I’ve been using Claude Code for about a month on a side project, and to say I’m impressed is an understatement. This feels like the next generation of developer assistant that goes far beyond a coding assistant.
When we think of a stereotypical rock-star developer, it’s not uncommon to imagine someone who works exclusively in the terminal and rarely touches an IDE. Most coding assistants we’ve seen so far work in an IDE, but Claude Code from Anthropic feels like a supercharged developer assistant that works in the terminal.
Because developers do more than write code. And most of the AI assistants aren’t as good at the other development tasks beyond coding. As an example, I noticed that my CloudFront costs were higher than anticipated this month, so I asked Claude to help me figure out why. Sure, I could have gone to the console and eventually figured it out, but as a developer assistant that knows the terminal inside and out, it was able to run some AWS CLI commands to gather the necessary information to help me find an issue in my code that caused excessive invalidations.
Beyond that, my git commit messages are more accurate than ever, and the ability for it to use the GitHub CLI to troubleshoot build failures—without me having to copy and paste or navigate between tabs—has been super productive.
I see companies trying to measure lines of code produced by AI, and that probably is a useful metric, but the potential is so much more than just generating code!