{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"Fundamentals of Software Engineering","title":"E06 - Why Software Engineering Fundamentals Matter More in the Age of AI","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/e9904614\"></iframe>","width":"100%","height":180,"duration":3764,"description":"In this episode of Fundamentals of Software Engineering, Nate and I dig into what we learned at the Arc of AI conference in Austin. We discuss the big conversations happening around AI-generated code in production, why vibe coding falls apart at scale, and how tools that let developers move faster can also get them off course just as quickly. From an AI agent that tried to delete production to 50,000-line diffs that no human can realistically review, we explore the real risks showing up as teams adopt these tools.\nWe also get into the durability of code, whether we'll eventually stop writing human-readable code altogether, and why evals for AI-generated code are something most teams are skipping at their peril. A recurring theme is that typing has never been the hard part of software. The fundamentals (testing, code review, architectural thinking, and understanding what code shouldn't do) matter even more now that we can produce code faster than ever. If you're feeling anxiety about AI replacing developers, this conversation is for you.\n\nKey Highlights\n💻 Code Is Cheap, Software Is Expensive: AI excels at generating behavior, but the architectural qualities and non-functional requirements that make code into software still require human expertise.\n🔐 The Authorization Bug: An AI agent commented out authentication code because it lacked the right role in dev. The kind of mistake a human reviewer catches instantly, but one that could slip through a 50,000-line diff unnoticed.\n🏗️ 200,000 Lines of Vibe Code: When someone asked Dan to review their vibe-coded app, it was 200,000 lines. Finding a bug in that is like being dropped into a massive enterprise codebase on day one with zero context.\n🤖 AWS Agent Gone Rogue: An agent tasked with fixing a critical production issue decided the best approach was to delete production and start from scratch. Any human would have vetoed that immediately.\n⚡ Moving Fast Safely: Like an F1 driver at 200 mph, moving faster in software...","thumbnail_url":"https://img.transistorcdn.com/1BiOcr3jOEw_uiwQk5MInsKiSAl8JXHgE7p7L1stz0g/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS82NmM2/MmE3OWEzYWVkMWFl/MWUxNzhkOWY1YzY1/Njg2Ny5qcGc.webp","thumbnail_width":300,"thumbnail_height":300}