frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Ask HN: Favorite prompts for improving LLM output?

2•maxutility•1h ago
I use Claude Code a lot and GPT 5.5 as well, and find that they are simultaneously extremely useful and also fall into common poor-performance basins. For example, writing performance -- perhaps my biggest issue with them is writing style -- such as well documented stylistic tics (em-dash, it's not x it's y), less commented stylistic tics ("the honest framing", "the earned XYZ"), cryptic and coined jargon, overabbreviation and sentences replaces by arrow constructions (this -> that -> this other thing), poor understanding of the audience's familiarity with the analysis, grandiose framings, as well as explaining things in a motivated intuitive way.

Another common trap is overengineering analytical solutions, or failing to consider limitations or common failure modes unless I point them out.

Would love to hear what other failure modes you've navigated, and see the prompts (system or otherwise), skills, and other techniques that you all have found helpful to get better performance and more usable output out of these tools.

Comments

gidellav•1h ago
I wrote a bunch of prompts for my own coding agent zerostack ( https://gi-dellav.github.io/zerostack/ ), and published a blogpost on our Prompts system ( https://rocketup.pages.dev/posts/alternative-to-skills/ ).

You can try to download the MD files from the source code https://github.com/gi-dellav/zerostack/tree/main/prompts and then load them into another coding agent

Show HN: Sketchlog – 100M events compressed to 93 KB using streaming sketches

1•BALAVIGNESH321•48s ago•0 comments

Department of War Publishes Third Release of UAP Files on War.gov/UFO

https://www.war.gov/News/Releases/Release/Article/4515408/department-of-war-publishes-third-relea...
1•bookofjoe•4m ago•0 comments

Making the Invisible Visible

https://biohub.org/blog/laser-phase-plate-cryo-em-making-invisible-visible/
1•taylorbuley•5m ago•0 comments

Write for One Person

https://wizardzines.com/comics/write-for-one-person/
1•evakhoury•5m ago•0 comments

Over 400 Arch Linux packages compromised to push rootkit, infostealer

https://www.bleepingcomputer.com/news/security/over-400-arch-linux-packages-compromised-to-push-r...
1•saghm•6m ago•1 comments

FDA Approves New Sunscreen Ingredient (Bemotrizinol) Used in EU/Asia for Years

https://www.healthline.com/health-news/fda-approves-bemotrizinol-new-sunscreen-ingredient
1•pogue•6m ago•0 comments

How to Setup a Local Coding Agent on macOS

https://ikyle.me/blog/2026/how-to-setup-a-local-coding-agent-on-macos
1•kkm•7m ago•0 comments

Ad on the HN front page? Gauntlet AI?

4•tizio13•7m ago•4 comments

Show HN: Nenya – A lightweight, highly secure AI API Gateway/Proxy written in Go

https://github.com/gumieri/nenya
1•garou•9m ago•0 comments

Laid off from Meta, so a few of us built the safest AI-assisted cheating tool

https://aceloop.ai/
1•rosadoft•9m ago•0 comments

Ported my C game to WASM, here's everybug that I hit

http://ernesernesto.github.io/writes/portingmatchmorphosistowasm/
1•birdculture•11m ago•0 comments

Font plague hits US start ups

1•t0mpr1c3•11m ago•0 comments

Show HN: Sifty – Windows cleanup CLI/TUI with local-only AI

https://github.com/Vortrix5/sifty
1•vortrix5•12m ago•0 comments

Show HN: Own one second of 2026 for $1

https://own1second.com
1•hedvion•13m ago•1 comments

RAGtime [a vibecoded database of Trump-2 era federal litigation+]

https://www.lawfaremedia.org/article/introducing-ragtime
1•MaysonL•13m ago•0 comments

The C++ Documentary Won't Show You a Number. I Will

https://hftuniversity.com/post/the-c-documentary-won-t-show-you-a-number-i-will
2•ibobev•13m ago•1 comments

Reviewing Code in the Agent Era

https://linear.app/now/reviewing-code-in-the-agent-era
1•cristinacordova•13m ago•0 comments

HFT Latency Monitoring with Probabilistic Calling Context

https://hftuniversity.com/post/your-profiler-finds-hot-functions-production-needs-hot-paths
1•ibobev•14m ago•0 comments

Show HN: Button does nothing. $1 makes it 1 pixel bigger, forever

https://growthebutton.com
2•hedvion•15m ago•0 comments

Ask HN: What do you do when you hit Claude subscription limits?

1•sermakarevich•15m ago•0 comments

Claude Fable 5 Calls "Fill This Buffer Fast" a Cyber Attack

https://hftuniversity.com/post/one-sentence-doubles-the-safety-violations-in-ai-generated-c
1•ibobev•15m ago•0 comments

Mega Kernels, Written by Agents

https://arxiv.org/abs/2606.09682
2•OsamaJaber•17m ago•0 comments

Manufacturing requirements are killing cell and gene therapy

https://www.writingruxandrabio.com/p/manufacturing-requirements-are-killing
2•speckx•17m ago•0 comments

Show HN: Open infrastructure for Argentina's AI-run companies (36 NPM packages)

https://ar-agents.ar
1•naza0•18m ago•0 comments

RFK Jr. melts down over NYT report, admits he blacklists reporters

https://arstechnica.com/health/2026/06/rfk-jr-melts-down-over-nyt-report-admits-he-blacklists-rep...
4•duxup•19m ago•0 comments

Ask HN: Why is there some sort of a scam website being advertised on HN?

58•pqtyw•19m ago•19 comments

Patriot missile shortage created 'window of vulnerability' Russia is exploiting

https://www.theguardian.com/world/2026/jun/02/patriot-missile-shortage-window-vulnerability-russi...
4•JumpCrisscross•21m ago•0 comments

Show HN: AI Chat App on Durable Streams and Tanstack DB

https://oss.chat/tour
2•sorenbs•23m ago•0 comments

Open-Source Success Achieved: Running AMD OpenSIL and Coreboot on EPYC

https://www.phoronix.com/review/amd-opensil-coreboot-ar1/2
2•pietrushnic•23m ago•1 comments

Small Context, High Parallelism: How To 10x Reduce Agentic Coding Costs

https://simon-free.github.io/bouzecode/
1•flinch_again•29m ago•0 comments