Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I've been using Claude Code for a couple of days

twitter.com

Post reply on HN
Print view View original

31–40 of 507 posts

Re: I've been using Claude Code for a couple of days

#31

Post by Yoric »

branko_d wrote: post #20
develoopest wrote: post #4

I must be the dumbest "prompt engineer" ever, each time I ask an AI to fix or even worse, create something from scratch it rarely returns the right answer and when asked for modification it will struggle even more. All the incredible performance and success stories always come from these Twitter posts, I do find value in asking simple but tedious task like a small refactor or generate commands, but this "AI takes the…

I have the same experience. Where AI shines for me is as a form of a semantic search engine or even a tutor of sorts. I can ask for the information that I need in a relatively complex way, and more often than not it will give me a decent summary and a list of "directions" to follow-up on. If anything, it'll give me proper technical terms, that I can feed into a traditional search engine for more info. But that's neve…

Same here. In some cases, brainstorming even kinda works – I mean, it usually gives very bad responses, but it serves as a good duck.

Code? Nope.

Re: I've been using Claude Code for a couple of days

#32

Post by owenthejumper »

999900000999 wrote: post #9

It’s ok. It’s expensive and is correct for the easy 90%. It messes up the hard 10%. I guess with it rapidly improving, I should wait 3 months. What’s frustrating is when you spend 20$ in credits on it writing non functional code. That said, every programmer needs to at least demo these tools

Feels like you shouldn't be using $20 of credits to produce non functional code.

I use Aider chat with Claude and my sessions are much smaller. You need to ask it much smaller tasks and work incrementally.

Re: I've been using Claude Code for a couple of days

#33

Post by mtlynch »

This is particularly interesting, as Steve Yegge works on (and I think leads) Sourcegraph Cody[0], which is a competitor to Claude Code.

Cody does use Claude Sonnet, so they do have some aligned interests, but it's still surprising to see Yegge speak so glowingly about another product that does what his product is supposed to do.

[0] https://sourcegraph.com/cody

Re: I've been using Claude Code for a couple of days

#34

Post by beezlewax »

abxyz wrote: post #26
develoopest wrote: post #4

I must be the dumbest "prompt engineer" ever, each time I ask an AI to fix or even worse, create something from scratch it rarely returns the right answer and when asked for modification it will struggle even more. All the incredible performance and success stories always come from these Twitter posts, I do find value in asking simple but tedious task like a small refactor or generate commands, but this "AI takes the…

I think it's probably the difference between "code" and "programming". An LLM can produce code and if you're willing to surrender to the LLMs version of whatever it is you ask for, then you can have a great and productive time. If you're opinionated about programming, LLMs fall short. Most people (software engineers, developers, whatever) are not "programmers" they're "coders" which is why they have a positive impres…

I'm waiting for artisan programming to become a thing.

Re: I've been using Claude Code for a couple of days

#36

Post by 999900000999 »

owenthejumper wrote: post #32
999900000999 wrote: post #9

It’s ok. It’s expensive and is correct for the easy 90%. It messes up the hard 10%. I guess with it rapidly improving, I should wait 3 months. What’s frustrating is when you spend 20$ in credits on it writing non functional code. That said, every programmer needs to at least demo these tools

Feels like you shouldn't be using $20 of credits to produce non functional code. I use Aider chat with Claude and my sessions are much smaller. You need to ask it much smaller tasks and work incrementally.

If you already have a moderately complex code base it starts making mistakes.

Re: I've been using Claude Code for a couple of days

#37

Post by noisy_boy »

The trick is not to get sucked into making it do 100% of the task and have a judgement of the sweet spot. Provide it proper details upfront along with the desired overall structure - that should settle in about 10-15 mins of back and forth. This must include tests that you have to review manually - again you will find issues and lose time again (say about 30-45mins). Cut your losses and close the lose ends of the test cide. Now run the tests and start giving it discreet tasks to fix the tests. This is easily 20-40 mins. Now take over and go through the while thing yourself because this is where you will find more issues upon in-depth checking (the LLM has done most of what it could) and this where you must understand the code you need to support.

Re: I've been using Claude Code for a couple of days

#39

Post by moomin »

develoopest wrote: post #4

I must be the dumbest "prompt engineer" ever, each time I ask an AI to fix or even worse, create something from scratch it rarely returns the right answer and when asked for modification it will struggle even more. All the incredible performance and success stories always come from these Twitter posts, I do find value in asking simple but tedious task like a small refactor or generate commands, but this "AI takes the…

I can definitely save time, but I find I need to be very precise about the exact behaviour, a skill I learned as… a regular programmer. Soper up is higher in languages I’m not familiar with, where I know what needs doing but not necessarily the standard way to do it.
Post reply on HN