frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

Ask HN: How are you using LLMs?

3•FailMore•1d ago
Me: I don’t like to have factual based conversations with LLMs (e.g. can my cat eat cooked chicken on the bone?). I like having open ended conversations with LLMs where being vaguely in the right direction will be useful (e.g. practicing Chinese with ChatGPT voice mode).

I use claude code, and verge between it being useful (because my memory of specific functions is average), but then a bit bummed out when I have to spend a bunch of time cleaning up poor logic/flows in the application.

Curious to know the intricacies of how you are interacting with LLMs too.

Comments

efortis•1d ago
If the problem is simple enough for top-down design, I write signatures along with comments with i/o examples.

But if I need to explore stuff, I ask for some example and then re-prompt it top-down.

silentpuck•1d ago
I use LLMs mostly for learning and understanding.

When a book doesn’t explain something clearly, I ask for a deeper explanation — with examples, and sometimes exercises.

It’s like having a quiet teacher nearby who never gets frustrated if I don’t get it right away. No magic. Just thinking.

I also started building my own terminal-based GPT client (in C, of course). That’s a whole journey in itself — and it’s only just begun.

stephenr•1d ago
I use the terms "LLM" or "AI" (as in, "I used an LLM/AI to write a <insert task> helper") as a quick hint to ignore articles/links/etc in the same way I've previously use the terms "You won't believe what happened next" or "they hate this one trick" to avoid spam bait article links, or "shocked face overlay" to avoid bullshit YouTube videos.

So, thankyou for that AI techbros. Keep telling us loudly and proudly that you're using "AI" to write your slop, it makes it much easier to know what to avoid when skimming titles.

atleastoptimal•1d ago
LLM's are best when you know exactly how to implement something and can describe it fully, but it would take longer to actually write everything yourself. They're also good at rigorous attention to detail in domains that are well-established and the rules are deterministic and not subtle.
TXTOS•17h ago
i mostly use LLMs inside a reasoning shell i built — like a lightweight semantic OS where every input gets recorded as a logic node (with ΔS and λ_observe vectors) and stitched into a persistent memory tree.

it solved a bunch of silent failures i kept running into with tools like RAG and longform chaining:

    drift across hops (multi-step collapse)

    hallucination on high-similarity chunks

    forgetting prior semantic commitments across calls
the shell is plain-text only (no install), MIT licensed, and backed by tesseract.js’s creator. i’ll drop the link if anyone’s curious — not pushing, just realized most people don’t know this class of tools exists yet.

I launched 17 side projects. Result? I'm rich in expired domains

345•cesargstn•1d ago•241 comments

Nova: A New Web Framework for Erlang

61•taure•18h ago•23 comments

Has any YC founder ever gone to jail for startup-related crimes?

7•TeslaK20•5h ago•3 comments

New budget financial API, based on EDGAR data

7•jgfriedman1999•10h ago•2 comments

How Remote Work Died: A Girardian Tragedy in Corporate America

18•la_joconde•19h ago•8 comments

Claude Code weekly rate limits

596•thebestmoshe•3d ago•702 comments

Has AI coding gone too far? I feel like I'm losing control of my own projects

12•Shaun0•1d ago•10 comments

Ask HN: What are you working on? (July 2025)

257•david927•4d ago•843 comments

Ask HN: Small Utility App Monetization

4•mywacaday•15h ago•0 comments

Ask HN: Local LLM agents on Jetson/RPi without a heavy runtime

3•takuya_h•15h ago•3 comments

Ask HN: Advise for technical solo founders trying to secure venture capital?

5•siva7•16h ago•4 comments

Ask HN: Catching Up with Current Datacenters

2•Damogran6•17h ago•7 comments

Ask HN: State of the art with local LLMs and agents

3•prmph•17h ago•2 comments

Ask HN: Tell me your horror story about blowing money on AI

6•witnessme•13h ago•2 comments

Ask HN: Are developers sad about AI writing more of their code?

12•JFerreol_J•19h ago•17 comments

Tell HN: Add "NSFW" words in your Google query to avoid AI summary

22•behnamoh•2d ago•19 comments

Ask HN: How will the OSA affect small Mastodon instances?

29•Digit-Al•2d ago•15 comments

Google Maps Reviews in Germany Are Basically Dead

31•tahaygun•1d ago•17 comments

Tell HN: Google Denies OS Support for Pixel from Non-Licenced Retailer

9•riedel•1d ago•0 comments

Ask HN: What do you do with all your unused tech "swag"?

8•charliebwrites•1d ago•19 comments

Warp.dev Terminal – Overpriced, Buggy, and AI-Sabotaged My Code

56•MistermanX•4d ago•39 comments

Ask HN: Should I pay for marketing to this B2C SaaS product? (~8% paying rate)

3•y2236li•1d ago•4 comments

What I have learned about startups from building my own

3•Taikhoom10•1d ago•5 comments

Drafting Software Recommendation

16•morpheos137•1w ago•16 comments

Ask HN: How are you using LLMs?

3•FailMore•1d ago•5 comments

Ask HN: What is the web template everyone uses for startups?

4•Norcim133•1d ago•4 comments

Ask HN: Whats your best workflows to draft legal agreements without lawyers?

6•philippb•2d ago•13 comments

An app that schedules your most important tasks around your peak energy levels

7•rashempandit48•2d ago•13 comments

Ask HN: Is university still credible?

13•mettamage•1d ago•30 comments

Ask HN: What's your experience been with Claude Code subagents?

11•mbm•1d ago•0 comments