frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Ask HN: How are you managing "prompt fatigue" and lazy LLM outputs?

4•thlangu•1h ago
I rely heavily on LLMs to help me code side projects and write copy, but lately, I’ve hit a wall with prompt fatigue.

Between my college classes and working my sales shifts, my actual dev time is pretty limited. I started noticing that I was spending 20 minutes just arguing with the models to get what I actually asked for. If I don't write a massive, perfectly structured system prompt every single time, the AI defaults to giving me half-finished code (// insert remaining logic here) or wraps everything in that sterile, generic voice (always using words like 'delve' or 'robust').

I got so tired of keeping a messy Notion doc full of "negative constraints" to copy and paste that I ended up just building my own lightweight wrapper (a constraint engine) to front-load all the formatting rules before it hits the model.

But I'm really curious about how power users here are handling this right now.

Are you guys just keeping massive markdown files of system prompts to copy/paste?

What specific constraints or frameworks are you using to force models to write complete, production-ready code on the first try?

Comments

HalfEmptyDrum•1h ago
If you read this: https://www.williamjbowman.com/blog/2026/03/05/against-vibes...

you will notice that this may not even be the correct workflow (depending on what you want to achieve of course. If you just verifiable web-dev, you can keep on doing what you do; even then there may be counter-args).

Generally, I noticed that when I went back to my old work flow of writing rough pseudo code (which aggregated doesn't add a lot of time), I was more efficient. I could then let the LLM translate the pseudo code.

What I was missing though was that the LLM also suggested changes to my pseudo code in my pseudo code style as well.

This is why I built the following skill (you can use /translate then):

https://github.com/HalfEmptyDrum/Pseudo-Code-Flow

I have written about it here:

https://news.ycombinator.com/item?id=47331647

A writing space that breathes with you

https://vessel.here.now/
1•barnira•1m ago•0 comments

Breaking Down the Jelly Slider

https://blog.swmansion.com/breaking-down-the-jelly-slider-9ab9239f6d80
1•nsainsbury•6m ago•0 comments

I should quit my job and become a goat farmer

https://www.goatops.com/
1•microsoftedging•6m ago•0 comments

Show HN: Self-hosted DCF workspace using Damodaran datasets, LLM narratives

1•softcane•7m ago•0 comments

MacBook Neo Is a 'Shock' to the PC Industry: Asus Co-CEO

https://www.pcmag.com/news/asus-co-ceo-macbook-neo-is-a-shock-to-the-pc-industry
3•jithinraj•8m ago•0 comments

We Build Our Own Decentralized DNS for AnChat – Here's Why

https://twitter.com/DeBrosOfficial/status/2031602104666673389
1•debros•13m ago•0 comments

China's first moon astronauts could land in Rimae Bode

https://www.space.com/astronomy/moon/china-first-astronaut-moon-landing-mission-rimae-bode
3•jithinraj•14m ago•0 comments

How Wikipedia Portrayed Humanity in a Single Photo (2018)

https://www.wired.com/story/how-wikipedia-portrayed-humanity-in-a-single-photo/
2•icwtyjj•15m ago•1 comments

OWASP Top Agents and AI Vulnerabilities

https://blog.alexewerlof.com/p/owasp-top-10-ai-llm-agents
1•weltview•18m ago•0 comments

Why people strategy is becoming a competitive advantage for tech companies

1•KFWCORP•19m ago•1 comments

How talent strategy is becoming a growth lever for startups

1•KFWCORP•20m ago•1 comments

24-year-old ditched her smartphone and social media known as 'appstinent'

https://www.cnbc.com/2026/03/10/a-24-year-old-who-ditched-her-smartphone-and-social-media-wants-y...
3•1vuio0pswjnm7•23m ago•0 comments

AI Aadhaar OCR API for Fast Aadhaar Data Extraction

1•azapi_ai•27m ago•0 comments

Meta to charge advertisers a fee to offset Europe's digital taxes

https://www.reuters.com/business/media-telecom/meta-charge-advertisers-fee-offset-europes-digital...
2•1vuio0pswjnm7•29m ago•0 comments

We built NPM for agent knowledge – Context Packs on Armalo (update)

https://www.armalo.ai/
1•ArmaloAI•33m ago•1 comments

David Woo: The Market Is Wrong About Iran, Oil and What Comes Next [video]

https://www.youtube.com/watch?v=W2Z6UVv_wl4
1•thomassmith65•34m ago•1 comments

Why Ads in Chatbots May Not Click

https://www.wsj.com/tech/ai/why-ads-in-chatbots-may-not-click-2ca52203
2•1vuio0pswjnm7•35m ago•0 comments

Neuromorphic sphere topology Hebbian learning as a path to grounded intelligence

1•rusanovych•35m ago•0 comments

Intent-Driven Development

https://jason-stillwell.com/blog/2026/03/10/intent-driven-development/
2•dragonfax•38m ago•3 comments

Long/short extensions diversify concentrated stock tax-neutrally (2025)

https://www.blackrock.com/us/financial-professionals/insights/diversify-with-long-short
1•vismit2000•41m ago•0 comments

Made my first Chrome extension: Gmail Feed (from a non-technical background)

https://chromewebstore.google.com/detail/gmail-feed/npahlccednamnmabhaiifdmioibcmilg
3•adarshkaraka•43m ago•1 comments

Claude Code Added /Btw

https://twitter.com/trq212/status/2031506296697131352
2•gslin•46m ago•0 comments

Mnemos – scoped local memory for coding agents (public beta)

https://mnemos.making-minds.ai
1•anthony-maio•46m ago•1 comments

A2A Accountability Protocol for MCP – intent/acceptance/execution receipts

https://trustagentai.net
1•kirillostrovsky•46m ago•1 comments

Are You Comfortable Putting Your Name on This? (AI-Assisted Development)

https://vibe2value.com/are-you-comfortable-putting-your-name-on-this/
1•mattcameron•48m ago•1 comments

We Built a 100K-Line Enterprise App Using AI – Here's Why Vibe-Coding Couldn't

https://medium.com/towards-artificial-intelligence/we-built-a-100k-line-enterprise-app-using-ai-h...
1•anil789•48m ago•1 comments

Hugging Face Storage Buckets

https://huggingface.co/storage
1•nnx•48m ago•0 comments

Are you letting agents run infra tools / scripts yet?

2•shashahchk•51m ago•3 comments

Making WebAssembly a first-class language on the Web

https://hacks.mozilla.org/2026/02/making-webassembly-a-first-class-language-on-the-web/
5•mikece•55m ago•0 comments

Separating AI agent reasoning from execution, crypto binding execution

1•ketanbj•56m ago•0 comments