Episode Details

Daniel has a working podcast pipeline: a language model writes scripts, retrieval tools feed it, and text-to-speech reads it out. But the occasional Hebrew word keeps breaking things, and the pipeline itself is trapped in his head. This episode digs into two questions at once — how to handle Hebrew code-switching in TTS, and how to package a chained AI workflow so it survives past the project that spawned it. We talk through script-based segmentation versus small classifiers, why unvocalized Hebrew makes TTS inherently unstable, the tradeoff between chaining small models and using one large one, and why containerization is the answer to dependency hell. Plus: the "recipe" as a distinct layer between a model and a project, and a pre-flight risk score that flags bad inputs before you burn compute. Episode #089350 — open it directly at myweirdprompts.com/089350