Ask HN: What hacks/tips do you use to make AI work better for you?

123 points · rupi · 28 days ago

I’ve found that breaking down a task into smaller parts before asking Cursor to help with each one makes a huge difference. Instead of tackling a big task all at once, guiding it step-by-step leads to much better results.

I also use ChatGPT to break the task into smaller tasks.

What’s a workflow tip or strategy you’ve picked up that others might benefit from?

101 comments

rongenre · 28 days ago

There's a couple uses cases (beyond the obvious) that I like with the chatbots

1. Brainstorming building something. Tell it what you're working on, add a paragraph or two of how you might build it, and ask it to give you pros and cons and ways to improve. Especially if it's mostly a well-trod design it can be helpful.

2. Treating it like a coach - tell it what you've done and need to get done, include any feedback you've had, and ask it for suggestions. This particularly helps when you're some kind of neurospicy and "regular human" responses sort of escape you.

Show replies

yen223 · 28 days ago

I like using LLMs for building non-critical tools that make me more productive. Things like shell scripts, Github actions, or one-off tools for visualising some problem space. The kind of thing where code quality doesn't really matter, but which will save you a ton of time in the long run

Show replies

wruza · 26 days ago

I use local LLMs and write system prompts.

For example, I have two “characters” for sh and cmd, who only produce valid shell commands or “# unable to generate”. No explanations, no important to remember crap. Just commands. I prompted them along the lines of “the output is connected to another program and can never be read by a human, no point in all that”.

Another useful (to me) character is Bob, a grumpy old developer who won’t disrespect a fellow dev with pointless explanations, unless a topic is real tricky or there’s a nuance to it. Sometimes he just says things like “why you need that?” or “sure, what suits you best” instead of pages of handwavy pros and cons and lists and warnings and whatever. Saves a whole lot of skimming through a regular chatgpt babbling, feels like a real busy person who knows things.

Another character (that I rarely use) is BuilderBot. It’s a generic bot “that builds things and isn’t afraid of dumping whole projects into the chat” or something like that. It really dumps code, most of the time. Not just snippets, whole files and folders.

I’m using “text generation web ui” frontend.

liuliu · 28 days ago

For more complex tasks, asking it to write in a more comfortable language for that task and then translate to target language helps. Example: if you ask Claude to write code that generates UDP datagram for mDNS in Swift, it will fail flat. But if you ask for C and translate, it can succeed.

BTW, Python is usually the more comfortable language.

ArtRichards · 28 days ago

In Cursor I try to mention specifically and only the files where the changes need to be made.

Also, I use HomeAssistant for my Dreame vacuum, Hue lights, electricity monitoring, all hooked up with a chat GPT plugin and TTS STT... its the default assistant on my phone and watch!

Over a year ago it was the only way to get GPT assistant, but now I prefer it :) i can customize it as needed through 'homeass'