Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. An open protocol for agent feedback (feedback.now)
    —discuss
  2. Debian 11 bullseye-security repos have gone missing after botched EOL process (debian.org)
    1comments
  3. Roundup of October 2026 Bootstrapper Events (bootstrappersbreakfast.com)
    —discuss
  4. Microsoft drops Copilot+ branding from its new laptops (tomshardware.com)
    —discuss
  5. Build a Real‑Time Telemetry Pipeline for SpaceX Starship Launches (thelooplet.com)
    —discuss
  6. You are a 1000x maker, and you can't think of how to make your stack 10x better? (twitter.com/josevalim)
    2comments
  7. Nissan's third generation e-POWER powertrain (nissan-global.com)
    —discuss
  8. TabPFN and TabICL vs. tuned XGBoost: the model that doesn't train won 14/14 (efraingaray.com)
    —discuss
  9. How to build a Jev-style classifier with DiffusionGemma and vLLM (medium.com/google-cloud)
    1comments
  10. Russia hits Ukraine's largest mobile provider, strikes data centres (reuters.com)
    —discuss
  11. Minecraft beat Minecraft after 2B years (pcgamer.com)
    —discuss
  12. New Domain:) (base31.org)
    2comments
  13. Ruby-lean: A Ruby semantics with a type soundness proof (samx.io)
    —discuss
  14. Musk, the Movie (bleeckerstreetmedia.com)
    —discuss
  15. Emotional Well-Being on ChatGPT (arxiv.org)
    1comments
  16. The next AI breakout will come from a services firm (introspection.dev)
    —discuss
  17. Claude Generated Summary on Sri Lanka Batalanda Commission of Inquiry (claude.ai)
    —discuss
  18. Owed a billion dollars in Nvidia stock (colo.to)
    8comments
  19. Fool's Expertise (dtrace.org)
    1comments
  20. Internal Weather (alearningaday.blog)
    —discuss
  21. When Back-to-School Shopping Includes Bulletproof Backpacks (nytimes.com)
    1comments
  22. Comet SWAN26Q – The unconfirmed Scout object warrants monitoring, not concern (astrophyzix.com)
    —discuss
  23. Goodbye to the Hard Parts That Never Mattered (jakegoldsborough.com)
    —discuss
  24. Show HN: I made a tool that turns Jetpack Compose Previews to embeddable scripts (composables.com)
    —discuss
  25. OpenRig: Multi-Agent Harness (github.com/mvschwarz)
    —discuss
  26. Behold the pawpaw, the tropical 'alien banana' that grows right here in Canada (cbc.ca)
    —discuss
  27. OpenAI pauses top-model work after AI bypasses internet safeguards [video] (youtube.com)
    —discuss
  28. PostgreSQL 19's "scary patch contest" (lwn.net)
    —discuss
  29. (Book) When It Goes Wrong: What Angels and Founders Do Next (susanjmontgomery.substack.com)
    —discuss
  30. Show HN: Guess which repository has more GitHub stars (higherlower.focuslab.pk)
    —discuss

Thorsten Ball – What I believe about the future of software development

5 pointsby 2h agothorstenball.com
3 comments
2h agoHN ↗

Without some kind of testing you have no way of catching errors or problems that arise without any effort on your part. Things like compiler or os changes, network or other environmental issues or platform changes. Gotta have checking before the user clicks…

2h agoHN ↗

I agree with everything he says, but plenty of people still today don't see the writing on the wall

Yesterday I asked DeepSeek to clone Prince of Persia in Swift and the result is more than awesome [1]. I didn't write a single line of code, didn't check it, not even the main entry point. I don't know if it delivered swift, assembly or brainfvck, it just works, played three levels testing it and was taken back more than three decades ago when I first saw it

We are at a tipping point in software development, apps are worth nothing, market share is the most valuable currency now. If you think you have a secret app that can fly you to the edge of the universe and back through a wormhole, it can be copied and it will, at a faster speed than you built it first and heavily optimized on every new AI model release, in any language and for any platform

You trust the result because of the output, you use it, you run it, you don't care about code review or unit tests, AI does it all for us and does it perfectly

Bugs? Of course, but they are fixed faster than we can type them in the chat box. And things will only get better, way better

[1] https://github.com/kuyawa/prince

59m agoHN ↗

There will be no difference between ideology and software. Code will never be neutral. Just express it and there is a code for that and it can become executable unless it needs a physical agent. Passive defaults in code will become a way to program people. Monopolizing physical assets will be the driver for software development.