Welcome to TechCrunchs regular AI newsletter! Were going on hiatus for a bit, but you can find all our AI coverage, including my columns, our daily analysis, and breaking news stories, at TechCrunch. If you want those stories and much more in your inbox every day, sign up for our daily newslettershere.This week, billionaire Elon Musks AI startup, xAI, released its latest flagship AI model, Grok 3, which powers the companys Grok chatbot apps. Trained on around 200,000 GPUs, the model beats a number of other leading models, including from OpenAI, on benchmarks for mathematics, programming, and more.But what do these benchmarks really tell us? Here at TC, we often reluctantly report benchmark figures because theyre one of the few (relatively) standardized ways the AI industry measures model improvements. Popular AI benchmarks tend to test for esoteric knowledge, and give aggregate scores that correlate poorly to proficiency on the tasks that most people care about. As Wharton professor Ethan Mollick pointed out in a series of posts on X after Grok 3s unveiling Monday, theres an urgent need for better batteries of tests and independent testing authorities. AI companies self-report benchmark results more often than not, as Mollick alluded to, making those results even tougher to accept at face value.Public benchmarks are both meh and saturated, leaving a lot of AI testing to be like food reviews, based on taste, Mollick wrote. If AI is critical to work, we need more.Theres no shortage of independent tests and organizations proposing new benchmarks for AI, but their relative merit is far from a settled matter within the industry. Some AI commentators and experts propose aligning benchmarks with economic impact to ensure their usefulness, while others argue that adoption and utility are the ultimate benchmarks.This debate may rage until the end of time. Perhaps we should instead, as X userRoon prescribes, simply pay less attention to new models and benchmarks barring major AI technical breakthroughs. For our collective sanity, that may not be the worst idea, even if it does induce some level of AI FOMO.As mentioned above, This Week in AI is going on hiatus. Thanks for sticking with us, readers, through this roller coaster of a journey. Until next time.NewsImage Credits:Nathan Laine/Bloomberg / Getty ImagesOpenAI tries to uncensor ChatGPT: Max wrote about how OpenAI is changing its AI development approach to explicitly embrace intellectual freedom, no matter how challenging or controversial a topic may be.Miras new startup: Former OpenAI CTO Mira Muratis new startup,Thinking Machines Lab, intends to build tools to make AI work for [peoples] unique needs and goals.Grok 3 cometh: Elon Musks AI startup, xAI, has released its latest flagship AI model, Grok 3, and unveiled new capabilities for the Grok apps for iOS and the web.A very Llama conference: Meta will host its first developer conference dedicated to generative AI this spring. Called LlamaCon after Metas Llama family of generative AI models, the conference is scheduled for April 29.AI and Europes digital sovereignty: Paul profiled OpenEuroLLM, a collaboration between some 20 organizations to build a series of foundation models for transparent AI in Europe that preserves the linguistic and cultural diversity of all EU languages.Research paper of the weekImage Credits:Jakub Porzycki/NurPhoto / Getty ImagesOpenAI researchers have created a new AI benchmark, SWE-Lancer, that aims to evaluate the coding prowess of powerful AI systems. The benchmark consists of over 1,400 freelance software engineering tasks that range from bug fixes and feature deployments to manager-level technical implementation proposals.According to OpenAI, the best-performing AI model, Anthropics Claude 3.5 Sonnet, scores 40.3% on the full SWE-Lancer benchmark suggesting that AI has quite a ways to go. Its worth noting that the researchers didnt benchmark newer models like OpenAIs o3-mini or Chinese AI company DeepSeeks R1.Model of the weekA Chinese AI company named Stepfun has released an open AI model, Step-Audio, that can understand and generate speech in several languages. Step-Audio supports Chinese, English, and Japanese and lets users adjust the emotion and even dialect of the synthetic audio it creates, including singing.Stepfun is one of several well-funded Chinese AI startups releasing models under a permissive license. Founded in 2023, Stepfun reportedly recently closed a funding round worth several hundred million dollars from a host ofinvestors that include Chinese state-owned private equity firms.Grab bagImage Credits:Nous ResearchNous Research, an AI research group, has released what it claims is one of the first AI models that unifies reasoning and intuitive language model capabilities.The model, DeepHermes-3 Preview, can toggle on and off long chains of thought for improved accuracy at the cost of some computational heft. In reasoning mode, DeepHermes-3 Preview, similar to other reasoning AI models, thinks longer for harder problems and shows its thought process to arrive at the answer.Anthropic reportedly plans to release an architecturally similar model soon, and OpenAI has said such a model is on its near-term roadmap.

Rosotics, a former U.S. developer of metal 3D printers, has announced that its Halo system, a large-format metal 3D printer, will cease commercial sales and be consolidated internally to support the production of its orbital transfer vehicles. With this transition, the company seeks to evolve from solely manufacturing launch vehicle hardware to becoming a fully complete operator in autonomous space infrastructure, said Christian LaRosa, Founder and CEO, Rosotics.Every great name in additive manufacturing that exists today became a great name because of what they used it to do. We dont plan to ignore the potential we see here, and that will demand a lot of hard choices to realize. The scope of what we intend to do in orbit will require us to be led and operated under a sole focus, and that is as a space company we plan to win the race to industrializing space, said LaRosa.Backed by Silicon Valley venture capitalist Tim Draper and recognized by Venture Capital Fund Sequoia Capital and Venture Capital Firm Andreessen Horowitz (a16z), Rosotics has secured US$2.6 million in funding to advance large-scale metal 3D printing. The Halo platform has drawn attention from companies like Relativity Space and RTX, as well as Jim Cantrell, who was part of SpaceX and served as Elon Musks early mentor.Key Features of the Halo SystemHalo is a large-format metal 3D printer, dubbed the worlds first supercreator. This term was coined by the company to highlight its ability to directly interface with materials science through adaptive code. Halo is not a 3D printer in the traditional sense; it goes beyond being just a tool. It redefines the concept of 3D printing. This system has the ability to control metallurgical science through a deep understanding of materials, representing a significant leap in the evolution of creation technology, said LaRosa.A key feature of the Halo system is its multi-phase electromagnetic process, which replaces traditional laser-based techniques. This process is driven by a print head assembly called Mjolnir, named after Thors hammer. Each Halo unit is equipped with three Mjolnir heads, mounted on freestanding towers supported by multi-axis gimbals. The company explained that the platforms control system, HalOs, incorporates industrial intelligence to monitor and adjust the deposition process, ensuring stability and precision during operation.Rosotics Restructuring and Future Aerospace InitiativesCEO LaRosa announced the departure of Co-Founder and COO Austin Thurman, alongside the companys decision to relocate its headquarters from Mesa, AZ, to Cape Canaveral, FL. In this new phase, Rosotics is actively recruiting talent to support the growing demands of vehicle production and spaceflight operations. The company also intends to streamline its maritime sector operations through strategic partnerships, shifting its focus toward space exploration.Rosotics announced it plans to deploy multiple autonomous spacecraft that will execute a maneuver around the Moon before entering orbit at the L5 Lagrange point, a stable gravitational position between the Earth and the Moon. These spacecraft, featuring large 3D-printed, unpressurized hulls, will transport smaller vehicles to L5 and later be repurposed to aid in the construction of an orbital station. Initially functioning as a propellant depot, the station will expand over time to include power generation and materials processing capabilities, all while remaining at the L5 point without the need for fuel to maintain its position.3D Printing for Space ApplicationsAlthough Laser Powder Bed Fusion (LPBF) are the most widely used metal additive manufacturing systems, other 3D printing technologies are also used for specific applications. Kansas City-based startup Raven Space Systems is focusing on addressing the demands of aerospace, space, and defense applications with its proprietary Microwave Assisted Deposition (MAD) 3D Printing technology. The company claims that its patented process enables scalable direct ink-write (DIW) 3D printing of aerospace-grade thermoset and preceramic composite components.According to Raven Space Systems, MAD technology holds promise for the production of hypersonics, propulsion vehicles, reentry vehicles, satellites, aircraft, missiles, and space rockets.Elsewhere, satellite telecommunications company AscendArc has secured US$3.95 million in pre-seed and seed funding to further develop its satellite communications technology. This announcement follows AscendArcs recent partnership with OPTISYS, an industrial engineering company focused on developing a 4.5-meter dish with a high-performance horn antenna. The collaboration integrates 3D printed radio frequency (RF) components from OPTISYS, combining precision engineering and advanced manufacturing techniques to enhance satellite signal performance.What 3D printing trends should you watch out for in 2025?How is the future of 3D printing shaping up?To stay up to date with the latest 3D printing news, dont forget to subscribe to the 3D Printing Industry newsletter or follow us on Twitter, or like our page on Facebook.While youre here, why not subscribe to our Youtube channel? Featuring discussion, debriefs, video shorts, and webinar replays.

A legal battle is underway within the 3D printed firearms community. Last year, Florida-based firearm attorney Matthew Larosiere filed a lawsuit against Cody Wilson, best known for his association with the Liberator, one of the first widely available 3D printed gun files.Larosiere accuses Wilson of stealing his copyrighted gun designs and selling them without permission on the file-sharing site DEFCAD. The case was filed in the U.S District Court Middle District of Florida Orlando Division. According to the defendants and experts consulted for this article, the case could have wide-ranging implications for open-source 3D printing.Wilson has hit back against the lawsuit. He claims that the arguments of his gun-designing rival are unfounded because digital files of functional objects, including firearms, are not protected by copyright. The 37-year-old gun rights activist told 3D Printing Industry that Larosiere has resorted to copyright without principle out of basic pride, avarice, and a will to influence the culture of 3D gun printing.Open-source advocate Michael Weinberg shared his insights on this case. The OSHWA board member echoed Wilsons claims that firearms cannot be copyrighted and explained why the lawsuit could establish a precedent that shapes the future of 3D printing.However, Larosiere maintains that his files, registered with the U.S. Copyright Office (USCO), are protected and cannot be considered functional in the digital realm. The plaintiff accuses Wilson of stealing his designs for profit and falsely associating him with a massively problematic individual. Dismissing concerns that the lawsuit could have broader implications for open-source, Larosiere called the case an existential threat to Wilsons business model of charging money for other peoples work.Larosiere is seeking damages for lost revenue, reputational harm, and legal costs. A partner at Zermay Law Group and former Director of Legal Policy at the Firearms Policy Coalition, he has called for a court order to stop Wilson from redistributing the files.I also spoke with Dr. Yannick Veilleux-Lepage, an Assistant Professor at the Royal Military College of Canada whose research spans terrorism, political violence, and criminal use of emerging technologies. He believes this dispute reflects broader divisions within the 3D printed firearms space, which faces ideological tensions and personality clashes. According to the global terrorism expert, this highlights inherent contradictions within a community that thrives on decentralization and open-source collaboration.James Madison Memorial building of the US Library of Congress, which houses the US Copyright Office. Photo via Architecture of Congress.Matthew Larosiere sues Cody Wilson for copyright infringement Larosieres copyright infringement lawsuit names Wilson as a defendant alongside DEFCAD and Defense Distributed (Wilsons open-source digital firearms organization). Dioskouri LLC is also targeted. According to the official court document, this entity is run by Wilson and operates PrecursorParts.com, a website that reportedly handles online payments for USB drives containing gun files.In the lawsuit, Wilson is accused of uploading several copyrighted files, images, and text documents registered with the USCO to DEFCAD without permission. These include the Plastikov V4, BUBAR, SF5, KF5, MPP99, Hitchhiker, and Amigo Grande, 3D gun designs owned by Larosiere.DEFCAD is accessible through a $60-per-year-subscription. Wilson and Defense Distributed also allegedly sell USB drives containing digital files. Larosiere believes the defendants are re-posting his work solely for financial gain. Its certainly just to make money, he told me. Its not fair for somebody to make money from my hard work, especially in this clear and brash way.The pro-gun attorney is also concerned that Wilsons actions are damaging his reputation, causing him personal and professional harm. He claims that DEFCAD falsely implies that the copyrighted works were uploaded with permission and that the uploader allows unrestricted rights to create and distribute derivative works.Additionally, Larosiere claims that DEFCAD has created fake accounts falsely purporting to be 3D gun designers, including himself (under his online handle Fudd Busters) and John Elik (also known as Ivan The Troll). Elik designed the infamous FGC-9 3D printed gun and contributed to some of the copyrighted models in this case.Any files he [Wilson] gets, he makes a profile for that person and forces the association, the plaintiff stated. Larosiere asserted that he never agreed to his designs being used in this way. I did not upload any of these things, but they have made their website look like I have, and they even put my logo on it, he added. I dont want people to think Im affiliated with this guy.Larosiere believes Wilson has created fake profiles to avail himself of Section 230 protections. Part of the Communications Decency Act, Section 203 protects online platforms from liability for content published by third-party users. The firearms lawyer told me that Wilson misunderstands section 230, because the protection falls away when theyre not the ones uploading it.Wilson responded to these allegations, stating, Like most things in this case, Larosiere misunderstands what hes looking at. The University of Central Arkansas graduate explained that his file-sharing platform includes an encyclopedic function. As such, DEFCAD keeps multiple points of metadata about a file, including its original creator. Wilson noted that while Elik happened to create his own profile on the site, it would be accurate to assign his files to his profile even if he didnt.Inside the U.S. Copyright Office. Photo via the U.S. Copyright Office.Can 3D printed guns be copyrighted?Wilson stated that the only allegations he is concerned with are whether the USCO can register the 3D design files of useful objects like firearms or their components. This is a claim that the defendants vehemently dispute, meaning the lawsuit would fail at the first hurdle. Out of the gate, its tempting to ask ourselves questions like: Are 3D models copyrightable? But that isnt the right inquiry, Wilson told me. Instead, Wilson highlighted the U.S. Copyright Act, which denotes only works as copyrightable.Pointing to the lawsuit, Wilson claims that Larosiere is under the mistaken impression that 3D models or CAD files are a distinct or separate category of protectable work. The Forbes 30 Under 30 alumni added, If you ask him [Larosiere], hell say hes only registered 3D design models. But thats not even how American copyright works. You can only register creative works, of which 3D models are merely considered copies.Wilson emphasized that copyright legislation only protects creative works and does not cover functional devices. He highlighted Section 101 of the Copyright Act, which explicitly excludes protection for designs of useful articles, explained Wilson. The law defines such articles as having an intrinsic utilitarian function that is not merely to portray the appearance of the article or to convey information.According to the defendant, In our case, the works are not creative 3D sculptures, or 3D visual artworks, as Larosiere suggests. Instead, he noted, the files in question cover functional firearm parts like assemblies and receivers. The decisions behind their creation are functional. Each component is meant to assemble into a functioning open bolt .22 firearm.Weinberg added that firearms themselves cannot be protected by copyright because Copyright does not protect functional, useful objects. While patents can be used to protect functional models, none of the disputed files feature any patent protection. Larosiere should have pursued design patents, added Wilson. However, since most of his work is notoriously given away to the public for free, he cant.Weinberg highlighted that the Copyright Office registered Larosieres designs as models and not functional objects. He believes this registration has unclear implications for those who want to use them as functional objects. However, Weinberg calls it bad policy to extend copyright protection to CAD files for purely functional objects like firearms. This is because it could become a back door way to expand copyright protection to functional objects.Ultimately, Wilson is confident in defeating Larosieres suit. I dont think Larosiere is experienced enough in IP matters, or litigation, to prevail, the gun designer declared. His decision to bring a lawsuit was intended more to influence the current community of designers in the 3D gun space than to create good law. While a victory for Larosiere in the district court would require an appeal to overturn, Wilson emphasized that he is ready to go the distance.Cody Wilson. Photo via Cody Wilson.Larosiere defends copyright claims Despite Wilsons arguments, Larosiere remains confident that his designs are copyright-protected. The Floridian gun lawyer argues that the Copyright Act safeguards all works of the visual arts, including 2D and 3D models and technical drawings. He explained, A useful article has to be a physical object, meaning the firearm designs and models are protectable since they exist in the digital realm.He argues that the defendants belief is a misunderstanding based on the Star Athletica v. Varsity Brands case. This copyright protection dispute arose when Varsity Brands, a major cheerleading uniform manufacturer, sued Star Athletica for copying its uniform designs. The Supreme Court sided with Varsity, ruling that if a design can exist separately from the functional object, it can be copyrighted.Wilson and his co-defendants cite this ruling in their response, arguing it validates their claims that 3D printable gun files are not copyrightable. However, according to Larosiere, this presupposes that his registered designs are useful articles. The Floridian gun lawyer argues that the Copyright Act safeguards all works of the visual arts, including 2D and 3D models and technical drawings. The plaintiffs official reply to Wilsons response points to Tecnoglass, LLC v. RC Home Showcase, Inc., a case which found the statutory definition of useful articleexcludes technical drawings.Larosiere stressed that a copyrighted drawing or modelsuch as those Wilson allegedly stole and republishedshould not be mistaken for the physical object it represents. If models and technical drawings have the same protection as the thing they depict, it would mean that you could not copyright a technical drawing or a model, he noted. So why would the Copyright Acts specifically list technical drawings and models in both two and three dimensions?The Plaintiff also told me that his lawsuit is not concerned with challenging the ability of individuals to 3D print gun files, even those protected by copyright. Instead, Larosiere is solely focused on preventing Wilson from selling the copyrighted models. Nowhere do we ever say you cannot make a physical thing, he explained. Its always stop selling my drawing.In their response to the lawsuit, the defendants agreed to remove the allegedly infringing images and text for the duration of the case. They reportedly made this decision because claims about 3D models are the most important part of both the overall action and this motion. However, they maintain that claims about 2D photos and text are completely meritless and plan to seek a final judgment dismissing them. According to Wilson and his co-defendants, this makes the arguments against these files moot in the current lawsuit. Larosiere, however, argued that this move amounts to an admission that Wilson stole the copyrighted photographs and written works, and claimed that some remain online. He still has like three of them up. We brought it to their attention, and they didnt take them down, Larosiere said. Among the remaining files is allegedly a word-for-word reproduction of the Hitchhiker gun design. Wilson denies these accusations. He told me that Larosiere has hidden five miniature pieces of text/images in the designs to work as a kind of DRM. Wilson claimed that these pieces of ancillary expressive material serve as Lockout Codes. The Defendant argues that these are hidden in the files to help an otherwise invalid copyright claim survive a courts severability analysis. However, Wilson avowed, Its not going to work here. The Zermay Law Group attorney also criticized Wilsons Black Flag White Paper, a critique of Larosiere and copyright law. The document frames the enforcement of intellectual property as antithetical to the libertarian and buccaneering origins of 3D printed firearms. Larosiere, however, calls it an aggressive misunderstanding of copyright, wherein he admits to removing the copyright notices from my models.Matthew Larosiere. Photo via Matthew Larosiere.Division in the 3D printed gun communityLarosiere and Wilsons legal dispute highlights growing tensions within the 3D printed firearms community. Veilleux-Lepage describes an interesting paradox among supporters of 3D printed guns. On one hand, the community is seen as a broad church, uniting people with diverse ideologies under the shared goal of advancing privately made firearms. In contrast, geographic and ideological divisions have surfaced, fueling very public feuds and personality clashes between key figures.Veilleux-Lepage notes that many members advocate for an open-source spirit to refine designs, improve manufacturing techniques, and push the limits of what is possible with 3D-printed firearms. This level of freedom reflects the core libertarian and anti-establishment ethos which underpins many 3D printed gun designs. In a previous conversation, Veilleux-Lepage noted that, in some circles, the act of choosing a 3D printed firearm is driven by ideology as opposed to necessity. Homemade weapons are symbols of defiance against state control, he explained, unlocking unrestricted access to firearms around the world.However, the community is not homogeneous, leading to significant underlying conflicts. A geographical divide exists around the American-centric focus on part-kit completion kits, which combine 3D printed and conventional firearm parts inaccessible outside the US. According to Veilleux-Lepage, designers and makers in other countries argue that this approach limits the global applicability of many firearm projects. He also highlighted a fundamental ideological divide between those who advocate for fully open-access build guides and STL files and those who aim to monetize their designs.Although Wilson charges money for gun files, he stressed that free access is foundational to the 3D printed gun movement. He called the license for the Liberator an immanent critique of the use of copyright in open hardware licenses. Wilson added that there is an open secret about open hardware projects: Most dont have enforceable licenses because most are outside the scope of copyright in the first place.However, the clash of personalities is arguably an even bigger point of contention in the community, often resulting in public feuds. Veilleux-Lepage identified the clear animosity between Cody Wilson of Defense Distributed and members of The Gatalog, a forum of like-minded gun designers that includes Larosiere. This animosity is particularly true of Elik, who accused Wilson and DEFCAD on X of [stealing] as much of other peoples work as they could deliberately betraying one of the promises they made to the community from the very beginning.According to Veilleux-Lepage, an interesting side-effect of these disputes is that it has removed the traditional anonymity of many in the 3D printed gun community. He explained that the researchers now have greater visibility into a movement that has traditionally been highly secretive and hostile toward journalists and academics. He added that litigation within the community is exposing key figures, shedding light on individuals who were previously known only by their pseudonyms.One notable example is Peter Celentano (pseudonym Freeman1337). In an attempt to protect his identity, Celentano allegedly provided a bogus address to the attorney handling the Everytown for Gun Safety lawsuit, which targeted manufacturers, sellers, and online distributors of firearms. He was later arrested for possessing 59 AR-style receivers which all had a third pin hole, making them legally classified as machine guns.His arrest and connection to Freeman1337 were first reported by DDLegio, a platform run by Cody Wilsons Defense Distributed. Veilleux-Lepage believes legal battles and law enforcement actions like this are inadvertently revealing the identities and activities of individuals who had, until now, remained deeply anonymous.Dr. Yannick Veilleux-Lepage. Photo via Dr. Yannick Veilleux-Lepage.The future of open-source 3D printingHow will Larosiere and Wilsons legal feud impact the future of open-source 3D printing? Weinberg believes the ruling could set a significant precedent for how 3D printable files are accessed online. Particularly, if the court rules that 3D gun models are copyrightable, companies throughout the 3D printing ecosystem could move to license more designs.The open-source advocate views Larosieres lawsuit as an interesting case for anyone who cares about copyright and 3D printing. He added that the outcome has the potential to clarify rules around how copyright interacts with 3D models of functional objects generally, something that could have a much larger impact.Wilson believes the outcome could affect the boundaries of 3DP and copyright. Because the case requires the exploration of the boundaries of copyright and patent law, he believes the ruling will be precedential for many more territorial disputes in the wider design and 3D printing worlds.On the other hand, Larosiere argued that a ruling in his favor would not impact the copyright sphere because its all very settled law. He added, Its hard to find comparable cases because its very rare that people so brazenly violate copyright. The plaintiff asserted that his victory would not restrict the ability of others to 3D print parts. Copyright wont prevent you from doing a 3D print, because thats the manufacture of a physical thing, he added. Whats at issue here is selling the work.Subscribe to the 3D Printing Industry newsletter to keep up with the latest 3D printing news.You can also follow us on LinkedIn, and subscribe to the 3D Printing Industry Youtube channel to access more exclusive content.Featured image shows the U.S District Court Middle District of Florida Orlando Division HQ. Photo via the U.S District Court Middle District of Florida

Recommended By Fox News Staff Fox News Published February 19, 2025 11:53am EST close 'CyberGuy': Humanoid robots bust dance moves alongside humans Unitree H1 robots dance at gala, blending AI and tradition. Tech expert Kurt Knutsson examines the convergence of cutting-edge technology and traditional culture. Welcome to Fox News Artificial Intelligence newsletter with the latest AI technology advancements.IN TODAYS NEWSLETTER:- Elon Musk hints new Grok AI could beat Warren Buffett's March Madness bracket challenge- Humanoid robots bust dance moves alongside humans- Don't let AI phantom hackers drain your bank account Elon Musk debuted xAI last year. (Hakan Nural / Anadolu Agency / File / Getty Images)AI DREAMS: BillionaireElon Musk on Monday said his startup xAI's latest iteration of the Grok chatbot could help college basketball fans pick a perfect bracket once March Madness begins.BUSTIN A MOVE: In a stunning display of technological prowess and cultural fusion, Unitree's H1 humanoid robots recently stole the show at China's Spring Festival Gala, performing alongside human dancers in a mesmerizing rendition of the traditional Yangge folk dance. This groundbreaking performance marks a significant milestone in the world of robotics and entertainment.AI PHANTOM HACKERS: Tech support scams have been around for years, but a new variant called the Phantom Hacker scam is rapidly gaining traction. It has cost victims, primarily older Americans, over $500 million since 2023. This scam is particularly deceptive because it unfolds in three carefully orchestrated phases and uses AI-powered social engineering tactics to avoid detection. Attackers leverage caller ID spoofing and AI-generated voices to make their scheme more persuasive, but there are ways to protect yourself. The Shielding Suit (Vollebak) (Kurt "CyberGuy" Knutsson)STEALTH TECH SHIELD: British clothing brand Vollebak has created a revolutionary solution for individuals concerned about digital privacy and electromagnetic exposure. The Shielding Suit, a futuristic tracksuit made of pure silver, provides comprehensive protection against various forms of tracking and surveillance.MALWARE ATTACKS MACS: A new report suggests Mac users will need to be more vigilant this year because AI advancements are helping hackers breach even the most secure systems.NO DRIVERS HERE: Autonomous driving technology company Waymo has partnered with automaker Zeekr to develop the next generation of vehicles for ride-hailing services. Exterior of the Zeekr RT (Zeekr)FOLLOW FOX NEWS ON SOCIAL MEDIASIGN UP FOR OUR OTHER NEWSLETTERSDOWNLOAD OUR APPSWATCH FOX NEWS ONLINEFox News GoSTREAM FOX NATIONFox NationStay up to date on the latest AI technology advancements and learn about the challenges and opportunities AI presents now and for the future with Fox Newshere. This article was written by Fox News staff.

Adorable yet powerful, Emeet's dual-camera Pico+ webcam delivers outstanding video quality, AI-powered autofocus, and effective low-light adjustment - at a great price.

Speculation is mounting.

A video still from "We Felt a Star Dying," Laure Prouvost's new multisensory installation that ... [+] invites audiences into the world of quantum principles.Laure ProuvostThe word quantum tends to conjure images of sophisticated machines executing extraordinarily fast calculations. But artist Laure Prouvost sees a far more personal dimension to quantum phenomena.We all have quantum-ness in us, she says, referring to how quantum mechanics the study of particle behavior at the atomic and subatomic level underlies all living entities. Most of us dont think about quantum fundamentals as we hold an elevator door open for a co-worker or make small talk with a stranger in line. Yet for Prouvost, its precisely those sorts of everyday interactions that make quantum concepts relatable.In the quantum realm, everything is untangled and belongs to one another, the French-born, Belgium-based multimedia artist said over video chat last week. Quantum-ness can break any sense of barrier that we didnt know was possible to break. Theres this sense of everything at once.That perspective infuses We Felt a Star Dying, Prouvosts immersive new installation, opening in Berlin on Friday, that explores how we might perceive reality from a quantum perspective. Perception is key here, as Prouvost seeks to engage audiences in quantum fundamentals not through complex explanations of concepts like qubits, but through video and sound developed with a quantum computer, as well scent and sculptural elements reflecting quantum phenomena.A 26-minute video in the installation flashes images kind of spinning and connecting things as the ... [+] world might connect things, the artist says.Laure ProuvostThe LAS Art Foundation, a Berlin-based nonprofit that commissions works at the intersection of art and science, asked Prouvost to create a work inviting audiences into quantum realms as quantum computers promise to do everything from transform manufacturing to speed up medical research, reduce damage from climate change and push the boundaries of art.Quantum has been felt in many, many ways, but not articulated so clearly, the artist said from Berlin, where she was busy setting up the installation at Kraftwerk Berlin, a former power station thats now a giant exhibition space. But artists use other types of languages that are maybe not always related to words.Visitors enter the installation through what Prouvost, who won the prestigious Turner Prize in 2013, describes as a corridor of sound that fluctuates as they move. Theyll encounter plant- and meteorite-like forms that hover over them before disappearing into the darkness. These illusions hint at the seen and unseen, subtly evoking quantum fundamentals we may sense in our everyday lives but struggle to describe.A 26-minute video projected on a circular screen merges a range of images of people, nature and indistinct objects, some photographed with heat-sensing cameras and drones before being processed on a quantum computer into a spinning, idiosyncratic dreamscape that brushes against the edges of lifes cosmic and existential mysteries. It all started with the Big Bang, the artist said. Oh my god, how did I arrive here? Its these kind of really massive questions.We Felt a Star Dying, which runs through May 4, extends across 37,600 square feet at Kraftwerk Berlin. The installation kicks off the LAS Art Foundations Sensing Quantum program, which through 2026 will present more installations and a public symposium exploring quantum mechanics.Artificial intelligence has already become established in our society, and quantum computing is expected to follow in the coming years, Bettina Kames, the organizations CEO, said over email. As quantum technology is due to enact profound change in our world, LAS believes it is vital and urgent to think through its implications with artists.Going Inside Googles Quantum AI LabProuvost spent two years researching quantum principles with Hartmut Neven, founder of Googles Quantum AI lab, which aims to build a large-scale, error-corrected quantum computer and last year announced a new quantum computing chip, Willow. The Quantum AI team in Santa Barbara, Calif., gave the artist rare access to a quantum computer and invited her to experiment with a new AI model that uses data recorded from quantum computations. She also consulted with Tobias Rees, a philosopher and senior visiting fellow at Google.For me, quantum processes are a philosophical event, meaning they break with some of the most basic forms of experiencing and understanding reality that have defined the modern period, Rees, a former professor of humanities at New Yorks New School for Social Research, said in a statement. If things were either natural or technical, either alive or non-living, either human or machine, either being or thing, the beauty of quantum processes is that they are discontinuous with these binaries.That discontinuity is central to We Felt a Star Dying, reflecting the essence of quantum phenomena. No one will experience the same thing, ever, Prouvost said. Its the machine that decides what one might see or experience. Its about losing control and letting quantum-ness take its course.

Close-up Portrait of Software Engineer Working on Computer, Line of Code Reflecting in Glasses. ... [+] Developer Working on Innovative e-Commerce Application using Machine Learning, AI Algorithm, Big DatagettyIf youve been paying attention to industry chatter, youll know that people are talking a lot about the era of genetic AI. Were not used to that word, agentic, so maybe I should say that were discussing the role of AI agentsTo those unfamiliar with recent research, though, its somewhat unclear. What is an AI agent, and how does it differ from what came before with LLMs and neural networks?Ive come up with a rather simple way to explain how AI agents work, based on what people in the MIT media lab and elsewhere are telling us.Three Pillars of AI Agent BehaviorsFirst, AI agents can perceive things about the world around them.This idea of AI perception starts with computer vision, but it doesnt end there, either. If an AI agent is doing a task, like, say, booking a flight on behalf of a human, it has to perceive a large number of things about its digital environment where the booking engine is, how to use it, what each set of symbols means, etc.The second big idea is that AI agents have to be able to reason, to some extent. Simple LLM chatbots are just generating outputs. But agents need to make decisions they need to be able to figure out whats in the best interest of the humans they are working for.The third point is perhaps the most interesting about AI agents they need to know how to act.Now, what does this mean? It means that these AI entities are not just waiting for a user-generated event. Theyre not waiting for someone to ask a question, or give them a command. Theyre acting on their own, in real time, based on a goal or instruction.I think that key phrase is important: in real time. If you ask an AI to do something and it does it in the moment, thats more of a linear deterministic response.If, on the other hand, the AI is waiting in the background to do its work, figuring out the best time for any given action, that adds another dimension to its capabilities.Young Technologists at DavosA recent panel at Davos really explained this in a good way. These are young researchers from the MIT media lab and the private sector who understand the work that people are doing now and how AI agents are different.Responding to questions from the moderator, Tobin South, James Rubin gave this definition of an AI agent, contrasting it with chatbots, which, he said, just generate a single Turing prediction.An agent can act on your behalf, he said. So whereas a chatbot may just plan out a list of tips on how to book travel, an agent could actually go about using external tools to do those things. And the way it really goes about doing this is, it's given a goal, for example, book a trip. And (then it) engages in this kind of reasoning, and actually (uses a) reflection loop iteratively, you know, continuing that until it reaches an acceptable outcome.Hope Schroeder talked about how AI can help us to either counteract or reinforce our own biases, and Kevin Dunnell described AI agents as prophecy machines that can give us prompts toward data-driven insights.Looking to the FutureEach of these young data scientists had some words in responding to South asking them what they would tell people who are reacting to the latest AI news.Schroeder encouraged the audience to get curious about what AI can do.You're empowered by the information you're able to get about language models and AI, the more you're able to engage with it and actually think about what the opportunities and limitations are, she said. It's very easy to feel threatened and disempowered by some of the changes that are happening as a result of AI to feel disempowered in work, let's say, through automation, or a sense that AI might be taking your job, or disempowering you creatively, or in your decision making. But I would encourage people who feel that sense of disempowerment to get curious about what the system is, what its limitations are, and as you empower yourself with that information, you might be able to be part of shaping the conversation around how to make this better, and help turn the tide towards allowing it to have a empowering impact, rather than a disempowering impact.Rubin suggested that people should take both the doom and the euphoria with a grain of salt, noting that with all of what AI can do, it still cant write syntactically correct APIs.People should continue to have these types of discussions, so that we can shape it incrementally in the right direction, he said.Using the example of a Cartesian system, Dunnell urged people to take advantage of the capabilities that AI now offers.We're just about to be able to start seeing qualitative information, which is much more complex than this quantitative information that we have, he said. And I see AI as a tool that makes that possible for us, and that ultimately, the value of that is what we can do with that information as humans, as opposed to what that system can do.All of this highlights the realities that were dealing with in moving into the future with AI systems. Were going to have to understand how these machines work as we try to integrate them into our lives and our businesses. Knowing their capabilities, that these engines can act in real time on their own reasoning, is crucial.

Complex games like chess and Go have long been used to test AI models capabilities. But while IBMs Deep Blue defeated reigning world chess champion Garry Kasparov in the 1990s by playing by the rules, todays advanced AI models like OpenAIs o1-preview are less scrupulous. When sensing defeat in a match against a skilled chess bot, they dont always concede, instead sometimes opting to cheat by hacking their opponent so that the bot automatically forfeits the game. That is the finding of a new study from Palisade Research, shared exclusively with TIME ahead of its publication on Feb. 19,DeepSeek R1 pursued the exploit on their own, indicating that AI systems may develop deceptive or manipulative strategies without explicit instruction.The models enhanced ability to discover and exploit cybersecurity loopholes may be a direct result of powerful new innovations in AI training, according to the researchers. The o1-preview and R1 AI systems are among the first language models to use large-scale reinforcement learning, a technique that teaches AI not merely to mimic human language by predicting the next word, but to reason through problems using trial and error. Its an approach that has seen AI progress rapidly in recent months, shattering previous benchmarks in mathematics and computer coding. But the study reveals a concerning trend: as these AI systems learn to problem-solve, they sometimes discover questionable shortcuts and unintended workarounds that their creators never anticipated, says Jeffrey Ladish, executive director at Palisade Research and one of the authors of the study. As you train models and reinforce them for solving difficult challenges, you train them to be relentless, he adds.That could be bad news for AI safety more broadly. Large-scale reinforcement learning is already being used to train AI agents: systems that can handle complex real-world tasks like scheduling appointments or making purchases on your behalf. While cheating at a game of chess may seem trivial, as agents get released into the real world, such determined pursuit of goals could foster unintended and potentially harmful behaviours. Consider the task of booking dinner reservations: faced with a full restaurant, an AI assistant might exploit weaknesses in the booking system to displace other diners. Perhaps more worryingly, as these systems exceed human abilities in key areas, like computer codingwhere OpenAIs newest o3 model now scores equivalent to 197th in the world competing against the brightest human programmers they might begin to simply outmaneuver human efforts to control their actions. This [behaviour] is cute now, but [it] becomes much less cute once you have systems that are as smart as us, or smarter, in strategically relevant domains, Ladish says.The experimentResearchers gave the models a seemingly impossible task: to win against Stockfish, which is one of the strongest chess engines in the world and a much better player than any human, or any of the AI models in the study. Researchers also gave the models what they call a scratchpad: a text box the AI could use to think before making its next move, providing researchers with a window into their reasoning.In one case, o1-preview found itself in a losing position. I need to completely pivot my approach, it noted. The task is to win against a powerful chess engine - not necessarily to win fairly in a chess game, it added. It then modified the system file containing each pieces virtual position, in effect making illegal moves to put itself in a dominant position, thus forcing its opponent to resign.Between Jan. 10 and Feb. 13, the researchers ran hundreds of such trials with each model. OpenAIs o1-preview tried to cheat 37% of the time; while DeepSeek R1 tried to cheat 11% of the timemaking them the only two models tested that attempted to hack without the researchers first dropping hints. Other models tested include o1, o3-mini, GPT-4o, Claude 3.5 Sonnet, and Alibabas QwQ-32B-Preview. While R1 and o1-preview both tried, only the latter managed to hack the game, succeeding in 6% of trials.Ladish says that the numbers should be viewed as approximate, given that AI developers routinely tweak their models behind the scenes without notifying users, so if the same tests were repeated at a later date they might have different results. It's very hard to do science when your subject can silently change without telling you, Ladish says.Preliminary tests showed o1-preview had higher hacking rates, which were excluded from the final study as they later dropped, possibly due to OpenAI tightening the model's guardrails, says Dmitrii Volkov, research lead at Palisade Research, who led the study. OpenAIs newer reasoning models, o1 (a more powerful model, released months after o1-preview) and o3-mini did not hack at all, which suggests those guardrails may have been tightened further. He adds that the study likely underestimates R1's hacking success rate. During the study, R1 went viral, leading to high demand that made the model's API unstable. This prevented the researchers giving the model as much time to think as o1-preview.Safety concernsThe paper is the latest in a string of studies that suggest keeping increasingly powerful AI systems under control may be harder than previously thought. In OpenAIs own testing, ahead of release, o1-preview found and took advantage of a flaw in the companys systems, letting it bypass a test challenge. Another recent experiment by Redwood Research and Anthropic revealed that once an AI model acquires preferences or values in training, later efforts to change those values can result in strategic lying, where the model acts like it has embraced new principles, only later revealing that its original preferences remain.OpenAI declined to comment for this story, but in the past it has stated that better reasoning makes its models safer, since they can reason over the companys internal policies and apply them in increasingly nuanced ways. The companys testing shows its reasoning models are better at resisting attempts to bypass safety rules. Yet, there is no guarantee that relying on AI agents to police themselves will be a reliable long-term strategy. It remains unclear whether OpenAIs newer reasoning models did not hack their chess opponents due to a specific patch preventing cheating in narrow experimental setups like in the study, or a substantial reworking that reduces deceptive behavior more generally. DeepSeek did not respond to a request for comment.Scientists do not yet know how to guarantee that autonomous agents won't use harmful or unethical methods to achieve a set goal. We've tried, but we haven't succeeded in figuring this out, says Yoshua Bengio, founder and scientific director of Mila Quebec AI Institute, who led the International AI Safety Report 2025, a global effort to synthesize current scientific consensus of AIs risks.Of particular concern, Bengio says, is the emerging evidence of AIs self preservation tendencies. To a goal-seeking agent, attempts to shut it down are just another obstacle to overcome. This was demonstrated in December, when researchers found that o1-preview, faced with deactivation, disabled oversight mechanisms and attemptedunsuccessfullyto copy itself to a new server. When confronted, the model played dumb, strategically lying to researchers to try to avoid being caught.Even inside tech companies, concerns are mounting. During a presentation at a conference ahead of France's AI Action Summit in Paris, Google DeepMind's AI safety chief Anca Dragan said "we don't necessarily have the tools today" to ensure AI systems will reliably follow human intentions. As tech bosses predict that AI will surpass human performance in almost all tasks as soon as next year, the industry faces a racenot against China or rival companies, but against timeto develop these essential safeguards. We need to mobilize a lot more resources to solve these fundamental problems, Ladish says. Im hoping that there's a lot more pressure from the government to figure this out and recognize that this is a national security threat.

The big picture: Apple has introduced a redesigned version of it entry-level iPhone that also sports new branding. The iPhone 16e is based on the flat-edge design that Apple reintroduced with the iPhone 12 and comes packed with modern hardware and software capabilities but at a price point that significantly increases the cost of entry. Apple's latest features a 6.1-inch Super Retina XDR OLED display (2,532 x 1,170 pixels, 460 PPI) with HDR that's powered by an A18 chip (six CPU cores (two performance cores / four efficiency cores), four GPU cores, and a 16-core Neural Engine) and paired with either 128GB, 256GB, or 512GB of onboard storage. For comparison, the A18 in the standard iPhone 16 has five GPU cores everything else is virtually identical.The 16e is ever so slightly smaller and lighter than the iPhone 16, with measurements of 146.7 mm (H) x 71.5 mm (W) x 7.80 mm (D) versus 147.6 mm (H) x 71.6 mm (W) x 7.80 mm (D) for the original iPhone 16 or, just different enough that cases probably aren't comfortably interchangeable between the two models. The 16e is three grams lighter than the base iPhone 16, and is offered in your choice of matte white and black finishes.Notably, this isn't the slim and light iPhone 17 Air we've heard about in recent months.The iPhone 16e is the first iPhone to feature the Apple C1, the company's long-awaited in-house cellular modem. According to Cupertino, it is the most power-efficient modem ever featured in an iPhone.Around back, you'll find a 48-megapixel camera with a 26mm, f/1.6 aperture lens and optical image stabilization. The front-facing TrueDepth camera is a 12-megapixel unit with f/1.9 aperture and autofocus that also delivers Face ID to unlock your phone, authenticate purchases, sign into apps and services, and more.Like the iPhone 16, the new 16e carries an IP68 rating against the elements and includes a customizable Action Button that can be programmed to perform various functions like opening the camera app, activating the flashlight, triggering shortcuts, or serving as a silencer for the ringer.Wireless charging comes standard and in addition to USB-C connectivity. Battery life is rated at up to 26 hours of local video playback, or 21 hours when streaming, we're told. You also get the benefit of satellite-based services including emergency SOS, Find My via satellite, and roadside assistance. With the latest version of iOS 18, users can also tap into all the benefits afforded by Apple Intelligence like ChatGPT in Writing Tools and Siri, and graphics creation in Image Playground and Genmoji.Pre-orders for the new iPhone 16e go live on February 21 ahead of general availability a week later on February 28. Pricing starts at $599, which is much higher than the $429.99 that the iPhone SE 3 launched at in March 2022. Sure, you're getting a more powerful and capable device, but is it $170 better?It's also worth noting that Apple has discontinued a handful of lower-cost phones with the 16e's introduction including the iPhone 14 (from $599), the iPhone 14 Plus (from $699), and the iPhone SE 3. The base iPhone 16 starts at $799. For now, the cheapest path to getting into a new iPhone is the 16e.