Hardware Podcasts

246: Sacrifice That Computer

Embedded.fm · 18 May 2018 · 1h 12m · Episode 246

What this episode covers

Dr. Bennie Lewis, a Senior Research Scientist at Lockheed Martin, discusses the current state of machine learning, deep learning, and robotics. He explains why deep learning has become viable now due to hardware advances but argues it is not yet mature enough for critical, real-time embedded systems compared to classical control theory. Lewis shares his experience using NVIDIA Jetson platforms, noting that while the hardware is powerful, the software ecosystem can be unstable and frustrating for production environments. He advises beginners to start with robust platforms like LEGO Mindstorms and learn C++ to understand the underlying mechanics of AI integration.

Most relevant to

large enterprises

Topics

Mentioned

Companies
Lockheed Martin
Products
NVIDIA Jetson, LEGO Mindstorms
Tools
TensorRT, CUDA, Caffe, Matlab
Also
SAMS C++ in One Hour a Day

Related episodes from other shows

Other shows on the same ground, closest first.

  • The Robo Success Podcast · 30m

    Ep. 14: Zero to H.I.R.O.: The Path to Becoming a Full-Stack Mechatronics Engineer

    Joseph Casebeer, Founder and CEO of Takamori Robot, discusses his educational robotics kit, Hero, which uses a Raspberry Pi Pico microcontroller to teach embedded systems and hardware fundamentals. He explains the transition from graphical programming to raw code, emphasizing that deep technical knowledge is essential for debugging and optimizing in an AI-driven industry. Casebeer also highlights a partnership with Real Engineers, a hiring platform that evaluates candidates based on project portfolios rather than resumes, arguing that hands-on experience with hardware is critical for career success.

    Microcontrollers & dev platformsCareers & education
    8 chapters

  • The Pragmatic Embedded Podcast · 39m

    Christmas Special

    Hosts Jeff Gable and Luca share war stories from their careers in embedded systems, highlighting the pain of manual testing in helicopter avionics, the chaos of managing requirements with Word and Excel in Dropbox for medical devices, and the technical debt of massive #ifdef blocks in legacy machine tool software. They discuss the industry's slow adoption of modern software practices like version control and CI/CD, contrasting embedded development with web development standards. The episode concludes with book recommendations including 'The Phoenix Project' and 'Team Topologies', and reflections on the future of IoT security and agile adoption in hardware.

    Debugging & failure analysisEmbedded firmware
    6 chapters

  • Coredump Sessions · 55m

    #007: AI, Open Source, and the Future of Embedded Development: How Much Code Will We Actually Write?

    François Baldassari, CEO of Memfault, and Thomas Sarlandie, Field CTO of Memfault, host a panel with Jonathan Beri, CEO of Golioth, and Dan Mangum, CTO of Golioth, to discuss the intersection of AI, open source, and embedded development. The group analyzes the maturity of edge AI for audio and video processing, the challenges of integrating LLMs into hardware due to rapid model iteration, and the current state of using AI for firmware development. They conclude that while AI is transforming code generation and review, open source remains critical for training models and providing trusted, complex infrastructure like Zephyr and FreeRTOS.

    Embedded firmwareAI & machine learning
    9 chapters

  • The Pragmatic Embedded Podcast · 49m

    Buzzword Bingo (or: which modern technologies make sense for embedded?)

    Hosts Jeff Gable and Luca Ingianni evaluate modern development buzzwords for embedded systems, discussing the applicability of web technologies like Kubernetes and Docker, the nuances of CI/CD pipelines, and the practical adoption of Rust and modern C++. They highlight specific tools such as Memfault, Mender, and Renode for observability and simulation, and analyze the role of AI in both device-side machine learning and developer productivity.

    Embedded firmwareMicrocontrollers & dev platforms
    8 chapters

  • Oxide and Friends · 1h 11m

    Okay, Doomer: A Rebuttal to AI Doom-mongering

    Bryan Cantrill and Adam Leventhal of Oxide Computer Company rebut AI doomerism by detailing the immense complexity of building physical hardware, arguing that AI cannot yet replicate the ingenuity required for hardware bring-up. They share specific debugging war stories from Oxide’s rack-scale computer development, including firmware bugs with AMD Milan processors, manufacturing errors with press-fit DIMMs and PCIe connectors, and the critical role of human desperation and creativity in solving these problems. The hosts conclude that the physical world's brittleness and the need for hands-on intervention make the idea of self-building AI infrastructure implausible.

    Debugging & failure analysisAI & machine learning
    7 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by Embedded.fm. The episode itself belongs to its publisher.