AI Impact on DevOps and CI/CD!
My organization recently gave us access to codex, claude and gemini pro to try and evaluate all for the daily workflows on both engineering and DevOps side. With a couple of weeks into it, here is my take as a DevOps Engineer-
Codex - Amazing at long running tasks. Handle huge context decently (When working with multiple repos). Skills come handy when you want to offload mundane stuff.
Gemini - Great at doing research, grasping errors from screenshots and working with google ecosystem. Have mostly used this with the Google's Anti-gravity IDE and sadly it not the best out there. The agent often fails with error on high load and needs to be nudged again and again. The auto-completion though works amazingly even across the files.
Claude - Great capabilities, unmatched results. Writes a very clean and modular code. Claude in chrome is amazing to troubleshoot pipeline running in Github and Gitlab (Waiting for the official plugins to move to GA). Only limitation - the amount of tokens it burns is insane.
As a DevOps engineer one of my primary duties is to build CI/CD pipelines which earlier used to take me a couple of hours and can now be completed in minutes(developed, tested, shipped) using AI tools.
My questions is -
1. How deep is AI adoption in you org. in a high trust domain such as CI/CD?
2. As an DevOps engineer how to keep yourself ahead of the curve so that you are not replaced by AI someday?