Bryan Zimmerman

The RISC-V vs ARM debate has become nothing more than tech tribalism that makes engineers feel virtuous while completely overlooking what actually determines project success in the real world. We've all encountered, 'RISC-V will democratize computing with its open-source model!' But let's be brutally honest: in 80% of commercial deployments, ecosystem availability beats architectural ideals every time. The countless hours spent battling incompatible RISC-V toolchains could have been invested in optimizing ARM-based solutions. And don't get me started on teams that select RISC-V for 'openness' only to find their custom ISA extensions have locked them into a single vendor more tightly than ARM's licensing ever could. I've observed directly projects hemorrhage 6 months just to get basic RISC-V compiler support working reliably for standard peripherals, while comparable ARM implementations reached production in 8 weeks leveraging battle-tested CMSIS drivers. The most ironic part? The 'open-source' narrative conveniently obscures that 70% of production RISC-V systems rely on proprietary extensions and commercial toolchains - you're essentially swapping ARM's licensing costs for RISC-V vendor support agreements. What we genuinely need is ruthless pragmatism: ARM when time-to-market and peripheral support are critical, RISC-V when custom accelerators or extreme cost constraints drive the design. This isn't compromise - it's engineering common sense. When your product launch is next quarter and you need functional WiFi, BLE, and security subsystems, every absent driver in the RISC-V ecosystem directly threatens your schedule. The next time you feel drawn to RISC-V for ideological reasons, ask yourself: Will this choice genuinely accelerate my development timeline, or am I volunteering to be a toolchain guinea pig? ⚖️ What single factor tipped the scales between RISC-V and ARM in your last project? Share your make-or-break decision criteria below - bonus points for how you'd evaluate the trade-off differently today! #RISCV #ARM #EmbeddedSystems #ProcessorArchitecture #EngineeringReality #EcosystemMaturity #TechChoices #DeveloperExperience #FPGA #SoC #CostOptimization #TimeToMarket

167

17 Comments

I had the pleasure of chatting with Kwabena Agyeman, co-founder of Open MV, about how they’re bringing advanced computer vision to microcontrollers, a field once considered impossible. Kwabena shared the story of how Open MV went from a side project to a game-changing platform, starting with bootstrapping in 2015 and now delivering products like the OpenMVCAM AE3 and N6. These devices pack a 600 billion instructions per second neural net accelerator into a tiny form factor. 💡 Key Takeaways: ✅ Microcontrollers, not application processors: Think tiny, low-power, heat-efficient devices that can run on batteries for years. ✅ MicroPython for easier coding, making it accessible for a new wave of Python developers. ✅ Real-world applications, from smart door locks with facial recognition (think year-long battery life) to industrial monitoring and automated event reading. ✅ No big Linux footprint, bare-metal firmware with MicroPython layered on top, bringing Python’s power to the embedded world. We also dove deep into topics like: Why C & C++ still matter for performance The potential of ML in future IoT applications The balance of privacy, power, and convenience in embedded AI The trade-offs of not recording video for battery life and security Kwabena’s passion for embedded systems is inspiring, and his insights highlight why this niche space is so important for the future of AI at the edge. 🔗 Watch the full video here: https://lnkd.in/gBF-ENwU I’d love to hear what you think about the intersection of embedded engineering and machine learning, and if you’re exploring similar applications in your projects! Check out the kickstarter project here: https://lnkd.in/ghwdStK8 #OpenMV #Microcontrollers #EmbeddedAI #MachineLearning #EdgeComputing #IoT #MicroPython #ComputerVision #TechInnovation

16

Later this quarter, we will publish the 2026 AV Talent Demand Report. Some things I am expecting to see: Machine Learning Engineers will continue to be the highest value technical job title. In particular, the data will highlight the challenge of finding staff and senior level engineers with autonomy related specializations like perception, motion planning and simulation. The robotaxi industry will remain a significant part of the demand, but an increasing percentage will come from other types of AVs, such as trucking, industrial and military. Operations hiring will have increased by 400% or more compared to last year. The new report will be available late March. For a reminder of last years whitepaper, visit: https://lnkd.in/eNmqH4uC

21

3 Comments

We’re helping a robotics team scale from “working prototype” → real-world deployment. They already have robots operating outside the lab. Now they need the engineers who make systems survive production. We’re #hiring multiple firmware engineers (senior + principal level) to work on: • Embedded systems running continuously in the field • Reliability, fault handling, and recovery, not just happy-path behavior • Hardware bring-up, debugging, and performance tuning • Sensors, actuators, and real-time communication • Closing the loop between field data → iteration → improved uptime This is not a research role. This is not a simulation-only role. This is for engineers who care about: keeping systems alive diagnosing failures fast and shipping updates that actually improve operations The environment is hands-on, multidisciplinary, and deployment-focused — the kind of place where firmware decisions directly impact real-world productivity every day. If you enjoy debugging issues that only happen at 2am in the field more than writing slides about architecture… you’ll probably like this. DM me or Varunn V. if you want details.

7

2 Comments

My honest opinion on #Rust: it’s cool that it can guarantee memory safety (or dangling references in #C++ speak), but it’s verbose. With modern C++ a reasonably skilled programmer can write provably correct, and terse, code. Also: why does Rust not have language-native variadic functions/templates?

5

🚨 BREAKING: Pentagon Invests $800M in “Agentic AI” The US Department of Defense just awarded $200M each to Anthropic, Google, OpenAI, and Elon Musk’s xAI to advance agentic AI, a new class of artificial intelligence that doesn’t just generate content, but takes action. These systems are designed to autonomously plan, decide, and execute tasks, marking a major shift from traditional AI toward operational deployment in government and defense. ✅ @xAI will adapt its Grok chatbot for secure government use ✅ Anthropic is rolling out ClaudeGov, focused on national security ✅ Google brings its Gemini + high-security cloud tools ✅ OpenAI was first to receive funding, now building “frontier AI” systems for the DoD Why it matters: This is the Pentagon betting big on commercial AI leaders not homegrown defense contractors to power the next wave of military-grade autonomy. The implications for national security, ethics, and tech regulation are enormous. Full story via Breaking Defense: #AI #DefenseTech #AgenticAI #Pentagon #Anthropic #OpenAI #xAI #Google #MilitaryAI #EmergingTech https://lnkd.in/eTvVnHu9

11

⚛️ Oak Ridge National Laboratory and Atomic Canyon signed a MoU to streamline licensing for nuclear power plants using #AI to speed up license application reviews. The agreement enables Atomic Canyon—a member of the NVIDIA Inception Program—to further develop its Neutron AI Platform, accelerated by NVIDIA, with enhanced cybersecurity features. ✨ Atomic Canyon developed its original FERMI AI models, which powers Neutron AI, using ORNL's Frontier supercomputer. ➡️ Read the announcement from Oak Ridge National Laboratory.

2

In this highly detailed “A History of Large Language Models,” Gregory Gundersen walks us through the intellectual lineage behind today’s large language models (LLMs). This article covers key building blocks like distributed representations, sequence transducers, and the progression of architectures that enabled modern AI systems. The author argues that what we often treat as “black box” models actually rest on decades of foundational research, and underscores why understanding that history is essential for responsibly shaping the future of AI. https://lnkd.in/e8zBex-W

3

Can Anyone Threaten NVIDIA Dominance in Autonomous Systems? Over the past decade, NVIDIA has established itself as the undisputed leader in the AI and compute backbone powering autonomous vehicles, robotics, aerospace, and defense applications. Its end-to-end stack—from silicon to software (CUDA, TensorRT, Drive, Isaac, Jetson, Orin, and now Blackwell)—has made it the go-to platform for companies building autonomy at scale. But with the rapid evolution of edge computing, real-time inference, and vertical-specific AI accelerators, the question arises: Can anyone challenge NVIDIA’s grip on the autonomous stack? Several companies are making moves: Qualcomm - the obvious one! Snapdragon Ride and Arriver stack have established inroads into automotive with power-efficient compute for ADAS and autonomy. Mobileye continues to build dominance with its custom EyeQ chips and tightly integrated software/hardware offering. Intel Corporation is attempting a return via Habana, Gaudi, and strategic investments, including Mobileye. Tesla has taken the in-house route with its FSD chip and Dojo supercomputer—betting big on vertical integration. Ambarella Inc, Mythic, and Tenstorrent are betting on power-optimized, domain-specific accelerators at the edge. AI-focused startups like d-Matrix, Hailo, and Recogni are designing AI inference chips optimized for latency, efficiency, and performance-per-watt. Still, the ecosystem advantage that NVIDIA commands—from developer community to software compatibility to scalable silicon—makes it incredibly hard to displace. 👉 Is there anyone who can realistically threaten NVIDIA’s dominance in the autonomous space? 👉 Will innovation come from within the silicon world—or from a radically different approach to autonomy? #AutonomousVehicles #EdgeAI #Semiconductors #NVIDIA #Autonomy #Robotics #AIDrivenFuture #Autonomous

8

Quick Highlight on the Robo-taxi scene! Waymo announced earlier this week a $16 billion (yes, with a B…) investment round, valuing the company at a whopping $126 billion. The focus now shifts to execution: expanding Waymo’s commercial services to 20+ new cities in 2026, including international markets like London and Tokyo. Specially in London, this could become a significant competitive edge over local player Wayve, which is also rolling out its platform in partnership with Uber this year. It’ll be interesting to see how the market reacts- and who ultimately comes out ahead.

27