←back to thread

What makes Claude Code so damn good

(minusx.ai)
469 points samuelstros | 6 comments | 23 Aug 25 19:07 UTC | HN request time: 0s | source | bottom
Show context
1zael ◴[23 Aug 25 20:46 UTC] No.44998995[source]▶
I've literally built the entire MVP of my startup on Claude Code and now have paying customers. I've got an existential worry that I'm going to have a SEV incident that will trigger a house of falling cards, but until then I'm constantly leveraging Claude for fixing security vulnerabilities, implementing test-driven-development, and planning out the software architecture in accordance with my long-term product roadmap. I hope this story becomes more and more common as time passes.
replies(6): >>44999056 #>>44999088 #>>44999121 #>>44999144 #>>44999151 #>>45001583 #
foobarbecue ◴[23 Aug 25 20:58 UTC] No.44999088[source]▶
[flagged]
replies(5): >>44999101 #>>44999317 #>>45001279 #>>45001766 #>>45077840 #
BeetleB ◴[23 Aug 25 21:33 UTC] No.44999317[source]▶
> I bet you could replace your slop pile with a tenth of the lines of clean code, and chances are it'd be less work than you think.

Actually, no. When LLMs produce good, working code, it also tends to be efficient (in terms of lines, etc).

May vary with language and domain, though.

replies(1): >>44999411 #
stavros ◴[23 Aug 25 21:49 UTC] No.44999411[source]▶
Eh, when is that, though? I'm always worrying about the bugs that I haven't noticed if I don't review the changes. The other day, I gave it a four-step algorithm to implement, and it skipped three of the steps because it didn't think they were necessary (they were).
replies(1): >>44999429 #
BeetleB ◴[23 Aug 25 21:52 UTC] No.44999429[source]▶
Hmm...

It may be the size of the changes you're asking for. I tend to micromanage it. I don't know your algorithm, but if it's complex enough, I may have done 4 separate prompts - one for each step.

replies(2): >>44999450 #>>44999451 #
1. foobarbecue ◴[23 Aug 25 21:55 UTC] No.44999450{3}[source]▶
Isn't it easier to just write the code???
replies(1): >>44999679 #
2. BeetleB ◴[23 Aug 25 22:37 UTC] No.44999679[source]▶
Depends on the algorithm. When you've been coding for a few decades, you really, really don't want to write yet another trivial algorithm you've written multiple tens of times in your life. There's no joy in it.

Let the LLM do the boring stuff, and focus on writing the fun stuff.

Also, setting up logging in Python is never fun.

replies(2): >>45000538 #>>45001453 #
3. foobarbecue ◴[24 Aug 25 01:34 UTC] No.45000538[source]▶
Right-- it's only really capable of trivial code and boilerplate, which I usually just copy from one of my older programs, examples in docs, or a highly-ranked recent SO answer. Saves me from having to converse with an expensive chatbot, and I don't have to worry about random hallucinations.

If it's a new, non-trivial algorithm, I enjoy writing it.

replies(1): >>45001623 #
4. a5c11 ◴[24 Aug 25 04:41 UTC] No.45001453[source]▶
> Also, setting up logging in Python is never fun.

import logging

replies(1): >>45001617 #
5. BeetleB ◴[24 Aug 25 05:21 UTC] No.45001617{3}[source]▶
Not fun at all.

Configuring it to produce useful stuff (e.g. timestamps, autologging exceptions, etc). Very boilerplate and tedious.

6. BeetleB ◴[24 Aug 25 05:23 UTC] No.45001623{3}[source]▶
For me, it's a lot easier getting the LLM to do it than browsing through multiple SO answers, or even finding some old code of mine.

Oh, and the chatbot is cheap. I pay for API usage. On average I'm paying less than $5 per month.

> and I don't have to worry about random hallucinations.

For boilerplate code, I don't think I've ever had to fix anything. It's always worked the first time. If it didn't, my prompt was at fault.

replies(1): >>45015583 #