radar

ONE Sentinel

smart_toyAI/PROMPT ENGINEERING

We Rewrote JSONata with AI in a Day, Saved $500K/Year

sourceSimon Willison
calendar_todayMarch 27, 2026
schedule2 min read
lightbulb

EXECUTIVE SUMMARY

Transforming JSONata with AI: A $500K Savings in Just One Day

Summary

The article discusses the successful rewrite of JSONata using AI, resulting in significant cost savings and efficiency improvements. The project showcased the benefits of vibe porting and leveraging existing test suites for rapid development.

Key Points

  • The JSONata expression language was rewritten in Go, a programming language known for its efficiency.
  • The project was completed in just 7 hours with a token spend of $400.
  • The initiative is part of a trend known as vibe porting, which emphasizes rapid adaptation of existing technologies.
  • The Reco team conducted a shadow deployment for one week to ensure the new Go implementation matched the original JSONata's behavior.
  • The project resulted in an estimated annual savings of $500,000.
  • JSONata is heavily associated with the Node-RED platform, which is used for wiring together hardware devices, APIs, and online services.

Analysis

This case study highlights the potential for AI-driven development to streamline processes and reduce costs in software engineering. By utilizing existing resources like test suites, teams can achieve rapid results while maintaining quality and functionality.

Conclusion

IT professionals should consider adopting vibe porting strategies and leveraging AI tools to enhance development efficiency and reduce operational costs. Emphasizing thorough testing during transitions can ensure reliability in new implementations.