frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open-source OWASP security testing for AI models and agents

https://github.com/tessera-ops/tessera
1•alysheaib•12s ago•1 comments

Show HN: HowMuchGameHeroesMake

https://howmuchgameheroesmake.vercel.app/
1•leonagano•1m ago•0 comments

I Tried to Invent a Better Replication Policy. It Failed

https://halil.cetiner.me/nearsight/
1•bayneri•2m ago•0 comments

A Eulogy for Vim

https://drewdevault.com/2026/03/25/2026-03-25-Forking-vim.html
2•mtts•3m ago•0 comments

TeamMind – persistent memory for Claude Code (no API key, runs locally)

https://github.com/natedemoss/Teammind
1•natedemoss•3m ago•0 comments

The Cost of Doing Business

https://pluralistic.net/2026/03/25/fact-intensive/
1•hn_acker•4m ago•0 comments

Show HN: Marco, a privacy-first, offline-first email client (IMAP-native, no AI)

https://marcoapp.io/
2•isaachinman•4m ago•0 comments

A 500K-parameter system that recovers invariant physics from observation alone

https://erebus.org/
1•ordinarily•4m ago•0 comments

Quantization from the Ground Up

https://ngrok.com/blog/quantization
2•samwho•4m ago•0 comments

Dan rewrote chardet, relicensed to MIT. Original author broke 15-year silence

https://www.elvex.com/podcast/he-rewrote-chardet-with-claude-the-internet-blew-up-heres-his-take
3•sak84•4m ago•0 comments

Show HN: Dbt-skillz compiles your dbt project into a Claude Code skill

https://github.com/atlasfutures/dbt-skillz
2•davidvgilmore•4m ago•0 comments

Lyria 3 Pro: Create longer tracks in more Google products

https://blog.google/innovation-and-ai/technology/ai/lyria-3-pro/
1•meetpateltech•5m ago•0 comments

Closing the knowledge gap with agent skills

https://developers.googleblog.com/closing-the-knowledge-gap-with-agent-skills/
1•xnx•6m ago•0 comments

RSA and Python

https://xnacly.me/posts/2023/rsa/
1•ibobev•6m ago•0 comments

Installing PyTorch with AMD ROCm on GNU/Linux

https://www.wedesoft.de/graphics/2026/03/24/rocm-torch-install/
1•ibobev•7m ago•0 comments

GitHub Nukes 900 Anime Piracy Repos and Forks, but Rejects Circumvention Claims

https://torrentfreak.com/github-nukes-900-anime-piracy-repos-and-forks-but-rejects-circumvention-...
1•t-3•8m ago•0 comments

Ask HN: Will juniors still learn coding the hard way?

1•QubridAI•8m ago•0 comments

Writing an LLM from scratch, part 32g – Interventions: weight tying

https://www.gilesthomas.com/2026/03/llm-from-scratch-32g-interventions-weight-tying
1•ibobev•8m ago•0 comments

Introducing DDD to Your Organization

https://docs.eventsourcingdb.io/blog/2026/03/26/introducing-ddd-to-your-organization/
1•goloroden•9m ago•0 comments

In Math, Rigor Is Vital. But Are Digitized Proofs Taking It Too Far?

https://www.quantamagazine.org/in-math-rigor-is-vital-but-are-digitized-proofs-taking-it-too-far-...
1•tzury•9m ago•0 comments

Prediction trading is coming to Canada

https://www.theglobeandmail.com/business/article-wealthsimple-clears-regulatory-hurdle-to-bring-p...
1•jprs•10m ago•1 comments

I Am the Bottleneck

https://www.bretmorgan.me/writing/2026/03/24/i-am-the-bottleneck/
1•strooltz•10m ago•0 comments

New American dream may come with a broker, a balance sheet and now an AI copilot

https://refreshmiami.com/news/the-new-american-dream-may-come-with-a-broker-a-balance-sheet-and-n...
1•lifenautjoe•10m ago•0 comments

The demise of public key encryption will come sooner than thought, Google warns

https://arstechnica.com/security/2026/03/google-bumps-up-q-day-estimate-to-2029-far-sooner-than-p...
1•ooboe•10m ago•1 comments

UK teenagers to trial six-week social media curbs for major study

https://www.theguardian.com/uk-news/2026/mar/25/hundreds-of-uk-teenagers-to-trial-six-week-social...
1•chrisjj•11m ago•0 comments

Should a conscious AI be given Human rights?

1•PalantirDroned•11m ago•0 comments

Miracle – old school cool [video]

https://www.youtube.com/watch?v=EVFmVPWKv4c
1•marysminefnuf•11m ago•0 comments

Mini Brains Just Learned to Solve a Classic Engineering Problem

https://singularityhub.com/2026/03/24/these-mini-brains-just-learned-to-solve-a-classic-engineeri...
1•Brajeshwar•11m ago•0 comments

Death by Clawd

https://deathbyclawd.com
2•speter•14m ago•0 comments

Google's extreme AI compression paper was on arXiv since April 2025

https://arxiv.org/abs/2504.19874
1•fadijob•15m ago•1 comments
Open in hackernews

OpenAI's latest repo has Claude as the third top contributor

https://twitter.com/CodeByNZ/status/2036723050197012771
56•mirzap•1h ago

Comments

mirzap•1h ago
Repo: https://github.com/openai/parameter-golf
taspeotis•1h ago

    OpenAI's latest repo has Claude as the third top contributor (twitter.com/codebynz)
    13 points by mirzap 19 minutes ago | flag | hide | past | favorite | 2 comments
#9 here?

https://github.com/openai/parameter-golf/graphs/contributors

aestrad7•1h ago
Lol, Claude Code genuinely is the best coding assistant as a product regardless of underlying model
devin•1h ago
Claude's contribution is 2 commits adding 4500 lines and removing 0.

I don't think people understand what this repo even is.

Claude's contribution is not significant any way you slice it.

Go look at the other users who are committers. There are giant logfiles. Other committers are +28,000 LOC on 2 commits.

Here's an example commit from another user: https://github.com/openai/parameter-golf/commit/985568880b60...

moralestapia•51m ago
Yeah, that's a great analysis you put together.

This is a company aiming for a trillion dollar public valuation, mostly built brand on and hype.

This tweet has been seen 100k times so far, plus it reached the front page of HN.

One would think they would be much more careful with their brand.

devin•49m ago
Who are you referring to? Claude? I don't understand what you're trying to say.
moralestapia•38m ago
I am happy to make my comment much more accesible.

Yeah, that's a great analysis you (@devin) put together.

This is a company (OpenAI) aiming for a trillion dollar public valuation, mostly built on brand and hype.

This tweet (the one linked in the submission, also adding it here [1]) has been seen 100k times so far, plus it reached the front page of HN (Hacker News).

One (indefinite pronoun to refer to a generic subject) would think they (OpenAI) would be much more careful with their brand.

If there's anything else you might think needs clarification, let me know and I'll be glad to help!

1: https://x.com/CodeByNZ/status/2036723050197012771

devin•26m ago
So you were just mocking the format of my post because it's reminiscent of AI-generated text? What I wrote was perfectly clear and in no way AI-generated.
Aurornis•44m ago
This repo is for a competition where they accept user-contributed code.

Everyone reading the Tweet or headline has incorrectly assumed that OpenAI employees were using Claude.

This is a perfect example of why you should always read past the headline or tweet at least a little bit.

andito•35m ago
I'm pretty sure they scrubbed the commit history to remove claude. Github caches the top contributors in the main page so even if you rewrite the history they remain. And in this repo you're right that if you go to the history claude's contribution shouldn't get it so high, so it follows that they scrubbed it.
datadrivenangel•1h ago
Claude code is impressively good. Some of the other models and harnesses get close, but it's a noticeable regression when I hit a limit and move to kilocode with gpt-oss:120b running locally.
messh•57m ago
but if you move to gpt-5.4 ?
sabareesh•59m ago
Codex usually dont add itself as contributor so this is misleading .
arjie•58m ago
* Having Claude add itself as co-author is a pretty smart marketing move.

* I have more respect for organizations that are willing to use competitor tools where necessary.

They should definitely be modifying Codex's commit messages to be similar.

Aurornis•54m ago
> * I have more respect for organizations that are willing to use competitor tools where necessary.

This repo is for an open competition. They accept PR submissions. The Claude submissions presumably didn’t come from OpenAI employees.

arjie•47m ago
Ah, thanks for that. Changes the story. So that leaves Meta as the only AI lab known to use a competitor lab's (Claude) products extensively.
tosh•51m ago
I would argue it is an anti-pattern and irritating the core audience they want to reach with Claude Code
ciroduran•20m ago
I have mixed feelings with this. I've just checked out my GitHub repo and just realised that Claude itself inserted as a contributor without giving me a heads up it would do it.
Aurornis•56m ago
They’re running an open competition. They merge user-contributed submissions into the repo. Claude was attributed for 2 of them.

This is a huge nothingburger.

Tiberium•36m ago
The only top comment in the thread that actually read past the title..
pier25•52m ago
https://xcancel.com/CodeByNZ/status/2036723050197012771
Growtika•50m ago
This feels like one of those stories that spreads because it’s too good not to.

People see that Claude is a top contributor in OpenAI related repo and jump to conclusions about what it means

Same with the Ballmer/Mac story. It sort of happened, but not really how people tell it. The Mac was there, but he wasn’t actually using it the way the story implies.

https://www.engadget.com/2008-04-30-steve-ballmer-uses-a-mac...

aubanel•48m ago
This tweet was just reusing an earlier tweet (same image) without attribution : https://x.com/andimarafioti/status/2036107240420032874

Update the link if possible?