meta released its open source Llama 2 model in partnership with Microsoft. Within six months, it was used to create more than 10,000 derivative models. A major new phase of open source AI development was underway.In this context, the AI Alliance established an impressive list of goals right from its inception. These goals included fostering open collaboration, establishing governance and guardrails for AI and developing benchmarking tools and clear policy positions. Additionally, the alliance prioritized extensive educational initiatives and nurturing robust hardware ecosystems. The AI Alliances strength is further demonstrated by the quality of its steering committee, which has a roster of well-known commercial organizations and universities.(Note: Moor Insights & Strategy has client relationships with about a dozen of the AI Alliance's 140-plus members.)Membership And Its StandardsCriteria for AI Alliance membershipIBMTo join the AI Alliance, an organization must meet four important standards. First, the potential member must be aligned with the mission to cultivate safety, open science and innovation. Second, members must be committed to work on significant projects aligned with the Alliances mission. Third, the potential member must be willing to contribute to the diversity of perspectives and cultures that exist within its global membership of 140-plus organizations. In the future, it is expected that the membership will grow even larger and even more diverse. The final expectation for membership is reputation; the AI Alliance seeks members with a recognized reputation as an educator, builder or advocate within the AI open source community.Alliance members generally fall into one of those three categories. Builders are responsible for models, datasets, tools and applications that use AI. Enablers evangelize the adoption of open AI technologies using tutorials, use cases and general community support. Advocates emphasize the benefits of the AI Alliance ecosystem, plus foster public trust and safety among organizational leaders, societal stakeholders and regulatory bodies.AI Alliance Focus AreasThe AI Alliance focuses its work on six major areas.Moor Insights & StrategyThe AI Alliance defines its long range priorities in terms of six focus areas. However, it doesnt restrict its activities just to these areas. The alliance takes a holistic approach to the entire AI ecosystem by encouraging its community members and developers and allowing them to participate in one or more these areas, then switch if interests or priorities change.Here are the six key focus areas of The AI Alliance:Skills and Education provide AI knowledge for both consumers and business leaders trying to evaluate the risks of using AI, as well as students and developers building AI applications. This area was designed to make it easier to locate expert guidance for a given area. It also contains a model evaluation initiative. Last year, the alliance created the Guide to Essential Competencies for AI. That guide was the result of an extensive survey done to understand roles needed in AI and what skills are needed in those roles. Even though the first version of the guide was published in mid-2024, it has already had nine revisions. Another follow-up survey is planned to clarify issues discovered in the first survey.Trust and Safety explores these important issues, which are necessary for all AI applications to be successful. Benchmarks, tools and methodologies are used to ensure that models and the applications that use them are high-quality, safe and trustworthy, including support for evolving standards of conduct and effective responses to risks. This work group gathers best-of-breed concepts about trust and safety, then points users to the expertise they need. The State of Open Source AI Trust and Safety End of 2024 Edition survey published on the AI Alliance website covered both needs and successes associated with trust and safety. There are some research and environmental gaps that are being addressed through research and development efforts by many AI Alliance members.Applications and Tools explore tools and techniques for building efficient and robust AI-enabled applications. This group is also building an AI lab to enable experimentation and testing of AI applications and to accelerate innovation.Hardware Enablement is fostering a robust AI hardware accelerator ecosystem by ensuring that the AI software stack is hardware-agnostic. Technologies like MLIR and Triton are key software tools to ensure high-performance hardware portability. These tools allow organizations to take advantage of whatever hardware they prefer. Increased flexibility and performance reduces dependency on proprietary systems.Foundation Models and Datasets focus on models for underserved areas, including multilingual, multimodal, time series, science and other domains. For example, the science and domain-specific models target climate change, molecular discovery and the semiconductor industry. Effective models and AI application architectures require useful datasets with clear governance and rights for use. The Open Trusted Data Initiative is clarifying requirements for such datasets and building catalogs of compliant datasets. This effort should largely eliminate concerns about legal, copyright and privacy issues.Advocacy of regulatory policies is needed to create a healthy and open AI ecosystem. All AI policies and regulations should represent balanced rather than biased viewpoints.Making AI Safer: New Tools, Tests And TransparencyTrust and Safety is an important and large field within the AI Alliance. It has many specialists working on tools to detect and reduce hate speech, bias and other harmful material. The Trust and Safety Evaluation Initiative is a major initiative for 2025 providing a unified view of the entire spectrum of evaluation, not just for safety, but also for performance and other areas where evaluating the effectiveness of AI models and applications is required. A sub-project is exploring specific safety priorities by domain for areas such as health, law and finance.In mid-2025, the AI Alliance plans to publish a Hugging Face leaderboard that will allow developers to search for evaluations that best fit their needs, to compare how open models perform against those evaluations, and to download and deploy those evaluations to examine their own private models and AI applications. That initiative will also provide guidance on important safety and compliance aspects of use cases.Not all invocations of AI models will use hosted commercial services. Some situations require air-gapped solutions. AI-enabled smart edge devices make it likely that more companies will begin deploying new small and powerful models on-premises, and sometimes without an internet connection. To support these use cases and facilitate large-scale model serving with flexible hardware configurations, the AI Alliance is developing hardware-agnostic software stacks.Collaboration Creates AI Alliance InnovationsTwo examples will illustrate how open collaboration between alliance members is providing significant benefits for everyone. The first example, SemiKong, was a collaborative effort among three alliance members. The members created an open-source large language model for the semiconductor manufacturing process domain. Manufacturers can use this model to accelerate the development of new devices and processes. SemiKong contains specialized knowledge about the physics and chemistry of semiconductor devices. In only six months, SemiKong captured the attention of the global semiconductor industry.SemiKong was created by fine-tuning a Llama 3 base model using datasets curated by Tokyo Electron. The tuning created an industry-specific generative AI model that had more knowledge about semiconductor etching processes than the generic base model. A technical report on SemiKong is available here.The second example is DANA the Domain-Aware Neurosymbolic Agents project. It is a joint development of Aitomatic Inc. (based in Silicon Valley) and Fenrir Inc. (based in Japan). DANA is an early example of the now-popular agent architecture, where models are integrated with other tools to provide complementary capabilities. While models alone can provide amazing results, many studies have shown that LLMs often generate incorrect answers. A 2023 study cited in the SemiKong paper measures typical LLM errors of 50%, while DANAs complementary use of reasoning and planning tools increased the accuracy to 90% for the target applications.DANA uses neurosymbolic agents that combine the pattern recognition of neural networks with symbolic reasoning that supports rigorous logic and rules-based capabilities to solve problems. Logical reasoning combined with tools for planning (such as for designing assembly-line processes) produce accurate and reliable results that are essential for industrial quality control systems and automated planning and scheduling.DANA can be used for multiple domains. For example, for financial forecasting and decision-making, DANA can understand market trends and make predictions based on complex theories, using both structured and unstructured data. That same ability can also be applied to retrieval and evaluation of medical literature and research information to ensure that existing diagnoses and treatments meet established medical protocols and practices. In short, DANA can enhance patient outcomes and reduce errors in critical patient applications.Accomplishments In 2024 And BeyondThe AI Alliance began 2025 in a strong position with members in 23 countries and a number of working groups focused on major AI issues. The AI Alliance has over 1,200 working-group collaborators working on 90-plus active projects. Internationally, the AI Alliance has participated in events held in 10 countries involving more than 20,000 people, and it has published five how-to guides on important AI topics to help researchers and developers build and use AI.The AI Alliance has published examples for using AI on models such as IBMs Granite family and Metas Llama models. Its growing curation of recipes utilize the most popular open libraries and models for popular application patterns, including RAG, knowledge graphs, neurosymbolic systems and emerging agent planning and reasoning architectures.In 2025, the AI Alliance is committed to scaling up its reach and impact tenfold. Two of its new major initiatives, discussed previously, are the Open Trusted Data Initiative and the Trust and Safety Evaluation Initiative. The AI Alliance also plans to create an industry-standard community lab for developing and testing AI application technologies. Its domain-specific model initiatives will also evolve. For example, the new Climate and Sustainability Working Group has plans to develop multimodal foundation models and open source software tooling to meet major challenges in climate change and its mitigation.By 2030, it is estimated that AI will help grow the global economy by $20 trillion. By then, it is forecasted that 70% of industrial AI applications will be run on open source AI. It is also expected that the shortage of AI professionals will become even more pronounced than it is today. AI Alliance members may be able to reduce that problem by collaborating with other members to gain access to diverse expertise and resource sharing.The AI Alliance is following a similar growth trajectory followed by other successful open-source organizations, such as the Linux Foundation, the Apache Software Foundation and the Open Source Initiative. These include:Comprehensive AI education and skills programsGlobal advocacy for responsible AICreating tools to ensure AI safety and trustworthiness, as well as ease of development and useCollaborative research with academic institutionsI believe the AI Alliance will continue to attract developers, researchers, business and government leaders as contributors and collaborators. The AI Alliances leadership has established scaling of global collaboration as its overarching mission for 2025. Everything considered, the AI Alliance has the foundation to grow into a dominant global force that shapes, improves and innovates the future of artificial intelligence.

Forward-looking: Recent big game releases appear to confirm that ray tracing and path tracing are the future of graphics rendering. However, these techniques remain prohibitively computationally expensive. The next major update to Microsoft's DirectX API aims to address this and facilitate broader adoption of the technology. Microsoft's GDC presentation this week offered a glimpse into the future of DirectX ray tracing support. The company claims that DirectX Raytracing 1.2 could help developers double the performance of ray tracing and path tracing.Two key features underpin the update: opacity micromaps (OMM) and shader execution reordering (SER). OMM can help path-traced games run up to 2.3 times faster by reducing shader invocations and optimizing opacity data to enhance rendering efficiency. Meanwhile, SER intelligently groups shader execution to minimize divergence, improving performance by up to 2 times.Although ray tracing first emerged as a flashy extra feature in games like Battlefield V and Cyberpunk 2077, recent titles have made it mandatory, indicating that it will soon become standard. Examples include Star Wars Outlaws, Indiana Jones and the Great Circle, and Assassin's Creed Shadows.Path tracing is a more advanced form of ray tracing that significantly improves the accuracy of dynamic lighting and shadows but comes with substantial performance costs. It typically requires high-end GPUs to run effectively in games such as Cyberpunk 2077, Alan Wake 2, Black Myth: Wukong, Indiana Jones and the Great Circle, and the recently released Half-Life 2 RTX demo. Doom: The Dark Ages, launching in May, will also require ray tracing and offer optional path tracing.Also see: Path Tracing vs. Ray Tracing, Explained // Related StoriesDirectX Raytracing 1.2 will become available to developers starting next month. As a result, ray tracing and path tracing could become significantly less demanding in new games over the next few years. Unsurprisingly, Nvidia's RTX graphics cards will support the API update first, while Microsoft is collaborating with AMD, Intel, and Qualcomm to expand support.Microsoft also shared more details on cooperative vectors and neural rendering, which aim to integrate AI workloads into real-time graphics rendering. A new foundational feature, Neural Block Texture Compression, significantly reduces memory usage potentially benefiting users with GPUs that have 12 GB of VRAM or less. Additionally, neural supersampling and denoising are expected to enhance image quality in path-traced games.DirectX neural rendering and cooperative vector support first appeared in January. Microsoft explained that neural rendering optimizes matrix-vector operations for AI training and enables smaller neural networks to run efficiently during GPU shading processes.

Gamers who are looking for monitor deals to upgrade their screen should set their sights on the 39-inch LG UltraGear OLED gaming monitor, which is available from LG with a $500 discount that slashes its price from $1,500 to only $1,000. Thats an excellent offer, but it gets better if you prefer a multi-monitor setup. If you buy two of the gaming monitors, youll get an additional 5% off, and if you buy three or more, you get 10% off. Youre going to have to hurry with your purchase though, as were not sure how much time is remaining before you miss out.The 39-inch LG UltraGear OLED gaming monitor will let you further enjoy the best PC games through the lifelike visuals and smooth animations provided by OLED technology, a 240Hz refresh rate, and a 0.03ms response time. The gaming monitor also enables complete immersion in the games that youre playing with its 800R curvature that fills your field of vision, and support for AMDs FreeSync Premium Pro and Nvidias G-Sync that prevent screen tearing and stuttering.For those who want a dual-monitor setup (or even more screens), the 39-inch LG UltraGear OLED gaming monitor is perfect because of its space-saving design, and the virtually borderless screen means theres not much of a barrier as you move your eyes across the multiple displays. You can also adjust the height, tilt, and pivot so that you can get the monitors at the perfect angle for you.RelatedIf you want to maximize the output of the machine you bought from gaming PC deals, youre going to need a powerful display such as the 39-inch LG UltraGear OLED gaming monitor. If you act fast, you can get it at $500 off from LG itself, which brings its price down to just $1,000 from $1,500. Whats more, if youre thinking about buying two of them, youll get an additional 5% discount, and if you buy three or more, youll get a 10% discount. Theres no time to waste if youre interested though, as the offers may already be gone if you delay your transaction to tomorrow.Editors Recommendations