Devs: Don't Kill My Vibe with Harald Kirschner | Ready. Or Not. Episode 4

In Ready. Or Not. Episode 4, host Nathan Macintosh sits down with Harald Kirschner, engineer and open source leader, to unpack the internet’s newest workflow phenomenon: vibe coding. From VCR nostalgia to building products by talking to AI, they explore what it means to “code with vibes,” why it can be incredibly useful for prototyping, and where it goes off the rails when people ship software they do not fully understand. The conversation breaks down how vibe coding actually works in practice, why it is taking off with non technical builders and fast moving teams, and what real engineering guardrails still matter when you want to move quickly without creating security and cost nightmares. Harald also shares how AI can be a productive critical thought partner, why he likes to ask models to “roast” code, and what it takes to keep quality high when the vibe gets messy. What This Episode Covers * What vibe coding is and where the term came from * Why natural language workflows are changing how software gets built * When vibe coding is great for prototyping and testing ideas fast * Why vibe coded apps can break badly when shipped without review * The guardrails that keep speed from turning into security risk * How code review, linting, and testing catch the problems vibes miss * Why observability and metrics matter once software is in production * How AI can help find issues if you use it as a critical partner * What “good vibes” really mean in product experience and usability Chapters 00:00 The Evolution of Technology and AI 02:47 Understanding Open Source and Vibe Coding 05:58 The Role of Vibe Coding in Prototyping 08:42 Best Practices and Guardrails in AI Development 11:31 Vibe Coding in Parenting and Everyday Life TL;DR Vibe coding can help teams explore ideas fast, especially early on. But without guardrails like review, testing, and security checks, it is easy to ship something you do not understand and pay for it later. Key Takeaways * Vibe coding is best for exploration and disposable prototypes, not production by default * Speed without guardrails can create security gaps and unexpected costs * Quality still depends on fundamentals like review, testing, and secure patterns * AI can be a strong partner when you ask it to critique, not just generate * Great product “vibes” come from usability, performance, and resilience under real conditions Who This Episode Is For * CIOs and Technology Executives * CISOs and Security Leaders * Engineering Leaders and Software Architects * DevOps and Platform Engineering Teams * Product Managers and Design Leaders * Builders exploring AI assisted development * Business leaders navigating AI adoption FAQ Q: What is vibe coding? A: It is building software by guiding an AI with natural language and judging outputs by whether the product feels right, often without deeply inspecting the underlying code. Q: Is vibe coding only for developers? A: No. It can be especially helpful for non technical teams to prototype ideas quickly and validate whether something solves a real problem. Q: Why does vibe coding get a bad reputation? A: Problems happen when people ship vibe coded apps to production without understanding how they work or adding basic engineering safeguards like tests, reviews, and security controls. Q: What guardrails matter most when using AI to code? A: Code review, linting, testing, secure coding patterns, and observability once deployed so teams can detect and fix issues quickly. Q: Can AI improve code quality instead of reducing it? A: Yes, if you use it as a critical thinking partner, for example asking it to critique, find edge cases, or identify security and reliability risks. Dive deeper into cyber resilience, AI risk, and modern readiness on the Readiverse: https://bit.ly/4af4jR2 #ReadyOrNot #VibeCoding #AIEngineering #DevSecOps #CyberResilience #SoftwareDevelopment #AIGovernance #OpenSource