Hacker News Reader

Comment by nico | Hacker News Reader

I used to feel similarly. But recently started using Claude Code and it does feel a lot better than Cursor for me

I'm not sure why. However, Claude does seem to know better where things are and knows not to make unnecessary changes. I still need to guide it and tell it to do things differently some times, but it feels like it's a lot more effective

Personally, I also like that usually it just presents to me only one change/file at a time, so it's easier for me to review. Cursor might open several files at once, each with a tons of changes, which makes it a lot harder for me to understand quickly

Btw, I use Claude Code in a terminal pane inside VSCode, with the extension. So Claude does open a file tab with proposed changes

View on HN

> Personally, I also like that usually it just presents to me only one change/file at a time, so it's easier for me to review

This is interesting. I haven't used Cursor, but one of my frustrations with Claude Code is that some of the individual changes it asks me to approve are too small for me to make a decision. There are cases where I almost denied a change initially, then realized Claude's approach made sense once I saw the full change set. Conversely, there are cases where I definitely should have stopped Claude earlier.

It doesn't help that Claude usually describes its changes after it has made the full series, instead of before.

...really, what I'd like is an easy way to go back in time, wherein going back to an earlier point in the conversation also reverted the state of the code. I can and do simulate this with git to some extent, but I'd prefer it as a layer on top of git. I want to use git to track other things.

> some of the individual changes it asks me to approve are too small for me to make a decision. There are cases where I almost denied a change initially, then realized Claude's approach made sense once I saw the full change set

Yes, I've adapted to just review quickly, then if it makes sense as part of the task, let it keep going until it's done with the whole thing. Most of the times, by the end it does the right thing

I love that it doesn't auto-commit everything, ala aider, so it's pretty painless to undo stuff

I also keep a TODO.md file with a plan of everything I want to do for the current ticket/PR. I tell CC to keep track of things there. CC takes stuff from there, breaks it down into its own subset of tasks and when finished I tell it to update the TODO.md with the progress. I also tell it to stage files and create commits

The way I use it, it feels like I'm still programming, but I don't need to write code or run commands by myself, nor get stuck googling stuff up. I can just tell CC to do almost anything for me. It takes away the tediousness of what I want to accomplish

> I love that it doesn't auto-commit everything, ala aider, so it's pretty painless to undo stuff.

Yeah, I'm definitely glad it doesn't commit for me. The main issue I have is that I'm never sure how granular to make my commits. Sometimes I make them very granular because I'm experimenting with Claude and I want to be able to revert to any point in the conversation—but now I have to write a message each time to keep track of which is which. Conversely, when I don't make the commits as granular I loose the ability to roll back, and sometimes regret it.

Also, sometimes Claude gets a bit too smart! Let's say I decide I want Claude to try again with a slightly different prompt. I save my current changes in a branch, roll back, to the previous state, and ask Claude to try again. Sometimes Claude will say "I see this is already implemented in the XX branch. Let me continue to build on that implementation."

Yeah, sometimes I love it when it checks the git log and properly finds reference code for what I’m trying to implement, but other times I really want it to just do it differently and can get annoying

Other times I’ll tell it about an issue that I want to solve and it will come up with a solution I don’t want. I’ll tell it to take a different approach and it will listen for a bit, then all of a sudden just try to go back to its first approach and I need to steer it again, multiple times even

> but now I have to write a message each time to keep track of which is which

I ask it to write my commits. Usually it’s also pretty smart about which files to include based on the most recently addressed tasks. I have it run git add and git commit under my supervision

A repo Im working on, has some rather annoying hooks that check linting when committing (and aggressively modify the files to fix formatting). If I forget to manually check before committing, then I end up with a “wrong” commit containing the incorrectly formatted files, and a bunch of uncommitted files with the formatting changes. CC most of the times will see the error messages, and automatically propose the proper commands to run for undoing or amending the commit to include the formatting changes

My commit messages have always been crap, so this doesn’t bother me much. I squash it all at the end anyway.

Still sometimes I forget to commit between tasks or it goes wild fixing compiler errors I didn’t want to address.

If you like Claude Code but either (1) prefer an agent that doesn't ask for review on each file edit or (2) miss the IDE for things like reviewing diffs, I'd humbly submit you try out Amp: https://ampcode.com. It has both a CLI and VS Code extension, and we built it from the ground up for agentic coding, so no asking for permission on each edit, a first-class editor extension (personally I spend more and more time reviewing diffs and VS Code's diff view is great), and it employs subagents for codebase search and extended thinking (using a combo of Sonnet and o3) to maximize use of the context window.

Thank you for the suggestion. Do you guys also have subscription plans? Or do I need to pay separately for the models/apis I use?

I rarely had Cursor do unwanted changes for me. Maybe this is about prompting? I am very particular with what I want and I try to be as verbose as I can and explain the context as well plus hint which files I believe it should put attention to.

How can you prove you’re not a bot? I get the feeling I’ve seen a comment like this somewhere before.

Tell us what you do—how, under what conditions, and in which programming languages. What exactly makes it better? Does it just search for files? Well, that’s hardly objective.

So far the impression is… well, you say it only seems better. But you won’t get far with that kind of reasoning.

Objectively, it now seems to me that Claude the cat is better, because everyone around says it’s better. Yet no one has actually explained why. So the hype is inflated out of thin air, and there are no objective reasons for it.

> How can you prove you’re not a bot? I get the feeling I’ve seen a comment like this somewhere before.

...you're replying to an account that was created in 2007. A bio is listed on the profile page.

Maybe you feel you've seen this comment before because it's an opinion lots of people share? Even if you do not.

There isn’t some sort of conspiracy here, man. Anthropic gets paid either way.

Reply on news.ycombinator.com