A day in the life of a developer

09:00 — The bug

It’s Monday morning. There’s a message from the tester in Teams: “User registration returns a 500 after the last deployment.”

Before, you’d open ChatGPT. You copy RegistrationService.cs into it. And UserRepository.cs, because it depends on that. And the request model, because otherwise it won’t understand the context. Three messages later you ask: “Can you see what’s wrong?” ChatGPT gives a suggestion. You copy it back, adjust it, run the tests manually. Takes you an hour.

Now you open your terminal in the project and type claude. You say:

“The registration endpoint returns a 500 after the last deployment. Here’s the exception:”

You paste the stack trace. Claude Code takes it from there — it reads the relevant files itself, using the exception as a starting point.

Claude Code reads the relevant files itself. It finds the bug: a null reference in the new validation logic. It fixes the code and runs dotnet test. All tests green. Done. It’s 09:20.

11:00 — The brainstorm

A notification system needs to be added. Users want to receive alerts when statuses change. You have an idea, but you want to bounce the approach off someone.

Before, you’d open ChatGPT and type: “How would you set up a notification system in an ASP.NET Core application?” You get a generic answer with example code that has nothing to do with your project. A SignalR example with a NotificationHub that you’d have to completely rewrite to match your structure.

Now you tell Claude Code:

“We want to add a notification system. Users should receive alerts on status changes. How would that fit into our current architecture?”

Claude Code knows your codebase. It sees that you already use an event-based pattern, that there’s already an IEventHandler interface, and suggests building on that. No generic story, but a proposal that fits what’s already there.

14:00 — The refactor

There’s a service that’s grown too large. Everyone knows it: OrderProcessingService.cs, 800 lines, does too much. It’s been on the backlog for weeks to split it up.

Before, you’d block out an afternoon. Manually go through the code, figure out which methods belong together, create new classes, wire everything up, hope you don’t break anything.

Now you say:

“OrderProcessingService is too large. Can you analyze what responsibilities it contains and propose a way to split it up?”

Claude Code reads the service, identifies three clear responsibilities — validation, price calculation, and status management — and proposes a split. After your approval, it executes the refactor, moves the methods, updates the dependency injection in Program.cs, and runs dotnet build and dotnet test to verify everything still works.

16:30 — The documentation

You want to document how the notification system works, for the rest of the team.

Before, you’d postpone this to next week. Or the week after. Or never.

Now you say:

“Write a short technical description of the notification system as we just set it up.”

Claude Code knows the code it just helped build and writes a concise description that matches reality. No generic story, but documentation you can put straight into the wiki.

17:00 — End of day

Four tasks. Bug fixed, feature designed, legacy code refactored, documentation written. Not because you worked harder, but because you spent less time providing context, copy-pasting code, and manually translating generic answers to your situation.

That’s the difference between ChatGPT and Claude Code. ChatGPT is a smart assistant in another tab. Claude Code is a colleague who lives in your project.

Want to get started?

npm install -g @anthropic-ai/claude-code

Open your terminal in your project, type claude, and run /init. Try it for a week. I don’t think you’ll want to go back.