Overclocker shows RTX Pro 6000 outperforms Nvidia’s best consumer GPUs - but at five times the price!

AMD has acquired AI software startup Brium, in a move potentially aimed at challenging Nvidia’s dominance in AI software and strengthening support for machine learning workloads on AMD hardware. “Brium brings advanced softwa

Stone PC Case, Cooler Master GPU, DIY Case from Scratch, and Metal FansJune 4, 2025Last Updated: 2025-06-04Cooler Master is doing some really interesting stuff with its new casesThe HighlightsCooler Master’s upcoming MF600, MF500, and MF400 reconfigurable frame cases are assembled from columns and cornersThe company also showed off interesting stone facade case front panelsCooler Master is working on a “GPU” with AsusTable of ContentsAutoTOC Grab a GN Tear-Down Toolkit to support our AD-FREE reviews and IN-DEPTH testing while also getting a high-quality, highly portable 10-piece toolkit that was custom designed for use with video cards for repasting and water block installation. Includes a portable roll bag, hook hangers for pegboards, a storage compartment, and instructional GPU disassembly cards.IntroWe visited Cooler Master’s booth at Computex 2025 where the company showed off several new cases. Arguably the most interesting one is a modular case. It comes with, we believe, 8 corners and 12 columns.Editor's note: This was originally published on May 20, 2025 as a video. This content has been adapted to written format for this article and is unchanged from the original publication.CreditsHostSteve BurkeCamera, Video EditingMike GaglioneVitalii MakhnovetsWriting, Web EditingJimmy ThangCooler Master MF CasesThe case comes with a front panel that has a dust filter in it. With it all assembled, it looks like the cases in the image above. The cases are the MF series, with the largest one being the MF600, which we assume translates to “Motherf***ing 600.” There’s also the MF500 and the smaller MF400. Initially, Cooler Master is basically going to be selling pre-configured models. Eventually, the company wants to allow people to customize the case on their site and have it assembled and shipped from around the City of Industry. It’s pretty cool as it’s a fully modular approach.The side panels are secured to the case via magnets, which is actually a nice touch. Internally, the MF600 we saw came with 3x140mm fans on the front and 1x120mm fan on the back. The motherboard tray is pretty standard for the most part. Exceptions include a rail system that provides numerous holes for screws to go in, which allows Cooler Master to reconfigure things. Inside the case towards the back, there’s also a rail system, which forms bits and pieces of the motherboard tray that allow for more customizability. Cooler Master has been kind of on-and-off in the DIY space over the years where they’ve had some really big wins and some really big losses. They were also kind of absent for a while, but these MF cases represent a better showing from what we’ve seen in a while from the company. According to Cooler Master, a pre-configured MF600 is supposed to cost $200. We expect to test and review the case. The MF500 is supposed to go for $165 and includes 2x200mm fans in the front and 1x120mm fan in the back. The smallest MF case, the MF600, which is a very large micro ATX box, is going for $150. In terms of fans, it has 2x120mm ones at the bottom coupled with a 1x120mm fan in the rear.  Cooler Master also showed off different panel types they’re experimenting with. One of them included a facade-style stone. One of the pre-built MF cases we looked at had stuff flipped around in an inverted layout. One of the benefits of its rail system allows the case to have a bar that screws in which can support the GPU. Looking into this system, you can see that the PSU is at the bottom next to a bottom intake fan. Updated Cosmos Visit our Patreon page to contribute a few dollars toward this website's operation (or consider a direct donation or buying something from our GN Store!) Additionally, when you purchase through links to retailers on our site, we may earn a small affiliate commission.Cooler Master’s updated Cosmos has the NVIDIA-like DGX style front. We also saw a variant of the Cosmos with thermal baffles in it. We have some criticisms of its execution, but overall, it’s an interesting idea.  The way the baffles are designed, Cooler Master is trying to bring air straight in through its channels. There’s a channel for the CPU that exposes the fin stack and Cooler Master's V8 CPU cooler. It conveys an idea similar to an engine cover. The GPU has a separate baffle beneath the CPU one. The company is trying to isolate air flow. In theory, this should work well and we would love this idea applied to more affordable cases, like the MF series, especially since they’re already kind of configurable. Looking at the back, fans can be mounted on the rear, which can help pull air out. We also saw another variant of the Cosmos case running liquid cooling with a distro block. It was coupled with 4x180mm fans and a “720” radiator, which pulled air into the case. Unfortunately, the air is blowing straight into the wall of a motherboard tray, but Cooler Master says the plan is to pull the air up and out of the case with additional 180mm fans on the top and to move the PSU towards the bottom of the case. Looking closer at the front of the special edition of the Cosmos cases, we can see the NVIDIA DGX shroud, which Cooler Master manufactures. It’s essentially like a sponge-like mesh. The special edition of the Cosmos doesn’t have a price yet, but the non-special edition variant is supposed to be around $400, which is before any potential tariffs. Cooler Master CoolersCooler Master showed off some CPU air coolers that had some 3D heat pipes, which had more heat pipes protruding from the center. The company also showed off its V8 cooler and a full-metal fan. The fan’s blades and frame are both aluminum.  Cooler Master Elite Series CasesCooler Master does some really cool sh*t but has a branding problem. For instance, the company’s “Elite” series cases, shown in the image above, are actually budget cases. From left to right, we believe they are called the Elite 482 ($50), Elite 600 ($65), Elite 490 Wood ($50), Elite 691 Wood ($60), Elite 693 ($60), Elite 692 ($70), Elite 302 ($40), and Elite 502 ($60). Our advice to Cooler Master here is for them to unf*ck these names.Most of the Elite series cases don’t come with fans with the exception of the Elite 302 and Elite 502, which come with 3 ARGB fans. MF360Next up are Cooler Master’s MF360 cases, which conveys that you can see inside the case from all sides. While it’s going to have some thermal challenges, to give the company credit, it’s actually really good looking. The MF360 is a showcase fish-tank style PC that you can see through from both sides. Inside the case, we saw a distro block and tubes routed through on both sides.Cooling XThe case in the image above, which goes by "Cooling X,” and uses the company’s new MF frame system. If you look at the corner, you can see the individual columns. At Computex, we saw it as a pre-built system.The top of the case has a magnetically attached panel, which just pulls right off. The panel itself provides really good porosity and the material is pretty nice. Removing the top panel exposes 2 offset fans. The back fan tries to pull in air with the front fan trying to exhaust air out of the top, which is why they’re offset. That’s kind of cool to see.  Cooler Master FansCooler Master showed off all-aluminum fans, which include the blades and frame. The MF120 XT is a 120mm model, is supposed to be $35, and the company says it goes up to 4,000 RPM. The fan’s RPM can also be button-controlled via an external remote and it uses a dual-ball-bearing solution. Cooler Master’s mixed fans, which use plastic blades coupled with an aluminum frame, come with fluid dynamic bearings (FDBs). The clearance between the fan blade tip and the frame is important as the smaller that clearance is, the better performance you get. The major downside is that as the fan ages, it can start to clip the interior of the frame. Having it too close can also negatively impact yields. The solution to this is LCP, which is incredibly expensive, or metal, because it doesn’t deform, but that’s also expensive. Right now, Cooler Master says it’s about a .8mm distance, which is pretty good. The company is targeting 0.6mm by the time the fan launches. Cooler Master Video Card Shroud Grab a GN15 Large Anti-Static Modmat to celebrate our 15th Anniversary and for a high-quality PC building work surface. The Modmat features useful PC building diagrams and is anti-static conductive. Purchases directly fund our work! (or consider a direct donation or a Patreon contribution!)Cooler Master also showed off some video cards, which is not something the company is typically involved with. Cooler Master created a GPU shroud with adjustable slats that can accommodate 15-30mm fans. This solution is geared towards pre-built PCs and isn’t planned to be sold separately.Examining one of the fans, we saw a standard 25mm-thick fan, which Cooler Master’s GPU shroud solution can adjust to via different notch options.Cooler Master is also using a vapor chamber, which is supported by 8x8mm heat pipes running through the shroud and a gigantic fin stack. In total, it weighs almost 7 pounds (3.2 kilograms).Cooler Master claims that, in terms of cooling, it performs similar to the 4-fan Astral solution at lower noise levels, but we don’t have those numbers. With 4,000 RPM fans running on a 600-watt heat load, Cooler Master claims a 5090 will run at about 49 degrees C or so for the GPU.

Over eight years after its predecessor launched and many rumors later, the Nintendo Switch 2 is finally available worldwide. It retails for $449.99 with a bundle available for $499.99 that includes Mario Kart World. Check out the launch trailer below. The Switch 2 offers several major improvements over the original, including a larger screen with HDR support. It can run in 120 frames per second in Handheld Mode and output to 4K resolution when docked. The Joy-Con 2 has also been revamped, sporting magnetic locks and mouse support. Other notable features include upscaling via Nvidia DLSS, though sadly, VRR support isn’t available in Docked Mode. Launch titles include Mario Kart World at a whopping $79.99, Nintendo Switch 2 Welcome Tour for $9.99, and several third-party games like Yakuza 0 Director’s Cut, Cyberpunk 2077: Ultimate Edition, Street Fighter 6 Years 1-2 Fighters Edition, and much more. Titles like The Legend of Zelda: Breath of the Wild and Tears of the Kingdom have also received Switch 2 Editions, with paid upgrades available for current owners. Of course, several existing titles have received free updates to help them perform better on the new console, including Super Mario Odyssey, Super Mario 3D All-Stars + Bowser’s Fury, and more. Head here for more details.

What just happened? Enthusiast-class discrete graphics cards typically dominate conversations about high performance, but integrated GPUs aren't far behind when it comes to overclocking records for clock frequency. The latest world record holder recently explained how he managed voltage and temperature levels to push an Intel iGPU past the 4GHz mark for the very first time. Overclocker Pieter Massman recently detailed how he set a new graphics clock frequency world record at Computex 2025. While most recent record holders have used Nvidia's flagship RTX 4090, Massman surpassed them using the integrated GPU from an Intel Core Ultra 9 285K. With help from Asus overclocker Peter "Shamino" Tan, Massman pushed the Arrow Lake processor's Xe2-LPG 64EU iGPU to 4.25GHz – more than double its stock boost clock. The team achieved the feat twice, validating the results in CPU-Z during a livestream in the early days of this year's Computex event in Taiwan. According to Massman's blog, Skatter Bencher, the achievement marks a new world record for both integrated GPU clock frequency and GPU clock frequency overall. Since the RTX 4090 launched in 2022, frequency records have steadily climbed from around 3.3GHz to 4.02GHz in 2023. Massman had previously set the iGPU record at 3.9GHz during an Arrow Lake launch event late last year – using the same chip he would later overclock at Computex. The overclock involved setting the GT ratio to multiply the default reference clock by a factor of 85, the highest available setting. Initially, Massman supplied 1.3V to the integrated GPU via a VccGT voltage rail dedicated to the CPU's graphics tile cores, but this only reached 3.1GHz. // Related Stories Pushing further required a delicate balance of overvolting and liquid nitrogen cooling, ultimately achieving 4.25GHz with 1.7V and a temperature of -170°C. However, measuring the iGPU's performance at those settings overwhelmed several common benchmarking tools. Furmark crashed after reaching 2,800 points in 1080p, 3DMark Speed Way halted around 650 marks, and GPUPI 1B only ran for about 17.9 seconds. To stabilize the system, the team overclocked the graphics die-to-die interface and increased the reference clock. While discrete and overall GPU clock frequency records have steadily risen since the early 2000s, progress with iGPUs only resumed recently. After breaking the 2GHz barrier in 2011, integrated GPU overclocking plateaued for nearly a decade before surpassing 3GHz in 2023. Whether graphics overclocking will stagnate as CPU frequency gains have since 2010 remains to be seen.

Tech AI robots help nurses beat burnout and transform hospital care Hospitals using AI-powered robots to support nurses, redefine patient care Published June 4, 2025 6:00am EDT close AI robots help nurses beat burnout and transform hospital care Artificial intelligence and robotics may help with nursing shortage. NEWYou can now listen to Fox News articles! The global healthcare system is expected to face a shortage of 4.5 million nurses by 2030, with burnout identified as a leading cause for this deficit. In response, Taiwan's hospitals are taking decisive action by integrating artificial intelligence and robotics to support their staff and maintain high standards of patient care. AI-powered Nurabot (Nvidia)Nurabot: The AI nursing robot changing patient careNurabot, a collaborative nursing robot developed by Foxconn and Kawasaki Heavy Industries with Nvidia's AI technology, is designed to take on some of the most physically demanding and repetitive tasks in clinical care. These include delivering medications, transporting samples, patrolling wards and guiding visitors through hospital corridors. By handling these responsibilities, Nurabot allows nurses to focus on more meaningful aspects of patient care and helps reduce the physical fatigue that often leads to burnout. AI-powered Nurabot (Nvidia)Using AI to build the hospitals of the futureFoxconn's approach to smart hospitals goes beyond deploying robots. The company has developed a suite of digital tools using Nvidia platforms, including AI models that monitor patient vitals and digital twins that simulate hospital environments for planning and training purposes.The process starts in the data center, where large AI models are trained on Nvidia supercomputers. Hospitals then use digital twins to test and train robots in virtual settings before deploying them in real-world scenarios, ensuring that these systems are both safe and effective.ARTIFICIAL INTELLIGENCE TRANSFORMS PATIENT CARE AND REDUCES BURNOUT, PHYSICIAN SAYS AI-powered Nurabot (Nvidia)AI robots in real hospitals: Results from Taiwan's Healthcare SystemTaichung Veterans General Hospital (TCVGH), along with other top hospitals in Taiwan, is at the forefront of this digital transformation. TCVGH has built digital twins of its wards and nursing stations, providing a virtual training ground for Nurabot before it is introduced to real hospital floors. According to Shu-Fang Liu, deputy director of the nursing department at TCVGH, robots like Nurabot are augmenting the capabilities of healthcare staff, enabling them to deliver more focused and meaningful care to patients. AI-powered Nurabot (Nvidia)Ways Nurabot reduces nurse burnout and boosts efficiencyNurabot is already making a difference in daily hospital operations. The robot handles medicine deliveries, ward patrols and visitor guidance, which Foxconn estimates can reduce nurse workloads by up to 30%. In one ward, Nurabot delivers wound care kits and educational materials directly to patient bedsides, saving nurses multiple trips to supply rooms and allowing them to dedicate more time to their patients. The robot is also especially helpful during visiting hours and night shifts, when staffing levels are typically lower.Nurses hope future versions of Nurabot will be able to converse with patients in multiple languages, recognize faces for personalized interactions and even assist with lifting patients when needed. For example, a lung patient who needs two nurses to sit up for breathing exercises might only require one nurse with Nurabot's help, freeing the other to care for other patients. AI-powered Nurabot (Nvidia)Kurt's key takeawaysWhen it comes to addressing the nursing shortage, Taiwan is demonstrating that AI and robotics can make a significant difference in hospitals. Instead of spending their shifts running errands or handling repetitive tasks, nurses now have robots like Nurabot to lend a hand. This means nurses can focus their energy on what matters most – caring for patients – while robots handle tasks such as delivering medication or guiding visitors around the hospital.It's a team effort between people and technology, and it's already helping healthcare staff provide better care for everyone.CLICK HERE TO GET THE FOX NEWS APPHow would you feel if a robot, not a human, delivered your medication during a hospital stay? Let us know by writing us at Cyberguy.com/Contact.For more of my tech tips and security alerts, subscribe to my free CyberGuy Report Newsletter by heading to Cyberguy.com/Newsletter.Ask Kurt a question or let us know what stories you'd like us to cover.Follow Kurt on his social channels:Answers to the most-asked CyberGuy questions:New from Kurt:Copyright 2025 CyberGuy.com. All rights reserved. Kurt "CyberGuy" Knutsson is an award-winning tech journalist who has a deep love of technology, gear and gadgets that make life better with his contributions for Fox News & FOX Business beginning mornings on "FOX & Friends." Got a tech question? Get Kurt’s free CyberGuy Newsletter, share your voice, a story idea or comment at CyberGuy.com.

html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd" Epic Games is rebranding RealityCapture, its professional desktop photogrammetry software, as RealityScan.RealityScan 2.0, due in the “coming weeks”, will unify the desktop application with the existing RealityScan: Epic Games’ free 3D scanning app for iOS and Android devices. The update will also introduce new features including AI-based mask generation, support for aerial Lidar data, and new visual tools for troubleshooting scan quality. A desktop photogrammetry tool for games, VFX, visualization and urban planning First released in 2016, RealityCapture generates accurate triangle-based meshes of real-world objects, from people and props to environments.Its core photogrammetry toolset, for generating 3D meshes from sets of source images, is augmented by support for laser scan data. The software includes features aimed at aerial surveying and urban planning, but is also used in the entertainment industry to generate assets for use in games and VFX. RealityCapture was acquired by Epic Games in 2021, which made the software available free to artists and studios with revenue under $1 million/year last year. Now rebranded as RealityScan to unify it with the existing mobile app RealityCapture 2.0 – or rather, RealityScan 2.0 – is a change of branding, with the desktop application taking its new name and logo from Epic Games’ existing mobile scanning app.First released in 2022, RealityScan was originally pitched as a way to make RealityCapture’s functionality accessible to hobbyists as well as pros. It’s a pure photogrammetry tool, turning photos captured on a mobile phone or tablet into textured 3D models for use in AR, game development or general 3D work. RealityScan 2.0: AI masking, new Quality Analysis Tool, and support for aerial Lidar data New features in RealityCapture 2.0 will include AI-powered masking, with the software automatically identifying and masking out the background of the source images.The change should remove the need to generate masks manually, either in RealityCapture itself or an external DCC app. In addition, the default settings have been updated to improve alignment of source images, particularly when scanning objects with smooth surfaces and few surface features. To help troubleshoot scans, a new Quality Analysis Tool displays heatmaps showing parts of the scan where more images may be needed to reconstruct the source object accurately. The update will also introduce support for aerial Lidar data, which may be used alongside aerial photography and terrestrial data to reconstruct environments more accurately. No information yet on how the new features break down between desktop and mobile It isn’t clear which of those new features will be included in the mobile app, although it will presumably also be updated to version 2.0 at the same time, since Epic Games’ blog post announcing the changes describes its aim as to “unify the desktop and mobile versions”.We’ve contacted Epic for more information, and will update if we hear back. Price, system requirements and release date RealityScan 2.0 is due in the “coming weeks”. Epic Games hasn’t announced an exact release date, or any changes to price or system requirements.The current version of the desktop software, RealityCapture 1.5, is available for Windows 7+ and Windows Server 2008+. It’s CUDA-based, so you need a CUDA 3.0-capable NVIDIA GPU. The desktop software is free to artists and studios with revenue under $1 million/year. For larger studios, subscriptions cost $1,250/seat/year. The current version of the mobile app, RealityScan 1.6, is compatible with Android 7.0+, iOS 16.0+ and iPadOS 16.0+. It’s free, including for commercial use. By default, its EULA gives Epic Games the right to use your scan data to train products and services, but you can opt out in the in-app settings. Read Epic Games’ blog post announcing that it is rebranding RealityCapture as RealityScan Read more about RealityCapture and RealityScan on the product website Have your say on this story by following CG Channel on Facebook, Instagram and X (formerly Twitter). As well as being able to comment on stories, followers of our social media accounts can see videos we don’t post on the site itself, including making-ofs for the latest VFX movies, animations, games cinematics and motion graphics projects.

Over eight years after its predecessor launched and many rumors later, the Nintendo Switch 2 is finally available worldwide. It retails for $449.99 with a bundle available for $499.99 that includes Mario Kart World. Check out the launch trailer below. The Switch 2 offers several major improvements over the original, including a larger screen with HDR support. It can run in 120 frames per second in Handheld Mode and output to 4K resolution when docked. The Joy-Con 2 has also been revamped, sporting magnetic locks and mouse support. Other notable features include upscaling via Nvidia DLSS, though sadly, VRR support isn’t available in Docked Mode. Launch titles include Mario Kart World at a whopping $79.99, Nintendo Switch 2 Welcome Tour for $9.99, and several third-party games like Yakuza 0 Director’s Cut, Cyberpunk 2077: Ultimate Edition, Street Fighter 6 Years 1-2 Fighters Edition, and much more. Titles like The Legend of Zelda: Breath of the Wild and Tears of the Kingdom have also received Switch 2 Editions, with paid upgrades available for current owners. Of course, several existing titles have received free updates to help them perform better on the new console, including Super Mario Odyssey, Super Mario 3D All-Stars + Bowser’s Fury, and more. Head here for more details.

NVIDIA is working with companies worldwide to build out AI factories — speeding the training and deployment of next-generation AI applications that use the latest advancements in training and inference. The NVIDIA Blackwell architecture is built to meet the heightened performance requirements of these new applications. In the latest round of MLPerf Training — the 12th since the benchmark’s introduction in 2018 — the NVIDIA AI platform delivered the highest performance at scale on every benchmark and powered every result submitted on the benchmark’s toughest large language model (LLM)-focused test: Llama 3.1 405B pretraining. The NVIDIA platform was the only one that submitted results on every MLPerf Training v5.0 benchmark — underscoring its exceptional performance and versatility across a wide array of AI workloads, spanning LLMs, recommendation systems, multimodal LLMs, object detection and graph neural networks. The at-scale submissions used two AI supercomputers powered by the NVIDIA Blackwell platform: Tyche, built using NVIDIA GB200 NVL72 rack-scale systems, and Nyx, based on NVIDIA DGX B200 systems. In addition, NVIDIA collaborated with CoreWeave and IBM to submit GB200 NVL72 results using a total of 2,496 Blackwell GPUs and 1,248 NVIDIA Grace CPUs. On the new Llama 3.1 405B pretraining benchmark, Blackwell delivered 2.2x greater performance compared with previous-generation architecture at the same scale. On the Llama 2 70B LoRA fine-tuning benchmark, NVIDIA DGX B200 systems, powered by eight Blackwell GPUs, delivered 2.5x more performance compared with a submission using the same number of GPUs in the prior round. These performance leaps highlight advancements in the Blackwell architecture, including high-density liquid-cooled racks, 13.4TB of coherent memory per rack, fifth-generation NVIDIA NVLink and NVIDIA NVLink Switch interconnect technologies for scale-up and NVIDIA Quantum-2 InfiniBand networking for scale-out. Plus, innovations in the NVIDIA NeMo Framework software stack raise the bar for next-generation multimodal LLM training, critical for bringing agentic AI applications to market. These agentic AI-powered applications will one day run in AI factories — the engines of the agentic AI economy. These new applications will produce tokens and valuable intelligence that can be applied to almost every industry and academic domain. The NVIDIA data center platform includes GPUs, CPUs, high-speed fabrics and networking, as well as a vast array of software like NVIDIA CUDA-X libraries, the NeMo Framework, NVIDIA TensorRT-LLM and NVIDIA Dynamo. This highly tuned ensemble of hardware and software technologies empowers organizations to train and deploy models more quickly, dramatically accelerating time to value. The NVIDIA partner ecosystem participated extensively in this MLPerf round. Beyond the submission with CoreWeave and IBM, other compelling submissions were from ASUS, Cisco, Dell Technologies, Giga Computing, Google Cloud, Hewlett Packard Enterprise, Lambda, Lenovo, Nebius, Oracle Cloud Infrastructure, Quanta Cloud Technology and Supermicro. Learn more about MLPerf benchmarks.

Humans learn the norms, values and behaviors of society from each other — and Bernt Børnich, founder and CEO of 1X Technologies, thinks robots should learn like this, too. “For robots to be truly intelligent and show nuances like being careful around your pet, holding the door open for an elderly person and generally behaving like we want robots to behave, they have to live and learn among us,” Børnich told the AI Podcast. 1X Technologies is committed to building fully autonomous humanoid robots, with a focus on safety, affordability and adaptability. Børnich explained how 1X Technologies uses a combination of reinforcement learning, expert demonstrations and real-world data to enable its robots to continuously learn and adapt to new situations. NEO, the company’s upcoming robot, can perform household tasks like vacuuming, folding laundry, tidying and retrieving items. It’s built with operational safety at its core, using tendon-driven mechanisms inspired by the human musculoskeletal system to achieve low energy consumption. Børnich highlights the potential for robots to enhance human productivity by helping handle mundane tasks, freeing people up to focus more on interpersonal connections and creative activities. Learn more about the latest in physical AI and robotics at NVIDIA GTC Paris, which takes place from June 10-12. Register to attend humanoid-related sessions, including: “An Introduction to Humanoid Robots” for a deep dive into NVIDIA Isaac GR00T “How Physical AI Is Shaping the Next Generation of Industrial Robots” to learn how robotics leaders are making robots more intuitive and effective across industries Time Stamps 05:18 – 1X Technologies’ approach to robot safety. 11:36 – How world models enable robots to search backwards from the goal. 16:51 – How robots can free humans up for more meaningful activities. 22:29 – NEO answers the door so Børnich can interview a candidate. You Might Also Like…  How World Foundation Models Will Advance Physical AI With NVIDIA’s Ming-Yu Liu AI models that can accurately simulate and predict outcomes in physical, real-world environments will enable the next generation of physical AI systems. Ming-Yu Liu, vice president of research at NVIDIA and an IEEE Fellow, explains the significance of world foundation models — powerful neural networks that can simulate physical environments. Roboflow Helps Unlock Computer Vision for Every Kind of AI Builder Roboflow’s mission is to make the world programmable through computer vision. By simplifying computer vision development, the company helps bridge the gap between AI and people looking to harness it. Cofounder and CEO Joseph Nelson discusses how Roboflow empowers users in manufacturing, healthcare and automotive to solve complex problems with visual AI. Imbue CEO Kanjun Qiu on Transforming AI Agents Into Personal Collaborators Kanjun Qiu, CEO of Imbue, explores the emerging era where individuals can create and use their own AI agents. Drawing a parallel to the PC revolution of the late 1970s and ‘80s, Qiu discusses how modern AI systems are evolving to work collaboratively with users, enhancing their capabilities rather than just automating tasks.