Ep. #39, Agents Take the Wheel with Zach Smith

On episode 46 of Open Source Ready, Brian Douglas and John McBride sit down with JT Archie. When an agent skips a command or grabs the wrong branch, a better prompt may only get you so far. JT explains how Steps uses structured pipelines to give agents clearer inputs and boundaries, then joins the hosts to explore SQLite, tool permissions, and the limits of fully automated development.

On episode 45 of Open Source Ready, Brian Douglas and John McBride speak with Dotta about Paperclip, an open-source platform for managing teams of AI agents. They explore what happens when agents produce more work than humans can review, why permissions need to extend beyond a separate computer, and how engineering judgment changes when writing code becomes easier.

On episode 44 of Open Source Ready, Brian Douglas and John McBride sit down with Mikyo King to explore why an agent can fail even when its model gets things right. They discuss observing the full agent workflow, using evaluations to uncover unexpected behavior, and testing whether skills actually improve results. Along the way, Mikyo shares how Arize approaches open source collaboration as coding agents reshape the contribution process.

On episode 43 of Open Source Ready, Brian Douglas and John McBride speak with Madelyn Olson. The Valkey maintainer and AWS principal engineer explains how AI is changing open source contributions, code review, and the work of keeping shared infrastructure reliable. They also explore Valkey’s origins, semantic caching, and why community governance still matters when anyone can customize the code.

On episode 42 of Open Source Ready, Brian Douglas and John McBride sit down with Will Cory, creator of Smithers. Will explains how workflows can simplify agent orchestration and why AI-assisted development could make specialized open source forks more practical. They also explore his planned TypeScript redesign, the trade-offs behind agent tooling, and what maintainers can learn from users who build their own solutions.

On episode 41 of Open Source Ready, Brian Douglas and John McBride sit down with Phil Estes. They explore why the definition of a container remains surprisingly fuzzy, how microVM-backed sandboxes could support the next generation of AI agents, and what AI-generated contributions mean for projects like containerd. Phil also explains why the future of open source depends as much on trusted maintainers and healthy communities as it does on code.

On episode 40 of Open Source Ready, Brian Douglas and John McBride sit down with Orhun Parmaksiz to explore Ratty, his experimental GPU-rendered terminal emulator that combines traditional command-line workflows with interactive 3D graphics. They discuss the future of terminal user interfaces, how AI is accelerating developer tooling, and what recent supply chain attacks on Arch Linux reveal about the evolving security landscape for open source software.

On episode 39 of Open Source Ready, Brian Douglas and John McBride speak with Zach Smith, creator of Kplane, about rethinking Kubernetes for an AI-driven future. Zach explains how virtualized control planes could enable isolated cluster experiences at massive scale, while the conversation explores developer experience, AI tooling, and the possibility that future AI agents may each require their own cloud.

On episode 38 of Open Source Ready, Brian Douglas and John McBride speak with Graham Christensen, CEO of Determinate Systems, about the evolution of the Nix ecosystem and why more organizations are embracing reproducible infrastructure. They discuss secure package management, enterprise adoption challenges, open source business models, and how AI tooling is rapidly reshaping software engineering workflows.

On episode 37 of Open Source Ready, Brian and John speak with Stormy Peters about the evolving relationship between AI and open source software. Together, they unpack the growing challenges maintainers face, why traditional “good first issues” may be disappearing, and how AI tools are changing the way developers contribute to projects. They also discuss open-weight models, inference costs, and why community health still matters more than lines of code.