Where Did Embedded Engineering Go?
Open almost any embedded systems news feed today and one topic dominates everything: generative AI.
Where did the rest of engineering go?
Generative AI: people promote it, criticize it, publish recipes, show what they generated in a few hours, warn about the risks, or explain why everyone else is using it incorrectly.
Where are Linux, ZephyrOS and FreeRTOS? BSPs and device trees? Advanced architecture? Real-time constraints? Power management? Secure boot? OTA updates? Reliability? Long-term maintenance? Cybersecurity? Even the Cyber Resilience Act, which was supposed to keep the whole industry busy, has become strangely quiet.
Are these subjects becoming less important?
Are embedded teams spending less time on boot issues, race conditions, hardware constraints, memory bandwidth, power domains, corrupted updates and difficult field failures?
Have processors stopped having timing constraints? Have peripherals become perfectly reliable? Has power stopped disappearing at inconvenient moments? Are products no longer expected to survive ten or fifteen years?
Or have these subjects simply become less interesting to talk about?
AI topics produce an extraordinary amount of visible content. Every new model, coding agent or spectacular demonstration generates reactions, predictions and controversies that feed the insatiable maw of social media. A BSP migration does not. Neither does fixing an intermittent crash after seventy hours of operation.
Are We Really Discussing New Engineering Problems?
There is another question too. When:
- software architecture becomes “how to structure an AI-generated project”
- code review becomes “how to review AI-generated code”
- cybersecurity becomes “can AI-generated code be trusted”
are we really discussing new engineering problems, or are familiar ones simply being repackaged around AI?
The AI angle changes the headline. The underlying engineering subjects are still there.
Which raises a simpler question: have Linux, RTOS, architecture, cybersecurity, reliability and maintainability really disappeared from the conversation, or are we now talking about them only through the prism of AI?
I’m asking because at Embedded Expertise, our purpose has not changed.
We still work to make embedded systems powerful, efficient, understandable, robust, secure and maintainable, whether that means architecture, embedded Linux, real-time systems, cybersecurity, performance, power constraints, reliability, or embedded AI/ML.
Generative AI is just one more tool in the tool chest. We use it when it helps, alongside the others.
Enjoyed this article?
Embedded Notes is an occasional, curated selection of similar content, delivered to you by email. No strings attached, no marketing noise.
