Embedded AI Podcast E12: Learning AI-powered development with Jacob Beningo

We sit down with Jacob Beningo, consultant and trainer for embedded systems, to talk about what we've learned teaching engineers to use AI for firmware development. Spoiler: it's not about typing faster. As it turns out, AI doesn't let you skip the hard parts of software engineering. If anything, it forces you to get better at requirements, architecture, and systems thinking—because if you can't clearly describe what you want, the AI will happily generate something you didn't ask for. We discuss common pitfalls, the surprisingly tricky skill of prompting, why IDE plugins might be leading us astray, and how AI works best as a demanding pair programming partner who never lets you get away with vague thinking. 0:00 Recording Start and Introductions 1:41 Meet Jacob Beningo 3:44 Beyond the Curly Brackets 6:06 Systems Engineering Matters 9:13 Training AI for Embedded 23:06 Validation and Trust Issues 28:05 The Cost of AI 31:06 Death of Agile Debate 36:19 Engineering Skepticism Required 42:37 AI for Development Tools 48:46 AI as Pair Programming 51:08 Resources and Closing