Hacker News Reader

The capabilities of the tool matter more. Claude Code, Codex, Cursor CLI all have different feature sets. This usually determines the choice more than base model capabilities.

View on HN

You think so? What features are differentiated enough to warrant that?

Seems like they all have tools for reading files, editing, running shell commands.

Cursor has linting access, but this can be added to claude with hooks I think.

Subagent personalities, skills, structured output generation, customizable context cleanup, etc.

Most of those seem kind of useless... I probably do not know how to use them correctly, so, what would you recommend by way of subagents and skills?

They felt similar enough to me. One feature that did make a difference to me is when follow up prompts are acted on; sooner or later.

If all you are doing is chatting with the model, they're effectively the same. If you are actually building agentic workflows, they differ immensely in supported features. Codex doesn't support subagent personalities, for example, and until this week Claude Code didn't support generating structured responses.

Reply on news.ycombinator.com