Live data from Hacker News

My experience with Claude Code after two weeks of adventures

sankalp.bearblog.dev

241–250 of 388 posts

Re: My experience with Claude Code after two weeks of adventures

#241

For me the best part about AI is that when I'm feeling lazy, I can tell the AI to do it. Whether it gives me gold or gives me shit, it doesn't matter, because I have now started my work.

At the end of the day when I'd be thrashing or reverting as much as I write, I can let Jesus take the wheel for a bit to refresh.

On small issues it's usually just a glance over the diff, on harder issues it's not challenging to scold it into a good path once something starts coming together, as long as it's a localized edit with a problem I understand. I'll often take over when it's 40-60% there, which is only possible b/c it does a good job of following todo lists.

I've moved into a good daily workflow that emphasizes my own crafting during the times I'm sharp like mornings, and lets AI do the overtime and grunt work while I prepare for tomorrow or do more thoughtful writing and design.

Re: My experience with Claude Code after two weeks of adventures

#242

For me the best part about AI is that when I'm feeling lazy, I can tell the AI to do it. Whether it gives me gold or gives me shit, it doesn't matter, because I have now started my work.

Even if I want to do it myself, I'll ask it to write it a plan and toss it in a markdown file.

It gave me a shit plan today. Basically I asked it to refactor a block of prototype code that consisted of 40 files or so. It went the wrong way and tried to demolish it from the bottom up instead of making it 100% backwards compatible. If it made a mistake we would have taken forever to debug it.

But yeah, it gave me something to attack. And the plan was fixed within an hour. If I tried to make one myself, I would have frozen from the complexity or gone in circles documenting it.

Re: My experience with Claude Code after two weeks of adventures

#243

Reading all these glowing reviews of Claude Code, I still get the feeling that either everyone’s been paid off or it’s just the die-hard fans of terminal windows and editors like Emacs and Vim. Using the terminal is right up their alley—it’s in their DNA. Every time I read comments saying Claude Code is far better than Cursor, I fire it up, pay for a subscription, and run it on a large, complex TypeScript codebase. F…

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

Re: My experience with Claude Code after two weeks of adventures

#244

Reading all these glowing reviews of Claude Code, I still get the feeling that either everyone’s been paid off or it’s just the die-hard fans of terminal windows and editors like Emacs and Vim. Using the terminal is right up their alley—it’s in their DNA. Every time I read comments saying Claude Code is far better than Cursor, I fire it up, pay for a subscription, and run it on a large, complex TypeScript codebase. F…

And so you don’t think I just dropped in to try it out: no, I’ve been using Cursor every day for many months. I set up the rules very precisely, make maps, write instructions and constraints so the LLM understands how to work with the current codebase—where to find things, what to do, what not to do, what things should look like. I try to write as concisely as possible so everything fits in the context window.

I have a monorepo, so different instructions and rules live in different corners of it, which I manually add as needed. Doing all this with Claude is hard: it kind of works, but it’s a kludge. It’s much easier to manage through the UI. As the saying goes, it’s better to see something once than to remember it ten times, type a command, and configure it.

I can say the same about Vim and Emacs users. No one has yet proven to me they can code and navigate faster than an ordinary programmer using a trackpad or a mouse and keyboard. It’s purely a matter of taste. I’ve noticed that people who love those tools are just bored at work and want to entertain their brains. That’s neither good nor bad, but it doesn’t give a real boost.

By the way, there’s a study (a bit old, admittedly) showing that working with an LLM still doesn’t provide a massive speed-up in programming. Yes, it spits out code quickly, but assembling a finished result takes much longer: you have to review, polish, start over, throw things out, start again. It’s not that simple.

Speaking of Claude, he’s a real master at shitting in the code. At the speed of sound he generates a ton of unnecessary code even when you ask him not to. You ask for something minimalist, nothing extra—he still slaps on a pile of code with useless tests that outsmart themselves and don’t work.

That’s the downside. Overall, it’s easier to program when you just say what to do instead of sitting there straining your own brain. But even then it’s not simple. If you yourself haven’t figured out very carefully what you’re doing, you’ll end up with crap. I almost always throw away what I wrote with an LLM when I was tired. I don’t think there’s ever been an exception.

The only way to write something decent is to dig in thoroughly: draw diagrams, write everything out, and only then tell the LLM exactly what to do. Then double-check, make sure it didn’t write anything extra, delete the excess. And only like that, in small chunks. The “make me a feature, I pressed a button and walked away” mode doesn’t work if the task is even slightly complex.

Yeah, I got carried away—had to let it out. I have a lot to say about LLMs and programming with them, and not just about Cursor and Claude. The tools completely upended the way we program...

Re: My experience with Claude Code after two weeks of adventures

#245

Earlier quoted context omitted.

I think you are perhaps missing the point. Investing into these techniques [2] enables you to do unhinged things. Such as building a compiler whilst you are AFK [1]. [1] https://x.com/i/broadcasts/1OyJALVOnEzGb [2] https://ghuntley.com/ralph

Sure if I want to just toy around for fun. Those are cool, but a production system is infinitely more complex.

What do you define as a production system? Are you aware that one can generate TLA+ specifications, then code generate from these specifications and assert that the implementation matches the TLA+ spec?

Re: My experience with Claude Code after two weeks of adventures

#246
post #243

Reading all these glowing reviews of Claude Code, I still get the feeling that either everyone’s been paid off or it’s just the die-hard fans of terminal windows and editors like Emacs and Vim. Using the terminal is right up their alley—it’s in their DNA. Every time I read comments saying Claude Code is far better than Cursor, I fire it up, pay for a subscription, and run it on a large, complex TypeScript codebase. F…

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…

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.

Re: My experience with Claude Code after two weeks of adventures

#247
post #184
post #68

Earlier quoted context omitted.

I like them for refactoring and “explain this massive codebase please”. Basically polishing or investigating things that already work. But I think we should expect the scope of LLM work to improve rapidly in the next few years. https://metr.org/blog/2025-03-19-measuring-ai-ability-to-com...

The bad news is that mostly, as far as we can see, that doubling of performance also requires (at least) doubling of resource usage, plus we're getting close to a point where planetary resources for doubling LLM resources are getting kind of low...

This species is going extinct. I finally accepted that when my dad died rather than change his lifestyle, despite being warned 10000x. My mom survived a heart attack, saw what happened to my dad, still hasn't changed her lifestyle.

Re: My experience with Claude Code after two weeks of adventures

#248
post #243

Earlier quoted context omitted.

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…

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…

> 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.

Re: My experience with Claude Code after two weeks of adventures

#249

Reading all these glowing reviews of Claude Code, I still get the feeling that either everyone’s been paid off or it’s just the die-hard fans of terminal windows and editors like Emacs and Vim. Using the terminal is right up their alley—it’s in their DNA. Every time I read comments saying Claude Code is far better than Cursor, I fire it up, pay for a subscription, and run it on a large, complex TypeScript codebase. F…

Some of it is taste. Personally, I dislike VSCode and its descendants; it feels bloated and like I’m always being advertised to inside the product. I want my editor to disappear from thought, and Sublime does this for me. In that context, CC is amazing.

A lot of people don't realize that you can completely change the vscode UI. All the side bars in particular - just get rid of them. You're left with a blank screen on which syntax-highlighted code appears with LSP and either the command palette or if you want to be more vim/emacs-y then basically every operation is key-bindable.

Re: My experience with Claude Code after two weeks of adventures

#250
post #171
post #65

PSA: you can use CC with any model via https://github.com/musistudio/claude-code-router The recent Kimi-K2 supposedly works great.

> The recent Kimi-K2 supposedly works great. My own experience is that it is below sonnet and opus 4.0 on capability - but better than gemini 2.5 pro on tool calling. It's really worth trying if you don't want to spend the $100 or $200 per month on Claude Max. I love how succinct the model is. > you can use CC with any model via Anthropic should just open source Claude Code - they're in a position to become the VS Co…

Where do you host Kimi-K2 ?
Post reply on HN