Stenobird

Episode

#472 Monorepos

Podcast
Python Bytes
Published
Mar 9, 2026
Duration seconds
1732
Processing state
processed
Canonical source
https://pythonbytes.fm/episodes/show/472/monorepos
Audio
https://pythonbytes.fm/episodes/download/472/monorepos.mp3
JSON
/v1/public/podcasts/python-bytes/episodes/472-monorepos
Markdown
/podcast/python-bytes/472-monorepos.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/python-bytes/episodes/472-monorepos/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/python-bytes/472-monorepos.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Explore the practicalities of managing Python monorepos using uv workspaces and the evolving landscape of AI-driven development. The episode covers everything from structural dependency management to the hidden costs of replacing SaaS with agentic workflows.

Topics

  • Python
  • uv workspaces
  • monorepos
  • FastAPI
  • VS Code
  • AI coding agents
  • software engineering
  • cattrs

Highlights

  • Main idea: Using uv workspaces requires a distinct top-level project name to avoid package name conflicts
  • Practical takeaway: Use the new FastAPI VS Code extension to navigate routes and functions directly via the command palette
  • Failure mode: Relying solely on AI for coding can bypass the essential learning of mental models and fundamental complexity management
  • Practical takeaway: Leverage AI as a conversational partner to understand why specific code patterns are preferred over just getting a working answer
  • Main idea: The shift toward agentic programming (like OpenClaw) may trade predictable SaaS subscription costs for unpredictable API and maintenance overhead

Chapters

  1. 1:05 Configuring uv Workspaces: Lessons learned from setting up a Python monorepo using uv, specifically handling top-level project names and dependency resolution.
  2. 7:20 Data Validation with Cattrs: A look at using cattrs for efficient structural validation and conversion between unstructured data and Python classes.
  3. 12:00 Learning to Code in the AI Age: How to use AI agents as tutors to build mental models and understand complex programming concepts without skipping the hard work.
  4. 16:15 FastAPI VS Code Extension: A walkthrough of the new official extension that allows developers to navigate API routes and function definitions via the command palette.
  5. 20:40 Agentic Programming Tools: Discussing the potential of tools like Claude Code and the Agent Client Protocol for interacting with development agents.
  6. 26:55 The Cost of 'SaaS is Dead': A critique of the idea that AI agents replace all software subscriptions, highlighting the rising costs of API usage and manual maintenance.