frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Ask HN: Should AI credits be refunded on mistakes?

3•ed_elliott_asc•1h ago
Something I’ve noticed (I’m using Claude subscription so no refunds but it applies to usage windows) is that sometimes AI makes mistakes so if something is important I tell Claude code to spin up a couple of sub agents and verify the information, often there will be a mistake and it gets rectified.

It feels unfair I have to pay (or lose some usage) for this.

Interested in other people’s thoughts.

Comments

sturza•1h ago
Do you pay your employer when you introduce bugs? I think you're lucky if you get usable output which you don't consider a mistake. Also, you might be mistaken if you think that you pay for a deterministic service.

edit: typo

meltyness•1h ago
It's interesting on the grounds of aligning incentives.

It's not interesting due to the fact that it suggests humans are still in the loop of some slow-cycle improvements. That'd never get by any board. In fact, selection of model modes implies it's your responsibility, so that meal was scraped into your flowerpot years ago.

I'd say fat chance.

verdverm•22m ago
No, you should know that no man or machine writes bug or mistake free code. You are paying for tokens (electricity and cooling), not what those tokens represent. How would you define mistakes in non code tasks?

Would you give money back to your employer when you make a mistake?

sloaken•2m ago
Interesting, you have just identified a potential market distinction. First we need a group (ala consumers report) to evaluate different services. Then different services would be motivated to perform the sub agent verification automatically as a Competitive Advantage,

How do you build a better, more decentralized web?

https://blog.archive.org/2026/04/02/how-dweb-camp-is-being-built-in-berlin/
1•the-mitr•2m ago•0 comments

QEMU Dev Starter guide, Part 1 [video]

https://www.youtube.com/watch?v=OCBLTMKLGAk
1•aleclm•2m ago•0 comments

Show HN: Scindo, AI agent that turns team discussions into shipped code

https://scindo.one
1•oneslashapps•5m ago•0 comments

LeaseGuard: Raft Leases Done Right

https://emptysqua.re/blog/leaseguard-raft-leader-leases-done-right/
1•g0xA52A2A•5m ago•0 comments

Free Chrome extension to see what employees send to AI tools

https://www.privent.ai/use-cases/employee-ai-monitoring
1•asilozyildirim•6m ago•0 comments

Catipult

https://www.catipult.ai
1•prequena•7m ago•0 comments

The Future of Everything Is Lies, I Guess

https://aphyr.com/posts/411-the-future-of-everything-is-lies-i-guess
2•pabs3•7m ago•0 comments

Show HN: Browser DevTools MCP Empire a tiny pixel world built in Cursor

https://github.com/serkan-ozal/browser-devtools-mcp-vscode-extension
4•erercan•7m ago•1 comments

Show HN: Live Cybernetics Maps, feedback loops in any organisation/problem

https://recursive.systems/
1•rgunther•8m ago•0 comments

Show HN: OpenFable – Open-source RAG engine using tree-structured indexes

https://github.com/alainbrown/openfable
1•alainbrown•8m ago•0 comments

TinyGo Boldly Goes Where No Go Ever Did Go Before

https://hackaday.com/2026/04/07/tinygo-boldly-goes-where-no-go-ever-did-go-before/
1•soypat•8m ago•0 comments

We Forgot How to Make Some Nuclear Bomb Parts

https://cdn.lanl.gov/files/nwj2-09_62c62.pdf#page=22
2•OhMeadhbh•9m ago•2 comments

We audited Claude Code and confirmed RCE via environment variable injection

https://audited.xyz/blog/claude-code
2•zack-eth•11m ago•1 comments

Tiny-TTS: The Smallest English TTS Model with Only 1M Parameters

https://github.com/tronghieuit/tiny-tts
1•thunderbong•15m ago•0 comments

LLM plays an 8-bit Commander X16 game using structured "smart senses"

https://pvp-ai.russell-harper.com
2•russellharper•16m ago•0 comments

Show HN: Yolt – safer YOLO mode for LLMs (recover deleted or overwritten files)

https://github.com/mvanderlinde/yolt
2•marc92•16m ago•1 comments

A systematic search for wormholes across all public astronomical data

https://www.blankline.org/research/hunting-for-real-wormholes-we-searched-every-available-dataset
2•DarenWatson•18m ago•0 comments

Show HN: I kept missing Claude Code prompts, so I built a tray dot in Rust

https://github.com/sprklai/agenttray
1•rakeshd•18m ago•0 comments

500 MWe Nuclear Fast Breeder Reactor built by India attains first criticality

https://www.pib.gov.in/FactsheetDetails.aspx?Id=150617
2•saharshpruthi•18m ago•0 comments

Fish Climbing in the Upper Congo Basin (Central Africa)

https://www.nature.com/articles/s41598-026-42534-8
1•Someone•19m ago•0 comments

Show HN: Heron is open-source security auditor that interviews your AI agents

https://github.com/theonaai/Heron
1•IlyaIvanov0•21m ago•1 comments

What Holds

https://www.kevinsdias.com/posts/what-holds.html
2•speckx•21m ago•0 comments

Show HN: Flowcost – know what your AI workflow is likely to cost

https://www.flowcost.ai/
1•erikms•21m ago•0 comments

Mythos Quest

https://thezvi.substack.com/p/ai-163-mythos-quest
1•7777777phil•23m ago•0 comments

Ask HN: What's Up with Azure DevOps

1•teekert•23m ago•0 comments

How to Set Up Your Monitoring System Alerts

https://blog.appsignal.com/2026/04/07/how-to-set-up-your-monitoring-system-alerts.html
1•malinavojvodic•25m ago•0 comments

Updu: Lightweight, self-hosted uptime monitoring in a single binary

https://github.com/nwpeckham88/updu
2•meysamazad•25m ago•0 comments

Used EV sales spike alongside gas prices

https://arstechnica.com/cars/2026/04/used-ev-sales-spike-alongside-gas-prices/
2•rbanffy•26m ago•1 comments

LLM inference engine from scratch in C++ – why output tokens cost 5x

https://www.anirudhsathiya.com/blog/transformer
6•ani17•28m ago•3 comments

Ask HN: How to learn web design in general

1•andrew-v•30m ago•1 comments