Talk Python To Me | Podcast on Spotify

Talk Python To Me: Recent Episode Highlights

This summary covers the latest episodes of the Talk Python To Me podcast, featuring technical deep dives into Python infrastructure, AI, and security.

• Episode #548: Event Sourcing in Python

  • Guest: Chris May.
  • Core Topic: Implementing event sourcing (storing state as an immutable sequence of events) in Python.
  • Key Takeaways: Discusses core patterns, library recommendations, and why event sourcing is effective for AI-assisted coding.
  • Mentioned Tools: eventsourcing package, valkey, diskcache, and Datastar.

• Episode #547: Scaling AI with Ray

  • Guests: Edward Oakes and Richard Liaw (Anyscale).
  • Core Topic: The evolution of Ray, a distributed execution engine for AI workloads.
  • Key Takeaways: Traces Ray’s history from UC Berkeley’s RISE Lab to its role in training large-scale models (like GPT-3). Covers Ray Data, KubRay, and how it compares to Dask and standard Python multiprocessing.

• Episode #546: The State of Self-Hosting (2026)

  • Guest: Alex Kretzschmar (Tailscale, Linuxserver.io).
  • Core Topic: Reclaiming digital privacy by self-hosting services.
  • Key Takeaways: Discusses the "enshittification" of cloud services, recommended apps (Immich, Home Assistant), and the role of Docker Compose and Tailscale in managing private infrastructure without opening ports.

• Episode #545: OWASP Top 10 (2025 Update)

  • Guest: Tanya Janca.
  • Core Topic: A breakdown of the updated OWASP Top 10 security risks.
  • Key Takeaways: Focuses on new categories like supply chain attacks and exceptional condition handling. Includes a practical demonstration using Claude Code to audit an open-source project.

• Episode #544: Wheel Next (Python Packaging)

  • Guests: Jonathan Dekhtiar (NVIDIA), Ralf Gommers (Quansight), and Charlie Marsh (Astral).
  • Core Topic: Addressing the limitations of current Python "wheels" regarding hardware optimization (e.g., AVX2, GPU support).
  • Key Takeaways: Introduces "Wheel Next," a set of PEPs (including PEP 817 and 825) designed to allow packages to declare hardware requirements, enabling installers like uv to automatically select the optimal build.