Podman Desktop for AI Lab | RHEL Presents Ep. 81

For episode 81, my cohost Brian Smith and I revisited Podman Desktop, this time to look at its new AI Lab extension, with developer advocate Cedric Clyburn and a product manager from the Podman Desktop team. Before we get into what AI actually is on this show, we wanted to show how you can run it somewhere other than a cloud bill, on your own laptop, in containers. That is exactly what AI Lab is built for.

A few things worth carrying away from the conversation:

• AI Lab runs large language models locally in containers, so your data stays with you. The guests leaned on the privacy angle, feeding sensitive documents through a model without shipping them off to an online service, pulling models from Hugging Face right onto your machine.
• A recipe catalog gives you ready-to-run, Kubernetes-shaped use cases. Recipes cover chatbots, summarizers, code generation, audio-to-text, and computer vision, and Cedric demoed an object-detection recipe running as a pod that was almost certain a photo contained a person.
• Local model serving turns into an app-ready API. Cedric served a model on a local port with a Swagger UI and generated client code in several languages, then closed with a Quarkus and LangChain insurance-claim app doing real prompt engineering against that local model.

If you want to get hands-on with AI without renting a fleet of GPUs, AI Lab on Podman Desktop is a genuinely approachable starting point.