Practical DevOps arguments for the AI era.
DevOps Paradox treats software delivery as a working problem, not a slogan. Darren Pope, Victor Farsen, and occasional guests test AI tools, cloud native patterns, and team habits against what actually breaks.
PRESS PLAY
DevOps Paradox is a working-room conversation about modern software delivery, cloud native infrastructure, and the operational consequences of AI. It does not treat DevOps as a polished category. It treats it as the place where code, tools, teams, security, documentation, and late-night incidents keep colliding. Darren Pope and Victor Farsen bring a quick, skeptical rhythm to the show, often starting with a strong claim and then testing it against daily engineering reality. The result is loose but useful. Recent episodes spend a lot of time on AI, but the concern is rarely novelty for its own sake. AI coding assistants, agents, Kiro-style workflows, MCPs, and prompt-driven software all become excuses to ask harder questions about review load, release control, dependency health, API exposure, and whether teams understand the systems they are accelerating. A simple app that uploads files to S3 becomes a conversation about buckets, encryption, transport, certificates, and KMS. A discussion of open source dependencies turns into a question about whether anyone can really assess the health of projects like Kubernetes or curl over a five-year horizon. The show also stays close to Kubernetes and cloud native mechanics. KubeCon EU 2026 leads into arguments about Kubernetes becoming an AI platform, Gateway API replacing Ingress, OCI artifacts as volume sources, and model weights that should not be glued awkwardly into container images. API design gets similar treatment. When AI models start using APIs directly, the show argues, those APIs stop being internal plumbing and become the actual company interface. Documentation takes its turn too, with GitOps and internal knowledge systems examined less as doctrine and more as messy replicas of what is actually running. The tone is often dry, sometimes exasperated, and usually practical. It suits listeners who enjoy technical disagreement when it produces better questions. It is especially strong for engineers, platform teams, DevOps practitioners, and technical leaders trying to separate useful change from fashionable vocabulary.
Made for: For engineers, platform teams, DevOps practitioners, and technical leaders who want blunt, practical conversations about software delivery, AI tooling, Kubernetes, APIs, and operations. It works best for listeners who already know the buzzwords and want them stress-tested.
What sets it apart: The show’s edge is its refusal to let abstract tooling claims stay abstract. AI, Kubernetes, feature flags, documentation, and API strategy are repeatedly brought back to reviews, security, maintainability, releases, and who gets stuck operating the result.
What is DevOps? We will attempt to answer this and many more questions.
Based on 6 episodes we listened to · September 2026
Darren Pope and Victor Farsen make the judgment behind software creation visible, especially where AI, maintenance and operational caution collide.
DevOps Paradox treats software delivery as a chain of human decisions, with Darren Pope and Victor Farsen testing each link through argument. Their conversations keep returning to the gap between producing code and responsibly putting it to work.
Darren and Victor turn AI-era engineering into the argument you wish your change-review meeting could handle.
Press play if you want
Short reviews from the PlayNext desk, based on the episodes we processed.
A grounded discussion of AI, open-source stewardship, and the difficult judgment calls behind dependency health.
DevOps Paradox episode 346 turns AI's arrival in open-source work into a question of stewardship. Darren Pope and Victor Farsen look at what happens when people wait on pull requests, AI forks projects, and maintainers have to guard their time more deliberately.
Kiro's speed feels exciting until the missing security questions and six-hour rabbit holes start to matter.
DevOps Paradox frames Kiro through a familiar but newly accelerated tradeoff: the leap from a chat prompt to a working chess application before the usual long stretch of learning has had time to happen.
A grounded, skeptical read on Kubernetes, AI inference, model storage, and infrastructure that will not simply move aside.
KubeCon EU 2026 turns into a useful lens on where Kubernetes may be heading, with Darren Pope and Victor Farsal tracing its pull toward AI inference, model storage, and the stubborn work of changing systems that already run a lot of production life.
A clear bridge metaphor turns AI's code boom into a practical question about feature flags, review, and release.
DevOps Paradox frames AI as a force that has widened the code highway while leaving review, security, and release capacity as the tighter path to production.
A practical case for winning over cautious operations teams through small, manageable technology changes.
Operations resistance comes across as learned caution, built over six straight weekends of 3 a.m. calls and war rooms after new features went live.
A pointed case for grounding DevOps architecture in Linux literacy and hands-on experience.
Darren and Victor make an unfashionable case for aspiring DevOps architects: learn the ordinary tools before chasing the title. Linux is their proving ground, beginning with commands such as ls, cat and echo, then moving into the Unix habit of linking small tools with pipes.
A practical conversation about making a conference session clear enough to earn attention.
Becoming a conference speaker starts here as a fear problem, then settles into a more useful craft problem: getting a session submission noticed. The conversation is strongest when it shifts from stage anxiety to the small but decisive work of the title.
Power Platform news with practical static.
A discussion on when to build custom software versus using off-the-shelf solutions, featuring guest Hugo Santos at KubeCon.
Conference speaking starts with dread.
Linux is still the floor.
Darren and Victor review KubeCon North America 2024, discussing the shift from new projects to best practices and platform engineering.
Hosts Darren and Victor discuss the challenges of manual dependency updates and introduce Renovate as an automation tool.
What happens to maintainers?
A simple app does not stay simple for long.
Kubernetes keeps getting dragged toward AI.
This episode discusses how AI is making APIs the primary interface for companies, replacing traditional websites, and the challenges that arise from APIs not being designed for direct end-user access.
Victor and Darren argue that most company documentation is useless for AI because it's poorly organized and not the true source of truth.
AI widened the highway.
Operations resistance starts with scars.
In this 300th episode of DevOps Paradox, hosts Darren and Victor discuss with Patrick Dubois how to become an AI native engineer in 2025, sharing personal experiences of building useful prototypes over weekends using AI tools.
This episode discusses how KitOps uses model kits to hand over AI artifacts between data science, DevOps, and application development teams.