frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Ask HN: What are your best ideas you will probably never build?

4•hsuduebc2•1h ago
I think a lot of people here have ideas they keep thinking about, but realistically will probably never build.

Maybe they are too small for a startup, too weird for work, too niche, too much effort, or just too far down the backlog.

I wonder if it would be fun to have a thread where people pitch those ideas.

If you have something you would like to see exist, but probably will not build yourself, post it here.

Could be software, hardware, data, art, local tools, research, weird experiments, anything.

I'm curious about your ideas!

Comments

davidloibner•1h ago
its a programming language, where code or its execution just behaves different depending on the case of use . a context aware compiler probably analyzes the environment, then it compiles in different ways. could also be hardware attached
davidloibner•1h ago
How about yours?
filup•1h ago
I want to build a modern version of a code book. Something that uses zero encryption at all, yet contains everything I could ever want to say.

It seems like once encryption took over the art of codebooks kinda went away but I still find the history of them really interesting and think a modern take would be fun to build.

toomuchtodo•1h ago
Use https://github.com/ndjordjevic/pin-llm-wiki as scaffolding (or a similar project to automate Karpathy's LLM Wiki pattern) to support keeping a wiki continually updated containing an HN user's comments updated as a Wiki knowledgebase, with the option to annotate or correct facts and similar context/knowledge accordingly. Automated knowledge base of sorts, grounded with the target user's comments (or multiple users, depending on whose comments you value).
mmarian•58m ago
Meetup alternative.

The two main Voynich Manuscript transcriptions disagree on 16% of lines

https://github.com/noah-chelednik/voynich-data
1•chedai__•2m ago•0 comments

ImpactArbiter – A PyTorch autograd trap for LLM-generated vLLM/SGLang invariants

https://github.com/msunda17/impactarbiter-cli
1•maniksundar•4m ago•0 comments

Empires Once Marched on Roads – AI marches on extension cords

https://earthchronicles.substack.com/p/empires-once-marched-on-roads
1•taguniversalm•5m ago•0 comments

Ex-DOGE Employees Raise $130 Mill for AI National Security Startup

https://www.vanityfair.com/story/doge-defense-startups
2•gaheehwjhwbs•6m ago•0 comments

Building Astro Websites with Almost No JavaScript – Introducing Webuum v0.x

https://webuum.dev/blog/webuum-0.x.html
1•lubomirblazekcz•8m ago•0 comments

Scott Alexander's AI Opinions

https://www.astralcodexten.com/p/my-ai-opinions
2•brettcvz•10m ago•0 comments

Can I Buy Your KV Cache?

https://arxiv.org/abs/2606.13361
2•MediaSquirrel•11m ago•0 comments

Before You Think: System 0, AI-Mediated Cognition and Cognitive Colonization

https://arxiv.org/abs/2606.13658
1•MediaSquirrel•13m ago•0 comments

What Is an LLM Control Plane?

https://blog.mozilla.ai/what-is-an-llm-control-plane/
1•angpt•15m ago•0 comments

3D Map That Acts as Commercial Vessel and Geopolitics Intel Platform

https://github.com/jamalrfordii-arch/Vanguard-Map
1•Lawyer24•15m ago•1 comments

Reasoning as Pattern Matching: Shared Mechanisms in Human and LLM Reasoning

https://arxiv.org/abs/2606.13607
1•MediaSquirrel•16m ago•0 comments

Show HN: Musefs – organize and tag music without touching the original files

https://github.com/Sohex/musefs
1•sohex•18m ago•0 comments

A Visual Guide to DiffusionGemma

https://newsletter.maartengrootendorst.com/p/a-visual-guide-to-diffusiongemma
1•speckx•18m ago•0 comments

Writing Constant-Time Rust Is Not Enough

https://emavan.com/blog/2026/constant-time-rust-llvm-aliasing/
1•emavan•19m ago•1 comments

The First Trillionaire Is a Killer

https://www.theverge.com/tech/949259/the-worlds-first-trillionaire-is-a-killer
4•cdrnsf•20m ago•1 comments

What Does It Feel Like to Live Under the Threat of Redundancy?

https://isrf.org/blog/what-does-it-feel-like-to-live-under-the-threat-of-redundancy
1•theanonymousone•20m ago•0 comments

NEURA: A Unified and Retargetable Compilation Framework for CGRAs

https://dl.acm.org/doi/10.1145/3808285
1•matt_d•21m ago•0 comments

Why the AI Renaissance Keeps Not Arriving

https://jamesfbaker.substack.com/p/why-the-ai-renaissance-keeps-not
4•jamesbaker1•23m ago•0 comments

Unified Contradiction‑Resolution Framework for Physics and Mathematics

https://zenodo.org/records/20671885
1•MatthewCarlo•26m ago•0 comments

NMOX Studio is being built by Fable

https://github.com/NMOX/NMOX-Studio
1•DavidCanHelp•28m ago•0 comments

Devirt.dev – generic JavaScript deobfuscator built as a compiler

https://devirt.dev/
2•vasie•30m ago•0 comments

General purpose LLMs outperform specialized clinical AI on medical benchmarks

https://www.nature.com/articles/s41591-026-04431-5
1•dnw•31m ago•0 comments

Show HN: Markdown Viewer for Mac Finder

https://quicklookmd.com/
1•jzone3•31m ago•0 comments

Swift at Apple: Migrating the TrueType Hinting Interpreter

https://www.swift.org/blog/migrating-truetype-hinting-to-swift/
1•DASD•31m ago•0 comments

China's Juno detector outpaces decades of research in 59 days (science.org)

https://www.science.org/content/article/first-results-put-neutrino-experiment-china-track-breakth...
3•Hypathia•36m ago•0 comments

Urban pollution in wealthy world still adding to heart damage, study finds

https://www.ft.com/content/2f86dced-7a1c-432e-9988-0592ebb8ac25
2•paulpauper•38m ago•0 comments

Parking Spot Is Free. Should It Be?

https://www.nytimes.com/interactive/2026/06/09/nyregion/nyc-street-parking.html
5•paulpauper•39m ago•0 comments

Kagi Magic

https://kagi.com/magic
38•amirmasoudabdol•39m ago•14 comments

Tyler Cowen: Is Mexico Safe Enough for the World Cup?

https://www.thefp.com/p/tyler-cowen-is-mexico-safe-enough
1•paulpauper•39m ago•0 comments

US and Iran have agreed to wording of a deal to end their war

https://apnews.com/article/iran-us-ceasefire-hezbollah-israel-12-june-2026-7085e386e1c40ee6cfe634...
3•geox•39m ago•0 comments