Walmart engineer Mike Pfaffenberger developed an internal AI coding assistant called Code Puppy to prevent vendor lock-in and maintain control over the company's software development and infrastructure costs.

Key Points

• Code Puppy is an AI coding agent built on the Pydantic AI library that allows developers to switch between dozens of models from various providers.
• The tool enables Walmart to distribute workloads across multiple AI suppliers, helping the company avoid usage limits and manage fluctuating token costs.
• By remaining platform-agnostic, the software ensures that Walmart’s internal codebases remain independent of any single third-party AI service provider.
• The project was created to mitigate the risks of "enshittification," where platforms become less reliable or more expensive as they prioritize investor returns over user needs.
• Code Puppy has seen widespread adoption across Walmart, with users ranging from engineering teams to store managers utilizing it for automation tasks.

Why it Matters

This initiative highlights a growing trend among large enterprises to prioritize technological sovereignty over reliance on a few dominant AI providers. By building flexible, model-agnostic tools, companies can protect themselves from sudden price hikes, service disruptions, and the long-term risks of vendor dependency.