Sparsity Technologies’ Post

Is Your Operating System About to Become Obsolete? Windows and macOS have ruled for decades—but a new era is emerging: Agentic Operating Systems. Instead of you managing apps and tasks, an Agentic OS uses proactive AI agents that understand your intent, automate workflows, and operate across apps without you touching a thing. And this isn’t science fiction: -AIOS – an open-source AI-native OS with LLMs embedded into the kernel -PwC’s Agent OS – orchestrates multiple AI agents for enterprise workflows -Microsoft – publicly shifting Windows toward an “agentic OS” Why legacy OS vendors should be worried: -Innovator’s Dilemma: Incumbents risk missing the next big shift -Legacy architecture: Windows/macOS weren’t built for AI-native computing -AI-native startups: Building fresh, faster, and without technical debt The future is agentic. This shift will be as big as the jump from command line to GUI—unlocking new productivity, creativity, and collaboration. So the real question is: Will Microsoft and Apple adapt, or become the next Nokia and Blackberry of the OS world? What do you think—are you ready for an OS that works for you? #AgenticOS #AI #FutureOfWork #TechInnovation #ArtificialIntelligence #Microsoft #Apple #Windows #macOS #Disruption #NextBigThing #Technology #Innovation #AIOS #PwC

The DC-ROMA RISC-V Pad II is a tablet announced by DeepComputing (a RISC‐V specialist) aimed at the developer / early adopter crowd. It uses a RISC-V based SoC: the SpacemiT K1 (8‐core 64-bit RISC-V) under the hood. It ships pre-installed with Ubuntu 24.04 Desktop (so a “full” Linux desktop experience, not just a mobile OS). _________________________________________ Specifications: * Display: 10.1‐inch IPS panel, resolution ~ 1920×1200, 10‐point touch. * SoC: SpacemiT K1 – 8 RISC-V cores, vector extensions (RVV), AI/NPU features. * RAM & Storage variants: from 4 GB RAM / 64 GB eMMC up to 16 GB RAM / 128 GB eMMC. * Battery: ~6000 mAh. * I/O: USB-C (USB3 + DisplayPort Alt Mode), 3.5mm headphone jack. Rear camera ~5 MP, front ~2 MP. * Price: starting at ~$149 for the base model. This isn’t a typical mass‐market consumer tablet — the emphasis is on supporting the RISC-V ecosystem. The company explicitly says the tablet is “for developers, early adopters” who want native RISC-V development and compilation on the device. RISC-V is an open‐standard instruction set architecture (ISA) that is royalty‐free and gaining traction. Running Ubuntu out of the box on RISC-V hardware is still fairly rare; this device signals more practical hardware availability for that platform. That gives Linux enthusiasts, developers and system hackers something interesting. Even though Ubuntu is supported on RISC-V, the overall ecosystem (drivers, apps optimized, touch/screen experience) may still be catching up. My friend 🗡⚔️🛡Yuning Liang from DeepComputing can help explain the product better than I Enjoy #linux 🐧

After the BSP post, let’s move on to one of the most interesting parts of embedded Linux: the boot process. When you power on an embedded board, Linux doesn’t start instantly. It goes through several steps before reaching user space, and each stage has a clear role. Stage 0 – Power-On Reset Power rails stabilize, the CPU resets, and BootROM execution starts. Stage 1 – BootROM This is manufacturer code stored in the SoC. It initializes basic hardware and loads the first bootloader (SPL or MLO) into SRAM. Stage 2 – SPL / MLO Runs from SRAM, sets up DDR and clocks, and loads the main bootloader (U-Boot). Stage 3 – U-Boot Initializes peripherals, loads the Linux kernel and device tree into DDR, and jumps to the kernel entry point. Stage 4 – Linux Kernel Decompresses itself, sets up memory management, initializes drivers, mounts the root filesystem, and starts the first process (init). Stage 5 – Init System Mounts /proc, /sys, /dev, and launches background services like udevd or systemd. Stage 6 – User Space & Applications At this point, the system is ready — network services start, logins are available, and user applications can run. Each stage prepares the next, moving from small internal memory to full DDR. Understanding this sequence helps a lot when debugging boot issues or porting Linux to new hardware. #EmbeddedLinux #Bootloader #Uboot #LinuxKernel #Firmware #SoC #IoT #LinuxDevelopment #SystemDesign #ProDiSynInnovations

 If I had a nickel for every AI demo that still requires manual document uploads, I wouldn't need any funding. We integrated AI directly into the OS 8 months ago. It can run on Apple's MacOS, Microsoft's Windows, and even on Qualcomm's ARM Processors #snapdragon. StarlightSearch is an Agent and Search integrated with the OS. #local #agent #search #generativeAI #rag

💻 Return to Linux: SBC Recommendations for Built-in Chrome AI Excited to be re-entering the world of Linux after a 10-year hiatus, largely driven by the purchase of three #unoQ boards! My prior Linux experience (circa 2015, using Ubuntu) was focused on heavy-duty, 3-day-long RNN machine learning models. At the time, I needed more power than a standard machine and dismissed not less power like a Raspberry Pi. My latest use for Linux has only been when the Windows Arduino IDE was too slow or didn't work properly. Not using Linux led me down the path of messing with what we now call WebAI, EdgeAI, and tinyML on Windows Now, I have a new challenge and a question for the community: What is the cheapest Single Board Computer (SBC) that can successfully run Chrome's Built-in-AI? The primary challenge is the minimum hardware requirement: Storage: At least 22 GB of free disk space for the AI model and a 4 GB download. I'm aware the NVIDIA Jetson Nano Orion would likely work, but I'm unsure about the Raspberry Pi 5. Any suggestions, or better yet, proof of a working setup, would be greatly appreciated! I've heard you can potentially use a USB thumb drive to allocate some of the required disk space—has anyone tried this for Chrome's AI model download? Does Chromium (the open-source version) support the Built-in AI features? #Linux #SBC #EdgeAI #TinyML #AIhardware #SingleBoardComputer #NVIDIAJetson #RaspberryPi If your are curious why Chrome-Built-In-AI is so important to #makers watch these ~3 min long videos in my Chrome Challenge 2025 playlist. I made all 9 of these apps in the last 6 weeks. https://lnkd.in/gJjREG_T Thomas Steiner, PhD, Connie Leung, Jason Mayes, Jd Fiscus, Kenji Baheux anyone else with an opinion?

Microsoft is encouraging users to “talk to their laptops” as it rolls out new artificial intelligence (AI) updates for Windows 11 while urging customers to move away from Windows 10. READ MORE: https://lnkd.in/ghfP74e8

PC Hardware and Software TUXEDO Unveils InfinityBook Max 15: A Linux Laptop with Ryzen AI 9 TUXEDO Computers has launched the InfinityBook Max 15, a new Linux laptop powered by the Ryzen AI 9 processor. Starting at €1,689, this device aims to deliver robust performance and AI capabilities for Linux users. Source URL: https://lnkd.in/gs6ZpcdT #PCHardware #Linux #RyzenAI #TechNews AI Developments CIQ: Rocky Linux Is Authorized Linux Distribution for NVIDIA AI Stack CIQ has announced expanded capabilities for Rocky Linux from CIQ (RLC), now including NVIDIA DOCA OFED support and authorizing it as a Linux distribution for the NVIDIA AI Stack. This integration enhances RLC's ability to support high-performance AI workloads. Source URL: https://lnkd.in/gCVDcMRU #AI #NVIDIA #Linux #HighPerformanceComputing #TechNews Linux Ryzen AI Software as AMD's collection of tools and libraries for AI inferencing on AMD Ryzen AI class PCs has Linux support with its newest point release. AMD's Ryzen AI Software now officially supports Linux with its latest release, providing a comprehensive suite of tools and libraries for AI inferencing on Ryzen AI-class PCs. This expands opportunities for Linux users in AI development and applications. Source URL: https://lxer.com/ #Linux #AI #AMD #OpenSource #TechNews

Those who know me have heard me praise Thinkpads and Arch linux since at least 2018~2019, it really is amazing to be in full control of your hardware and Software, and be able to customise it to your tune. I've however moved away from pc/linux for my development main driver because simply the best cpu out there is no longer Intel or AMD, it's in fact Apple silicon, and this is what matters to me since I buy a laptop to last me 10 years. The reason I'm sharing this is the whole hype around Omarchy, DHH did a great job packaging a great distro for devs and many have flocked there, which is alright, but I don't agree that its a better alternative to Mac, at least not today; you are settling for lower performance with the PC platform today (specifically CPU, GPU is a whole different story) , that might change in the future but today PC is in second place. I realise to many devs, peak performance isn't a top priority, at some point good performance is enough, but if you do care about your equipment's longivity and have a performance hungry dev ecosystem, you might want to consider more than the operating system.

Microsoft just shared what Windows might look like by 2030—and it’s all about AI and voice control. They’re imagining a future where computers see what we see, hear what we hear, and respond naturally. Think less mouse and keyboard, more talking and gesturing. That might sound futuristic, but it’s already starting. Windows 11 has AI tools popping up, and newer laptops are being built with special AI chips to support them. Still, not every idea sticks—remember when Copilot was supposed to change all your settings just by typing a request? Yeah, not quite. I don’t think the mouse and keyboard are going away anytime soon, but it does feel like we’re heading toward a world where voice and AI are part of how we use computers—maybe not replacing the old ways, but adding new layers on top. Would you talk to your PC if it actually understood you? Or are you just fine with clicking and typing like always?

Linux Just Got Better – AMD vs NVIDIA, KDE Gets Lean ... Linux has seen notable advancements this week, featuring a new Ryzen AI release that significantly boosts AI capabilities on the platform. Users can also look forward to an optimized KDE Plasma, now saving memory, and ongoing efforts by Ubuntu to enhance the overall Linux desktop experience. Source: https://lnkd.in/gwvjrEQn #Linux #AMD #NVIDIA #KDE #Ubuntu #RyzenAI #TechNews High Performance Computing Recent News: How Linux Optimizes AI Hardware Acceleration Discover how Linux is at the forefront of optimizing AI hardware acceleration, enhancing performance for machine learning workloads. This article delves into the critical role Linux plays in maximizing the efficiency of AI infrastructure. Source: https://lnkd.in/g_BJs97U #AI #Linux #HardwareAcceleration #MachineLearning