My daily workflow involves using competing AI assistants to build better code through creative tension.
Most mornings, I start with two windows and two assistants to tackle a frustrating problem with the Stark Insider server. On the left, I have VS Code with Claude Code, which is my go-to for commit hygiene and helpful suggestions for succinct titles and explanations.
On the other hand, I use Cursor, which is built on VS Code but prioritizes AI functionality over editing capabilities. I switch between models like GPT-5 Codex, Auto, or other options to get a unique perspective.
People whisper that some of the “stealth” options map to bigger-name models (New version: Gemini 3???!).
Author's summary: Using two AI assistants improves code quality.