Hacker News Reader

Comment by verdverm | Hacker News Reader

The night and day difference after adopting Claude code prompts in my custom agent leads me to the opposite conclusion.

Everyone is not getting the same results. This is evident in the wide reporting in usefulness. Some people are producing production code while others claim they can't get the AI to to even basic things without error.

Something is def different. If we then look to human history and tool usage, never has there been a tool that is equally wielded by all members of the species. It's not magic, it's a tool. You know two plumbers can produce very different outcomes with the same tools, why do people expect developers with coding agent tools to exhibit different human-tool outcomes?

View on HN

That's a valid point - I largely agree with you, I can make it generate mostly corrected & acceptable code with a solid level of quality, given smaller & scoped features and detailed prompts. I'm just not at all convinced that it's a net productivity boost - you must take your time for these detailed prompts and then verify the output, which is less of the case when you write the thing from scratch.

It certainly speed up some things, slows down others; for learning - a great resource! For generating code I'm on the fence, still experimenting; for now, I write some code manually, some with Claude, working on the hybrid setup. My intuition tells my that a flexible use of this tool will prove to be the most optimal - writing some code manually, some with LLms, depending on both the task and the programmer knowledge, experience and skills.

my daily prompts are very short, I don't have to be as detailed

my system prompts and agents.md are where I put the effort in

we all have different experiences with our tools, even before Ai, there was a difference in developer productivity, gonna always be like that

My experience has been a clear gains in both the short & long term, I don't vibe code because that is lunacy

I think it’s mostly the difference between breaking things down into small focused parts .. and people who are trying to accomplish many features with a single prompt and just expecting the ai to read their mind.

Reply on news.ycombinator.com