One of the biggest shifts I made to my workflow this week was using parallel AI agents (I've been using Amp, Codex and Claude in Superset).
I've been using AI for a long time but being able to work on multiple things at once and at different levels of complexity is a game changer.
It's also a bit spicy, with great power comes something something, I can see how easy it is to get into a mess quickly since you still have to hold a mental map of whats going on in your head even if it's not the entire implementation detail.
I also still battle with the idea that I am giving up some implementation detail in my head in favour of cadence, but it's working and I feel very productive.
commit 6145eb4
The "holding a mental map" thing is what I am still struggling with but trying to improve... sometimes need to slow down 😅