[giantg2]

We started with GitHub Copilot, but it was largely replaced by Claude Code. The major benefit was that Claude could execute scripts and Copilot could not. That might have just been the way my organization configured it though.

[FloatArtifact]

Claude really utilizes tools or develops its own tools well. If OpenAI could get their sandbox situation ironed out... I could switch.