frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Ask HN: What's a good 3D Printer for sub $1000?

213•lucideng•3d ago•273 comments

Ask HN: Walled garden dwellers: What keeps you there?

6•FlyingAvatar•1h ago•4 comments

Ask HN: How were graphics card drivers programmed back in the 90s?

4•ferguess_k•6h ago•7 comments

We built automated testing for vibe-coded apps

2•MatveyF•4h ago•2 comments

Tell HN: Apple Broke Fitts' Law in Tahoe

30•dmd•11h ago•19 comments

Ask HN: LLM Prompt Engineering

3•Scotrix•10h ago•3 comments

I launched a Mac utility; now there are 5 clones on the App Store using my story

127•tTarnMhrkm•2d ago•132 comments

Ask HN: What Are You Reading?

9•ImPleadThe5th•1d ago•32 comments

Ask HN: What Terminal apps (via homebrew) support 24 bit color on macOS Tahoe?

4•amichail•1d ago•8 comments

Ask HN: How can we reliably determine if text was written by AI?

4•denis_dolya•7h ago•6 comments

Paid $2400 to Cloudflare, support refuses to help

142•thekonqueror•3d ago•29 comments

Ask HN: Generalists, when do you say "I know enough" about any particular topic?

32•AbstractH24•2d ago•85 comments

Ask HN: Dark Mode for HN?

43•todotask2•6h ago•41 comments

Ask HN: How to be ambitious/hungry again?

9•Poomba•20h ago•15 comments

Is the era of personal software portfolios over?

10•justanotherunit•1d ago•9 comments

Ask HN: How to deal with fake job applicants?

16•rswerve•1d ago•24 comments

Ask HN: Is it immoral not to correct someone else's grammar on social media?

2•amichail•1d ago•27 comments

Ask HN: Why isn't capability-based security more common?

12•killerstorm•2d ago•21 comments

Ask HN: Is Claude Code less useful in recent weeks for you?

9•vintagedave•1d ago•11 comments

Advertising in Microsoft Excel

12•BLKNSLVR•1d ago•8 comments

You've reached the end!

Open in hackernews

Ask HN: LLM Prompt Engineering

3•Scotrix•10h ago
I’m working on a project where I need to extract user intents and move them to deterministic tool/function/api executions + afterwards refining/transforming the results by another set of tools. Since gathering the right intent and parameters (there are a lot of subtle differences in potential prompts) is quite challenging I’m using a long consecutive executed list of prompts to fine tune to gather exactly the right pieces of information needed to have somewhat reliable tool executions. I tried this with a bunch of agent frameworks (including langchain/langgraph) but it gets very messy very quickly and this messiness is creating a lot of side effects easily.

So I wonder if there is a tool, approach, anything to keep better control of chains of LLM executions which don’t end up in a messy configuration and/or code execution implementation? Maybe even something more visual, or am I the only struggling with this?

Comments

thekuanysh•9h ago
What kind of IO do you have? JSON or plain language?
Scotrix•8h ago
I input text and preferably I output JSON but doesn’t matter much as long as it’s somewhat structured.

Ultimately I’d like to extract information like date ranges, specific indications of tool usages (e.g. I have a bunch of data apis with their own individual data and semantic meaning which need to be picked and then a combination of tools to transform the data)

zerodayai•6h ago
I am creating something along these lines, https://github.com/zero-day-ai, it's meant for security testing, but probably has most of the functionality you need (and you can write plugins fairly easily if not); you can create a prompt repository, defined by a schema that are organized my domains (again, security testing domains, but they can be expanded). If you have any features you'd like to see, or have an ideal workflow feel free to ping me: anthony@zero-day.ai