Meditation is great, but there might also be other options available when it comes to dealing with the effects of context switching.

For me, I find the Pomodoro technique really effective, but there's a key caveat, which is that you try to focus on one thing at a time (which flies in the face of using multiple AI agents working on x number of things at the same time).

Another angle to explore is how much of the process of software development that we do manually can we automate, particularly the parts that still require human input (like code reviews). That may also help with reducing the cognitive load.

▲

jplusequalt 4 days ago | parent [-]

>Another angle to explore is how much of the process of software development that we do manually can we automate, particularly the parts that still require human input (like code reviews). That may also help with reducing the cognitive load.

Get your vibe coded software with zero human reviewers away from the public infrastructure.

	▲	paulbjensen 3 days ago \| parent [-]
		I’m talking about CI pipelines where you have: - Unit tests - End to End testing with Cucumber and Puppeteer (or Playwright these days) - Code Coverage analysis - Lint and formatting checks - Tools like knit to remove unused code - Security checking tools like Synk and Socket - Size checking If you look at what those tools do, they automate aspects of reviewing PRs. The question then becomes - what still gets done manually, and are there ways to automate that.