frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Smithy API models for all AWS services

https://aws.amazon.com/blogs/aws/introducing-aws-api-models-and-publicly-available-resources-for-aws-api-definitions/
1•stich0602•48s ago•0 comments

I built a tool called BitCache – that backs up locally BitLocker keys from Entra

https://github.com/pawellakomski/bitcache
1•pawellakomski•1m ago•1 comments

Show HN: Lightweight Durable Workflows Built on Postgres

https://github.com/dbos-inc/dbos-transact-python
1•qianli_cs•1m ago•0 comments

Medroid: AI Doctor, Always on Call

https://app.medroid.ai/join-founders
1•raja159•3m ago•1 comments

Country Meal

https://kivuhub.net/job/street-child-country-meal-manager-kinshasa/
1•aganzeeliud•3m ago•0 comments

How do you wash your clothes in space?

https://www.bbc.co.uk/newsround/57751277
1•thunderbong•5m ago•0 comments

Blood clotting discovery opens "whole new chapter in vascular biology"

https://newatlas.com/disease/new-blood-clotting-mechanism-hemolysis-necroptosis/
3•Brajeshwar•12m ago•0 comments

America tried to ban fake photos in 1912

https://www.freethink.com/the-digital-frontier/fake-photo-ban-1912
2•Brajeshwar•13m ago•0 comments

Some thoughts on human-AI relationships

https://reservoirsamples.substack.com/p/some-thoughts-on-human-ai-relationships
1•tosh•14m ago•0 comments

The Chicxulub Asteroid Impact and Mass Extinction

https://www.science.org/doi/10.1126/science.1177265
3•simonebrunozzi•18m ago•0 comments

Wise goes to the US. Will its founder's supercharged voting rights follow?

https://www.theguardian.com/business/nils-pratley-on-finance/2025/jun/05/wise-goes-to-the-us-will-its-founders-supercharged-voting-rights-follow
1•colinprince•21m ago•0 comments

First Map Made of a Solid's Quantum Geometry

https://www.quantamagazine.org/first-map-made-of-a-solids-secret-quantum-geometry-20250606/
1•rbanffy•21m ago•0 comments

Show HN: Poseidon, an Oncall AI agent (open source alt to resolve.ai)

https://github.com/harshadmanglani/poseidon
1•harshadmanglani•22m ago•0 comments

Big Mouth Billy Bass now talks back with AI

https://boingboing.net/2025/06/06/big-mouth-billy-bass-now-talks-back-with-ai.html
2•cmsefton•24m ago•0 comments

Ask HN: Will something replace transformers and GPUs in the next 10 years?

1•melenaboija•26m ago•1 comments

Breakthrough in search for HIV cure leaves researchers 'overwhelmed'

https://www.theguardian.com/global-development/2025/jun/05/breakthrough-in-search-for-hiv-cure-leaves-researchers-overwhelmed
6•colinprince•26m ago•0 comments

Indirect strengthening through reactivation of contextually bound memories

https://www.nature.com/articles/s44271-025-00250-5
1•PaulHoule•27m ago•0 comments

Defence Tech: Collective Property, Private Control

https://www.lrb.co.uk/the-paper/v47/n10/laleh-khalili/collective-property-private-control
1•mitchbob•27m ago•1 comments

Odd Couple: Gas Giants and Red Dwarfs

https://www.centauri-dreams.org/2025/06/06/odd-couple-gas-giants-and-red-dwarfs/
2•JPLeRouzic•29m ago•0 comments

Fresh strain of pro-Russian wiper flushes Ukrainian critical infrastructure

https://www.theregister.com/2025/06/06/pro_russia_wiper_ukraine/
2•rntn•29m ago•0 comments

Show HN: A minimal analytics tool, using Redis and Node.js

https://github.com/orangecoloured/analog
1•rckt•31m ago•0 comments

JPMorgan has had enough of grads accepting future-dated roles elsewhere

https://fortune.com/2025/06/06/jamie-dimon-jpmorgan-fire-talent-who-accept-roles-within-18-months/
2•thehoff•32m ago•1 comments

Show HN: Lovable/Bolt Security Scanner

https://vibe-eval.com/
1•alex_mia•32m ago•0 comments

An AI Tools Website for All Those Tools You Dreamt of as a Child

https://klutz.netlify.app
1•Jeff9James•32m ago•1 comments

Shared Memory Across GenAI Tools

https://base0.ai/
1•surya_m95•32m ago•1 comments

Show HN: Tape/Z – a toolkit for analysing z/OS assembler (HLASM) code

https://github.com/avishek-sen-gupta/tape-z
2•armorer•36m ago•0 comments

IRS open-sources Direct File tax software amid political and industry pushback

https://www.zdnet.com/article/irs-open-sources-direct-file-tax-software-amid-political-and-industry-pushback-heres-why/
2•CrankyBear•39m ago•1 comments

Controversial genetics testing startup Nucleus Genomics raises $14M Series A

https://techcrunch.com/2025/01/30/controversial-genetics-testing-startup-nucleus-genomics-raises-14m-series-a/
1•ZeljkoS•39m ago•0 comments

See how a dollar would have grown over the past 94 years [pdf]

https://www.newyorklifeinvestments.com/assets/documents/education/investing-essentials-growthofadollar.pdf
12•mooreds•39m ago•12 comments

Ask HN: How are you using Markdown files these days?

1•fred_terzi•41m ago•2 comments
Open in hackernews

Show HN: Ask-human-mcp – zero-config human-in-loop hatch to stop hallucinations

https://masonyarbrough.com/blog/ask-human
105•echollama•18h ago
While building my startup i kept running into the issue where ai agents in cursor create endpoints or code that shouldn't exist, hallucinates strings, or just don't understand the code.

ask-human-mcp pauses your agent whenever it’s stuck, logs a question into ask_human.md in your root directory with answer: PENDING, and then resumes as soon as you fill in the correct answer.

the pain:

your agent screams out an endpoint that never existed it makes confident assumptions and you spend hours debugging false leads

the fix:

ask-human-mcp gives your agent an escape hatch. when it’s unsure, it calls ask_human(), writes a question into ask_human.md, and waits. you swap answer: PENDING for the real answer and it keeps going.

some features:

- zero config: pip install ask-human-mcp + one line in .cursor/mcp.json → boom, you’re live - cross-platform: works on macOS, Linux, and Windows—no extra servers or webhooks. - markdown Q\&A: agent calls await ask_human(), question lands in ask_human.md with answer: PENDING. you write the answer, agent picks back up - file locking & rotation: prevents corrupt files, limits pending questions, auto-rotates when ask_human.md hits ~50 MB

the quickstart

pip install ask-human-mcp ask-human-mcp --help

add to .cursor/mcp.json and restart: { "mcpServers": { "ask-human": { "command": "ask-human-mcp" } } }

now any call like:

answer = await ask_human( "which auth endpoint do we use?", "building login form in auth.js" )

creates:

### Q8c4f1e2a ts: 2025-01-15 14:30 q: which auth endpoint do we use? ctx: building login form in auth.js answer: PENDING

just replace answer: PENDING with the real endpoint (e.g., `POST /api/v2/auth/login`) and your agent continues.

link:

github -> https://github.com/Masony817/ask-human-mcp

feedback:

I'm Mason a 19yo solo-founder at Kallro. Happy to hear any bugs, feature requests, or weird edge cases you uncover - drop a comment or open an issue! buy me a coffee -> coff.ee/masonyarbrough

Comments

throwaway314155•17h ago
Not certain that your definition of hallucination matches mine precisely. Having said that, this is so simple yet kinda brilliant. Surprised it's not a more popular concept already.
loloquwowndueo•17h ago
- someone sets up an “ask human as a service mcp” - demand quickly outstrips offer of humans willing to help bots - someone else hooks up AI to the “ask human saas” - we now have a full loop of machines asking machines
TZubiri•16h ago
This is pretty much already possible in any economy, but quite a waste.

Not much is stopping you from buying products from a retailer and selling them at a wholesaler, but you'd lose money in doing so.

olalonde•11h ago
I built this - but mostly as a joke / proof-of-concept: https://github.com/olalonde/mcp-human
aziaziazi•10h ago
Cool project! Naive question: does mechanical turk uses llm now?
lordmauve•10h ago
Finally, the "AI" turns out to be 700 Indians. We now have the full loop of humans asking machines asking humans pretending to be machines. Civilisation collapses
franky47•5h ago
AI stands for Actual Indians.
kajkojednojajko•3h ago
please do the promptful
conception•17h ago
What sort of prompt are you using for this?
kordlessagain•4h ago
The prompt is (mostly) built using the tool loads in the MCP server. In Python, the @mcp.tool() decorators provide the context of tool to the prompt, which is then submitted (I believe) with each call to the LLM.
rgbrenner•17h ago
Sounds similar to `ask_followup_question` in Roo
kjhughes•16h ago
Cool conceptually, but how exactly does the agent know when it's unsure or stuck?
Groxx•16h ago
The same way it knows anything else.

So not at all, but that doesn't mean it's not useful.

TZubiri•16h ago
So we are just pushing the issue to another, less debuggable layer. Cool.
kjhughes•16h ago
I'll try to give you credit for more than dismissing my question off-hand...

Yes, it may not need to know with perfect certainty when it's unsure or stuck, but even to meet a lower bar of usefulness, it'll need at least an approximate means of determining that its knowledge is inadequate. To purport to help with the hallucination problem requires no less.

To make the issue a bit more clear, here are some candidate components to a stuck() predicate:

- possibilities considered

- time taken

- tokens consumed/generated (vs expected? vs static limit? vs dynamic limit?)

If the unsure/stuck determination is defined via more qualitative prompting, what's the prompt? How well has it worked?

Groxx•15h ago
I don't believe[1] any of those are part of the MCP protocol - it's essentially "the LLM decided to call it, with X arguments, and will interpret the results however it likes". It's an escape hatch for the LLM to use to do stuff like read a file, not a monitoring system that acts independently and has control over the LLM itself.

(But you could build one that does this, and ask the LLM to call it and give your MCP that data... when it feels like it)

So you'd be using this by telling the LLM to run it when it thinks it's stuck. Or needs human input.

1: I am not anything even approaching deeply knowledgeable about MCP, so please, someone correct me if I'm wrong! There do seem to be some bi-directional messaging abilities, e.g. notification, but to figure out thinking time / token use / etc you would need to have access to the infrastructure running the LLM, e.g. Cursor itself or something.

threeseed•13h ago
You are trying to control a system that is inherently chaotic.

You can probably get some where by indeed running a task 1000 times and looking for outliers in the execution time or token count. But that is of minimal use and anything more advanced than that is akin to water divining.

kordlessagain•4h ago
The system is only nondeterministic (and a model of nondeterminism at that) when it's emitting tokens. It (the system) becomes completely deterministic when it calls a tool and a result is returned from the tool.

This is little different than how I wrote this. Now it is deterministic, when I hit reply.

echollama•13h ago
the reasoning aspect of most llms these days knows when its unsure or stuck, you can get that from its thinking tokens. It will see this mcp and call it when its in that state. Though this could benefit from some rules file to use it, although cursor doesn't quite follow ask for help rules, hence making this.
kjhughes•12h ago
Does all thinking end up getting replaced by calls to Ask-human-mcp then? Or only thinking that exceeds some limit (and how do you express that limit)?
aziaziazi•10h ago
I had the same question reading your post:

> (problem description) your agent […] makes confident assumptions

> (solution description) when it’s unsure

I read this as a contradiction: in one sentence you describe the problem as an agent being confident while hallucinating and in the next phrase the solution is that the agent can ask you if it’s unsure.

You tool is interesting but you may consider rephrasing that part.

mgraczyk•16h ago
If you are answering these questions yourself, why not just add something like this to your cursor rules?

"If you don't know the answer to a question and need the answer to continue, ask me before continuing"

Will you have some other person answer the question?

deadbabe•16h ago
Having another person answer the question is pretty much the obvious route this will go.
mgraczyk•16h ago
But then that means they are editing a markdown file on your computer? How is that meant to work?

I like the idea but would rather it use Slack or something if it's meant to ask anyone.

echollama•13h ago
this is mainly meant as a way to conversate with the model while you are programming with it. This is not meant to pull questions to a team but more to pair program. a markdown file is best for syntax in an llm prompt and also just easiest to have open and answer questions with. If i had more time and could i would build an extension into cursor.
mgraczyk•12h ago
Why not have the model ask in the chat? It's a lot easier to just talk to it than open a file. The article mentions cursor so it sounds like you're already using cursor?
echollama•11h ago
would probably work better, this is just how i threw it together as an internal tool a long time ago. i just improved it and shipped it to opensource it.
multjoy•1h ago
Conversate is not a word.
echollama•57m ago
yes it is
bckr•15h ago
I’ve tried putting “stop and ask for help” in prompts/rules and it seems like Cursor + Claude, up to 3.7, is highly aligned against asking for help.
ramesh31•13h ago
>If you are answering these questions yourself, why not just add something like this to your cursor rules?

What you are asking for is AGI. We still need human in the loop for now.

mgraczyk•12h ago
What I'm describing is a human in the loop. It's just a different UX, one that is easier to use and closer to what the model is trained to use.
ramesh31•2h ago
Human in the loop means despite your best efforts at initial prompting (which is what rules are), there will always be the need to say "no, that's wrong, now do this instead". Expecting to be able to write enough rules for the model to work fully autonomously through your problem is indeed wishing for AGI.
mgraczyk•1h ago
In my example, the human would be in the loop in exactly the same way as the technique in the article. The human can tell the model that it's wrong and what to do instead.

Tools like th one in the article are also "rules".

superb_dev•16h ago
This site is impossible to read on my phone. Part of the left side of the screen is cut off and I can’t scroll it into view
lobsterthief•15h ago
Same here
tyzoid•14h ago
Completely blank for me on mobile (javascript disabled)
banner520•14h ago
I also have this problem on my phone
rfl890•13h ago
Switching to desktop mode fixed it for me
kbouck•13h ago
Rotate phone to landscape
multjoy•2h ago
lol, no
threeseed•13h ago
> an mcp server that lets the agent raise its hand instead of hallucinating

a) It doesn't know when it's hallucinating.

b) It can't provide you with any accurate confidence score for any answer.

c) Your library is still useful but any claim that you can make solutions more robust is a lie. Probably good enough to get into YC / raise VC though.

echollama•13h ago
reasoning models know when they are close to hallucinating because they are lacking context or understanding and know that they could solve this with a question.

this is a streamlined implementation of a interanlly scrapped together tool that i decided to open-source for people to either us or build off of.

geraneum•12h ago
> reasoning models know when they are close to hallucinating because they are lacking context or understanding and know that they could solve this with a question.

I’m interested. Where can I read more about this?

threeseed•11h ago
> reasoning models know when they are close to hallucinating because they are lacking context or understanding and know that they could solve this with a question

You've just described AGI.

If this were possible you could create an MCP server that has a continually updated list of FAQ of everything that the model doesn't know.

Over time it would learn everything.

xeonmc•3h ago
Unless there is as yet insufficient data for meaningful answer.
exclipy•12h ago
Would be great if it pinged me on slack or whatsapp. I wouldn't notice if it simply paused waiting for the MCP call to return
spacecadet•6h ago
Easy enough to do with smolagents and fastmcp, its 20 lines of code.
atoav•9h ago
I am running an electronics/medialab in an university, the amount of fires bad electronics advice from LLMs caused already is probably non-zero.

It is amazing how bad LLMs are when it comes to reasoning about simple dynamics within trivial electronic circuits and how eager they are to insist the opposite of how things work in the real world is the secured truth.

spacecadet•5h ago
If the model responds with an obvious incorrect answer or hallucination, start over. Rephrase your input. Consider what output you are actually after... Adding to original shit output wont help you.
ddalex•5h ago
Why wouldn't a rag-enabled ai be faster and better then humans at answering these documentation-grounded questions ?
kordlessagain•4h ago
The same technique can be had by creating a "universal MCP tool" for the LLM to use if it thinks the existing tools aren't up to the job. The MCP language calls these "proxies".