[Oni]

One of my biggest uses for AI has been in-depth QA across the 100-plus network environments hosted on my dedicated server. It helps me review configurations, validate lab workflows, identify inconsistencies, troubleshoot failures, and repeat tests much faster than I could manually. I use it moderately during the day, but I run it overnight every night to continuously do QA over the server.

I'm curious, did you use it much for visual QA? While I found AI to be excellent for code, I didn't find it to be as good when it came to visual QA. Namely when comparing mockups to staging sites.

I'm sure it's advanced a bit since I last tried and I only tested out Codex and Claude.