ChatGPT, Coding Agents and Workflows Are Different Tools

I used to think about AI development tools mostly in terms of which model was better.

Now I find it more useful to think about what kind of work I am trying to do.

ChatGPT is still where I like to think. I can discuss an architecture, challenge an idea, explore alternatives, understand a technology or turn a vague thought into something concrete.

A coding agent is different. Once I know what I want, it can inspect the repository, modify files, run tests and carry a bounded engineering task through to completion.

Then there are workflows and automations. These are useful when something should happen repeatedly without me starting the same conversation every time: checking something, processing incoming data, generating a report or reacting to an event.

They overlap, but I do not see them as interchangeable.

My rough mental model is:

conversation → think and decide

coding agent → change the software

workflow → repeat a defined process

Sometimes the same task moves through all three. I might first discuss an idea in ChatGPT, turn the result into a precise implementation task for a coding agent, and eventually automate part of the finished feature.

The important change for me is that I no longer try to force one AI interface to do everything.

Choosing the right mode of interaction often matters more than choosing between two models that are both already capable enough.