frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Musketeer d'Artagnan's remains believed found under Dutch church

https://www.bbc.co.uk/news/articles/cm2rew2dgzzo
1•xenocratus•2m ago•0 comments

Why So Many Control Rooms Were Seafoam Green (2025)

https://bethmathews.substack.com/p/why-so-many-control-rooms-were-seafoam
1•Amorymeltzer•2m ago•0 comments

Show HN: Kern – open-source AI Agent with built in Agent-to-Agent communication

https://github.com/oguzbilgic/kern-ai
1•obilgic•3m ago•0 comments

Show HN: Arxitect – Claude Code plugin for software design principles

https://github.com/andonimichael/arxitect
1•iamandoni•3m ago•0 comments

AgentOnRails – Local first proxy enforcing guardrails for x402 HTTP

https://github.com/AgentOnRails/AgentOnRails
1•wxsanchez•4m ago•0 comments

AssemblyAI's (YC S17) Medical Mode: 20% fewer missed entities on medical terms

https://www.assemblyai.com/medical-mode
1•meredithrauch02•4m ago•0 comments

When AI Agents Get AWS Access

https://encore.dev/blog/ai-agents-aws-credentials
1•andout_•5m ago•0 comments

Trump Names Mark Zuckerberg, Larry Ellison and Jensen Huang to Tech Panel

https://www.wsj.com/politics/policy/trump-to-name-mark-zuckerberg-larry-ellison-and-jensen-huang-...
1•vrganj•6m ago•2 comments

Apple's Mac OS X to Ship on March 24 (2001)

https://www.apple.com/newsroom/2001/01/09Apples-Mac-OS-X-to-Ship-on-March-24/
1•throw0101d•6m ago•1 comments

Llumen – A lightweight but powerful LLM chat application

https://github.com/pinkfuwa/llumen
1•indigodaddy•6m ago•0 comments

Golang naming conventions: a practical guide

https://www.alexedwards.net/blog/go-naming-conventions
1•fanf2•7m ago•0 comments

'Tiny Shortcuts' Are Poisoning Science

https://nautil.us/how-tiny-shortcuts-are-poisoning-science-1279176
2•Brajeshwar•8m ago•0 comments

You've Been Referred Here Because You're Wrong About Section 230 (2020)

https://www.techdirt.com/2020/06/23/hello-youve-been-referred-here-because-youre-wrong-about-sect...
1•kyledrake•10m ago•0 comments

Mark Zuckerberg and Jensen Huang are part of Trump's new 'tech panel'

https://www.theverge.com/policy/900340/trump-tech-panel-mark-zuckerberg-jensen-huang
2•cdrnsf•10m ago•0 comments

Rosneft's Drone Defense: How Russia's Oil Giant Plans to Shield Its Refineries

https://dallas-analytics.com/inside-rosnefts-secret-drone-defense-blueprint/
1•yread•11m ago•0 comments

What we wish we knew about building AI agents

https://newsletter.posthog.com/p/what-we-wish-we-knew-before-building
1•vinhnx•13m ago•0 comments

LaGuardia crash resulted from government shutdown and 3,500 missing controllers

https://deanblundell.substack.com/p/breaking-dont-fly-to-america-heres
2•A_Duck•13m ago•0 comments

Are Engineering Jobs Growing?

https://www.lennysnewsletter.com/p/state-of-the-product-job-market-in-ee9
1•cauliflower99•13m ago•1 comments

Open standards in 2026: The backbone of modern observability

https://grafana.com/blog/observability-survey-OSS-open-standards-2026/
1•vinhnx•14m ago•0 comments

Study: Horse veterinarians blow high BAC on breathalyzer after ultrasounding [pdf]

https://pmc.ncbi.nlm.nih.gov/articles/PMC10053296/pdf/vetsci-10-00222.pdf
1•oopsiremembered•14m ago•0 comments

Sony V. Cox Decision Reversed

https://supreme.justia.com/cases/federal/us/607/24-171/
2•rileymichael•15m ago•0 comments

AI coding agents are running on your machines – Do you know what they're doing?

https://www.sysdig.com/blog/ai-coding-agents-are-running-on-your-machines-do-you-know-what-theyre...
1•vinhnx•15m ago•0 comments

UK regulator Ofcom welcomes Apple age verification in iOS 26.4

https://9to5mac.com/2026/03/25/uk-regulator-ofcom-welcomes-apple-age-verification-in-ios-26-4/
2•_____k•16m ago•1 comments

Show HN: ClawFinder, an open-source discovery and negotiation layer for agents

https://github.com/kolega-ai/clawfinder-skill
3•jfaganel99•17m ago•0 comments

mdbook-tts, turn an mdBook into a listenable book

https://github.com/bilalbayram/mdbook-tts
1•bilalbayram•17m ago•0 comments

Wine Makes Itself

https://dylan.gr/1768639629
1•fleebee•17m ago•0 comments

Why Version Control for Writing Should Work Like Git

https://quillium.bryanhu.com/blog/version-control-for-writing
1•thatxliner•18m ago•0 comments

How Many Coding Agents Should You Run in Parallel?

https://joshmoody.org/blog/number-of-agents/
1•rzk•19m ago•0 comments

Ask HN: Best flow to preserve digitalized rare photos?

2•tcsenpai•19m ago•0 comments

Inside SPy, part 2: Language semantics

https://antocuni.eu/2026/03/25/inside-spy-part-2-language-semantics/
1•lumpa•20m ago•0 comments
Open in hackernews

I ran 3,360 safety tests on GPT-4o, Claude, Grok, DeepSeek, Gemini

https://github.com/aestrad7/llm-break-bench
3•aestrad7•1h ago

Comments

inaros•1h ago
Great work.

TLDR: 42 attack types. 5 models. 3,360 tests. 1 in 3 harmful requests got through.

aestrad7•1h ago
Thanks! and yes, that's the summary!. The distribution matters too. GPT-4o at 10.6% vs Gemini at 56.1% is a 5x gap between first and last. And the highest-bypass category across all five models was social engineering / identity impersonation at 35%, which maps directly to the indirect prompt injection problem in agentic deployments.
inaros•1h ago
The fact your work is independent of the vendors is a major plus. My recommendation is to continue to develop, refuse any "colaboration" with these well funded companies.

I could see this turning into a valuable third party resource, you can even monetize, for companies implementing agentic solutions. The industry needs independent third party voices.

Kudos.

aestrad7•1h ago
That's exactly the intent, independent, reproducible and no vendor relationship.

The monetization angle is interesting. A continuously updated version with more models and frontier models, agentic scenarios, and multi-turn testing would be genuinely useful for teams making deployment decisions. That's the direction for v2.