Softspell

I've been running Claude Code, Cursor, and Codex pretty heavily for the last few months and I keep hitting the same loop:

1. Start a task in one agent

2. Switch to something else (Slack, Twitter, another terminal)

There seems to have two types of developers:

• Human in the loop: Those who like to control the behavior of their agents as it works, looking at the context usage, reading reasoning blocks, and approving individual file edits.

• Agent first: Those who prefer to review the output of agents, rather than individual actions, and run one or more sessions in parallel.

What type of developer are you when working with AI coding agents?

I've been using AI tools for everything lately; writing, coding, design, research. On paper, they should be massive productivity boosters. Instead of spending hours on tasks, I can get decent results in minutes.

But I'm starting to notice a weird pattern. I spend way too much time tweaking prompts, trying different AI tools for the same task, and comparing outputs. Sometimes I'll spend 30 minutes getting the "perfect" AI generated result when I could've just done it myself in 20 minutes.