Tweet by jarredsumner
July 15, 2025
I think there’s a workflow here where you combine code coverage + faker-like data and you have a semi-automated fuzzer that fails if the LLM hardcodes fake data or the LLM fails to properly test code. Still need to experiment with this
- Author
- jarredsumner
- Date
- July 15, 2025
- Canonical URL
- /tweets/jarredsumner-1944949936447173042-dbc2f0