Serverless CrAIc Ep84 AI-Generated Code Is a Liability: Technical Debt & Engineering Excellence

Send us Fan MailIs AI-generated code creating more value — or more liability?In this episode of Serverless Craic, David Anderson, Mark McCann, and Michael O’Reilly explore why one of software engineering’s oldest principles is suddenly more relevant than ever in the age of AI:“Code is a liability. The system is the asset.”As agentic AI and code generation tools accelerate development, teams are producing more code, more tests, and more complexity than ever before.But:Does more code actually mean better outcomes?Are organisations creating massive technical debt without realising it?What happens when AI accelerates poor engineering practices?And how do you maintain confidence, security, and quality in probabilistic systems?This episode explores:AI-generated code and technical debtValidation, verification, and testing strategiesObservability and evaluation frameworksSecurity vulnerabilities and unmanaged codeCritical thinking in modern software engineeringWhy “lines of code” ≠ business valueThe return of XP and foundational engineering principlesChapters00:00 – Introduction00:24 – Old engineering principles returning in the AI era00:51 – The return of Extreme Programming (XP)01:43 – “Code is a liability” explained02:46 – AI-generated code and growing technical debt03:32 – Why engineers must review AI-generated code carefully05:16 – The history of generated code and technical debt06:28 – Why more code doesn’t mean more value07:12 – AI hype, supply chains, and unmanaged complexity08:30 – AI accelerates weak engineering practices09:02 – Why teams still struggle with testing strategies10:39 – Observability and deploying with confidence11:54 – Evaluation frameworks for probabilistic systems12:55 – System boundaries and verification13:11 – Engineers are still accountable for AI-generated code13:50 – Critical thinking in probabilistic systems14:44 – Security vulnerabilities and unmanaged legacy code16:27 – Commodity systems vs unnecessary custom code17:25 – AI models finding security vulnerabilities18:38 – Exploration, testing, and security charters20:02 – Why code liability matters more than ever20:24 – Engineering excellence as competitive advantage20:44 – Final thoughtsResources & References📘 Concepts & People MentionedWard Cunningham — Technical DebtKent Beck — Extreme Programming (XP)Dave Farley — Continuous Delivery & modifiable systemsDan North — Engineering practices & architectureElizabeth Hendrickson — “Testing = Checking + Exploring”📚 Topics DiscussedTechnical debtAI-generated codeAgentic AI workflowsEvaluation frameworks (evals)ObservabilityContinuous verificationSecurity scanningProbabilistic systemsPlatform engineeringServerless architectureEngineering excellenceServerless CrAIc from The Serverless EdgeCheck out our book The Value Flywheel Effect Follow us on X @ServerlessEdgeFollow us on LinkedIn Subscribe on YouTube