0 Comentários
0 Compartilhamentos
Diretório
Diretório
-
Faça Login para curtir, compartilhar e comentar!
-
WWW.TECHRADAR.COMWho is ISoftStone? Mysterious tech firm set to become one of the biggest PC vendors in China, beating Huawei, HP and ApplePlucky regional player ISoftStone has recorded astounding recent growth.0 Comentários 0 Compartilhamentos
-
WWW.TECHRADAR.COMGrok's mobile app is here and it might not be very carefulxAI releases Grok mobile app for iOS.0 Comentários 0 Compartilhamentos
-
WWW.FASTCOMPANY.COMStarbucks strike will expand to over 300 U.S. stores on Christmas EveA strike at Starbucks will expand to over 300 U.S. stores on Tuesday, with more than 5,000 workers expected to walk off the job before the five-day work stoppage ends later on Christmas Eve, the workers union said.Starbucks Workers United, representing employees at 525 stores nationwide, said more than 60 U.S. stores across 12 major cities, including New York, Los Angeles, Boston and Seattle, were shut on Monday.Talks between Starbucks and the union had hit an impasse with unresolved issues over wages, staffing and schedules, leading to the strike that began on Friday.With a union contract, improved benefits, wages, and working conditions could be in writing and cant be reduced without bargaining, according to the Workers United website.The Christmas Eve strike on Tuesday was projected to be the largest ever at the coffee chain, the union added. These strikes are an initial show of strength, and were just getting started, an Oregon barista said in a union statement.When asked for a response, a Starbucks spokesperson referred to a company statement it released on Monday.It said that 97% to 99% of Starbucks stores will continue to operate and serve customers, adding that it expects a very limited impact to overall operations. Starbucks has over 10,000 company-operated stores across the U.S.We are ready to continue negotiations when the union comes back to the bargaining table, the company said.The Seattle-headquartered firm had previously claimed that the union delegates prematurely ended the bargaining session.Earlier this month, the workers group rejected an offer of no immediate wage raises and a guarantee of a 1.5% pay increase in future years.The union also said that Starbucks has yet to present its workers with a serious economic proposal.Bipasha Dey, Shubham Kalia, Gursimran Kaur and Savyata Mishra, Reuters0 Comentários 0 Compartilhamentos
-
WWW.YANKODESIGN.COMBenedict Radcliffe crafts a limited edition glowing 3D sculpture of the iconic Porsche 959Benedict Radcliffe is an established name in the European design circles attributing to his off-beat 3D geometric wireframe sculptures. The designer has crafted life-sized versions of iconic vehicles and his acclaimed creations that transform the environment around them.His latest creation caught my eye, because you already know Im a Porsche fan. Yes, this is the Porsche 959 recreated in a scaled version with a neon green glowing steel finish. This revered model from the German automaker is celebrating its 40th anniversary in 2025, and Benedict has got his timing right on point. Motorheads who cant get their hands on the real thing can sport this intricate sculpture on their work desk or living room shelf with pride.Designer: Benedict Radcliffe and Studio INKThis desktop-worthy sculpture is made out of composite resin and has a cool glow to it that will make you stick to your office chair for longer than usual. The 1:15 scale model is limited to just 15 pieces, and can be purchased from the artists official website for 2500 GBP. The wireframe model grows on the success of the already popular interpretations of the Ferrari F40, Porsche 934, and the classic London taxi that got exposure at Heathrow Airport in 2014.The London-based artist has put together the Porsche 959 and collaborated with studio INK who have made an animated teaser of this glowing steel tube sculpture in the digital world. Interested Porsche fans can view the car in a 360-degree world, since the studio has taken 3D scans of the designers creation. The captivating animation shows the assembly process and then heads to the final form of the fluorescent sculpture. Coming onto the history of the Porsche 959, well, the car debuted in 1985 and quickly grabbed all the attention for being the worlds fastest street-legal production car at that time. The sportscar could toggle the torque between the rear and front wheels in any kind of weather conditions.The post Benedict Radcliffe crafts a limited edition glowing 3D sculpture of the iconic Porsche 959 first appeared on Yanko Design.0 Comentários 0 Compartilhamentos
-
APPLEINSIDER.COMMerry Christmas, from the entire AppleInsider teamAll of us at AppleInsider wish you a happy and peaceful holiday season, and thank you for reading us throughout a roller-coaster 2024.Jony Ive helped design the decorations for this tree, displayed in London's Claridge's store in 2018Every year for the last 25 has been a big one for Apple, and you'd expect it for a company of its size. It launched the Apple Vision Pro, it joined the AI arms race with Apple Intelligence, and even found some time to release new products like the iPhone 16.With the end of the year fast approaching, the Apple News cycle is temporarily winding down, but AppleInsider won't be taking a break. Throughout the holidays, we will be with you, sharing the latest news stories and using our expertise to discuss how the news of the day impacts our favorite company, as it happens. Continue Reading on AppleInsider | Discuss on our Forums0 Comentários 0 Compartilhamentos
-
APPLEINSIDER.COMEnding Google search partnership would hamstring Apple, says Eddy CueIn a declaration filed on Monday, Apple's senior vice president of services, Eddy Cue, asserted that Apple has no interest in creating its own search engine and would prefer to continue to use Google.In 2022, Google paid Apple $20 billion to be the default search engine on its devices. The result was a boon for both Google and Apple, with Apple receiving about 36% of the ad revenue generated from Safari search result advertising.The declaration, spotted by Reuters, is Apple's attempt at defending Google. Cue's stance is the same one that the company has defended for quite some time now.The declaration lists several reasons Apple is satisfied with its deal with Google. It also explains why the Cupertino tech giant has no interest in creating its own search engine. Continue Reading on AppleInsider | Discuss on our Forums0 Comentários 0 Compartilhamentos
-
GAMINGBOLT.COMHelldivers 2, Space Marine 2, Baldurs Gate 3 Were Among the Bestselling Games on Steam in 2024Valve has released its annual year-end round-ups, and in the top 100 sellers list for 2024 based on their gross revenue (including microtransactions), the Platinum category which groups together the top 10 games in random order list a number of major new titles that released this year. Namely, those games areHelldivers 2, Warhammer 40,000: Space Marine 2, Palworld, Call of Duty: Black Ops 6,andBlack Myth: Wukong.Baldurs Gate 3is also included in the Platinum category, as isElden Ring,no doubt driven by the release of its expansion,Shadow of the Erdtree. Meanwhile, free-to-play megatons likeApex Legends, Dota 2,andCounter-Strike 2are also featured.Several other new 2024 releases are included across multiple categories in the top 100, with Gold (13th-24th) includingEA Sports FC 25, Path of Exile 2,andDragon Ball: Sparking! ZERO, Silver (25th-50th) includingS.T.A.L.K.E.R. 2: Heart of Chornobyl, Granblue Fantasy: Relink, Tekken 8,andDragons Dogma 2,and Bronze (51st-100th) including the likes ofPersona 3 Reload, Metaphor: ReFantazio, Dragon Age: The Veilguard, Silent Hill 2, Balatro, andHades 2.Interestingly,Monster Hunter Wilds, which is releasing in February, is also included in the Bronze category based on its pre-order sales alone, as isMonster Hunter Worlds.Fallout 4can be found in the same category, though curiously,Starfieldis nowhere to be found in the top 100. Meanwhile, some notable 2024 releases are also surprisingly missing, such asLike a Dragon: Infinite Wealth.You can check out the full top 100 sellers list below.PLATINUM:Apex LegendsCounter-Strike 2Helldivers 2Warhammer 40,000: Space Marine 2Baldurs Gate 3PalworldCall of Duty: Black Ops 6Black Myth: WukongDota 2Destiny 2Elden RingPUBG BattlegroundsGOLD:EA Sports FC 25The First DescendantEnshroudedPath of Exile 2EA Sports FC 24Warframe 1999Dead by DaylightCyberpunk 2077Naraka: BladepointWar ThunderThrone and LibertyDragon Ball: Sparking! ZEROSILVER:Forza Horizon 5S.T.A.L.K.E.R. 2: Heart of ChornobylFallout 76Manor LordsHogwarts LegacyGhost of TsushimaThe Elder Scrolls OnlineYu-Gi-Oh! Master DuelGranblue Fantasy: RelinkFinal Fantasy 14Lethal CompanyGrand Theft Auto 5Rainbow Six SiegeStardew ValleyOnce HumanCall of Duty: WarzoneLast EpochRustCall of Duty: Modern Warfare 3Dragons Dogma 2Gray Zone WarfareDiablo 4Red Dead Redemption 2Lost ArkTekken 8Forza Horizon 4BRONZE:Fallout 4Path of ExileSilent Hill 2It Takes TwoTeam Fortress 2Crusader Kings 3American Struck SimulatorMonster Hunter WorldNBA 2K25Black DesertFarming Simulator 25BalatroThe FinalsV RisingBattlefield 2042Dragon Age: The VeilguardNo Mans SkyArk: Survival AscendedTCG Card Shop SimulatorPhasmophobiaRimworldStreet Fighter 6Sons of the ForestHorizon Forbidden WestSekiro: Shadows Die TwiceSatisfactoryWallpaper EngineHades 2Ready or NotStellarisFactorioHearts of Iron 4Overwatch 2The Sims 4Metaphor: ReFantazioBodycamSupermarket SimulatorContent WarningTotal War: Warhammer 3Liars BarChained TogetherWorld of WarshipsSea of ThievesMonster Hunter WildsEurotruck Simulator 2Persona 3 ReloadCivilization 6Hunt: Showdown 1896Project Zomboid0 Comentários 0 Compartilhamentos
-
VENTUREBEAT.COMOpenAIs o3 shows remarkable progress on ARC-AGI, sparking debate on AI reasoningo3 solved one of the most difficult AI challenges, scoring 75.7% on the ARC-AGI benchmark. But does it really mean we're closer to AGI?Read More0 Comentários 0 Compartilhamentos
-
WWW.MARKTECHPOST.COMSalesforce AI Research Introduces AGUVIS: A Unified Pure Vision Framework Transforming Autonomous GUI Interaction Across PlatformsGraphical User Interfaces (GUIs) play a fundamental role in human-computer interaction, providing the medium through which users accomplish tasks across web, desktop, and mobile platforms. Automation in this field is transformative, potentially drastically improving productivity and enabling seamless task execution without requiring manual intervention. Autonomous agents capable of understanding and interacting with GUIs could revolutionize workflows, particularly in repetitive or complex task settings. However, GUIs inherent complexity and variability across platforms pose significant challenges. Each platform uses distinct visual layouts, action spaces, and interaction logic, making creating scalable and robust solutions difficult. Developing systems that can navigate these environments autonomously while generalizing across platforms remains an ongoing challenge for researchers in this domain.There are many technical hurdles in GUI automation right now; one is aligning natural language instructions with the diverse visual representations of GUIs. Traditional methods often rely on textual representations, such as HTML or accessibility trees, to model GUI elements. These approaches are limited because GUIs are inherently visual, and textual abstractions fail to capture the nuances of visual design. In addition, textual representations vary between platforms, leading to fragmented data and inconsistent performance. This mismatch between the visual nature of GUIs and the textual inputs used in automation systems results in reduced scalability, longer inference times, and limited generalization. Also, most current methods are incapable of effective multimodal reasoning and grounding, which are essential for understanding complex visual environments.Existing tools and techniques have attempted to address these challenges with mixed success. Many systems depend on closed-source models to enhance reasoning and planning capabilities. These models often use natural language communication to combine grounding and reasoning processes, but this approach introduces information loss and lacks scalability. Another common limitation is the fragmented nature of training datasets, which fail to provide comprehensive support for grounding and reasoning tasks. For instance, datasets typically emphasize either grounding or reasoning, but not both, leading to models that excel in one area while struggling in others. This division hampers the development of unified solutions for autonomous GUI interaction.The University of Hong Kong researchers and Salesforce Research introduced AGUVIS (7B and 72B), a unified framework designed to overcome these limitations by leveraging pure vision-based observations. AGUVIS eliminates the reliance on textual representations and instead focuses on image-based inputs, aligning the models structure with the visual nature of GUIs. The framework includes a consistent action space across platforms, facilitating cross-platform generalization. AGUVIS integrates explicit planning and multimodal reasoning to navigate complex digital environments. The researchers constructed a large-scale dataset of GUI agent trajectories, which was used to train AGUVIS in a two-stage process. The frameworks modular architecture, which includes a pluggable action system, allows for seamless adaptation to new environments and tasks.The AGUVIS framework employs a two-stage training paradigm to equip the model with grounding and reasoning capabilities:During the first stage, the model focuses on grounding and mapping natural language instructions to visual elements within GUI environments. This stage utilizes a grounding packing strategy, bundling multiple instruction-action pairs into a single GUI screenshot. This method improves training efficiency by maximizing the utility of each image without sacrificing accuracy.The second stage introduces planning and reasoning, training the model to execute multi-step tasks across various platforms and scenarios. This stage incorporates detailed inner monologues, which include observation descriptions, thoughts, and low-level action instructions. By progressively increasing the complexity of training data, the model learns to handle nuanced tasks with precision and adaptability.AGUVIS demonstrated great results in both offline and real-world online evaluations. In GUI grounding, the model achieved an average accuracy of 89.2, surpassing state-of-the-art methods across mobile, desktop, and web platforms. In online scenarios, AGUVIS outperformed competing models with a 51.9% improvement in step success rate during offline planning tasks. Also, the model achieved a 93% reduction in inference costs compared to GPT-4o. By focusing on visual observations and integrating a unified action space, AGUVIS sets a new benchmark for GUI automation, making it the first fully autonomous pure vision-based agent capable of completing real-world tasks without reliance on closed-source models.Key takeaways from the research on AGUVIS in the field of GUI automation:AGUVIS uses image-based inputs, reducing token costs significantly and aligning the model with the inherently visual nature of GUIs. This approach results in a token cost of only 1,200 for 720p image observations, compared to 6,000 for accessibility trees and 4,000 for HTML-based observations.The model combines grounding and planning stages, enabling it to perform single- and multi-step tasks effectively. The grounding training alone equips the model to process multiple instructions within a single image, while the reasoning stage enhances its ability to execute complex workflows.The AGUVIS Collection unifies and augments existing datasets with synthetic data to support multimodal reasoning and grounding. This results in a diverse and scalable dataset, enabling the training of robust and adaptable models.Using pyautogui commands and a pluggable action system allows the model to generalize across platforms while accommodating platform-specific actions, such as swiping on mobile devices.AGUVIS achieved remarkable results in GUI grounding benchmarks, with accuracy rates of 88.3% on web platforms, 85.7% on mobile, and 81.8% on desktops. Also, it demonstrated superior efficiency, reducing USD inference costs by 93% compared to existing models.In conclusion, the AGUVIS framework addresses critical challenges in grounding, reasoning, and generalization in GUI automation. Its purely vision-based approach eliminates the inefficiencies associated with textual representations, while its unified action space enables seamless interaction across diverse platforms. The research provides a robust solution for autonomous GUI tasks, with applications ranging from productivity tools to advanced AI systems.Check out the Paper, GitHub Page, and Project. All credit for this research goes to the researchers of this project. Also,dont forget to follow us onTwitter and join ourTelegram Channel andLinkedIn Group. Dont Forget to join our60k+ ML SubReddit. Asif RazzaqAsif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences. [Download] Evaluation of Large Language Model Vulnerabilities Report (Promoted)0 Comentários 0 Compartilhamentos