Fireside Chat with Sam Newman and Neal Ford

Neal Ford and Sam Newman have each spent decades thinking about software architecture at its highest level. Neal coauthored _Head First Software Architecture_ and introduced architectural fitness functions, while Sam's _Building Microservices_ became required reading for distributed systems engineers worldwide. Now they're applying that accumulated expertise to one of the most urgent questions in software today: what does it actually mean to architect systems with agentic AI, and what's about to go badly wrong? This fireside chat from Software Architecture Superstream covers why agentic AI is stuck at the "advanced beginner" stage of the Dreyfus model, capable of trying recipes but incapable of judging whether they're actually good ones. Watch as Neal and Sam make the case that agents have no sense of value, judgment, or right and wrong, which makes them fundamentally nothing like a junior developer. They argue for modular thinking and boundary-level verification as the right architectural response, explain why architectural fitness functions are the guardrails agentic workflows desperately need, and introduce the "architecture quantum" as the correct scope for agentic code regeneration. They also draw sharp parallels to the SOA era, warning that transactional coupling is about to trip up agentic architectures all over again. Follow O'Reilly on: LinkedIn: https://www.linkedin.com/company/oreilly/ Facebook: http://facebook.com/OReilly Instagram: https://www.instagram.com/oreillymedia BlueSky: https://bsky.app/profile/oreilly.bsky.social